Cellular autophagy is a prosurvival mechanism in the kidney against ischemia-reperfusion injury (IRI), but the molecular pathways that activate the autophagy in ischemic kidneys are not fully understood. Clusterin (CLU) is a chaperone-like protein, and its expression is associated with kidney resistance to IRI. The present study investigated the role of CLU in prosurvival autophagy in the kidney. Renal IRI was induced in mice by clamping renal pedicles at 32°C for 45 min. Hypoxia in renal tubular epithelial cell (TEC) cultures was induced by exposure to a 1% O2 atmosphere. Autophagy was determined by either light chain 3-BII expression with Western blot analysis or light chain 3-green fluorescent protein aggregation with confocal microscopy. Cell apoptosis was determined by flow cytometric analysis. The unfolded protein response was determined by PCR array. Here, we showed that autophagy was significantly activated by IRI in wild-type (WT) but not CLU-deficient kidneys. Similarly, autophagy was activated by hypoxia in human proximal TECs (HKC-8) and WT mouse primary TECs but was impaired in CLU-null TECs. Hypoxia-activated autophagy was CLU dependent and positively correlated with cell survival, and inhibition of autophagy significantly promoted cell death in both HKC-8 and mouse WT/CLU-expressing TECs but not in CLU-null TECs. Further experiments showed that CLU-dependent prosurvival autophagy was associated with activation of the unfolded protein response in hypoxic kidney cells. In conclusion, these data suggest that activation of prosurvival autophagy by hypoxia in kidney cells requires CLU expression and may be a key cytoprotective mechanism of CLU in the protection of the kidney from hypoxia/ischemia-mediated injury.
- kidney ischemia-reperfusion
- acute kidney injury
ischemia is a hypoxic condition of disrupting blood flow to the tissue, and the subsequent reperfusion is the restoration of blood flow to the ischemic tissue, which together result in tissue injury, termed “ischemia-reperfusion injury” (IRI). Renal IRI induces acute kidney injury (AKI) and is seen in many clinical situations, including kidney transplantation (80, 81) and cardiac surgery-associated AKI (29, 42, 78). In kidney transplantation, severe ischemic injury in donor kidneys from donation after cardiac death is a primary reason for either exclusion (2) or delayed graft function (10, 91) and is a high risk factor for both acute and chronic allograft rejection (75, 86) and graft loss (39, 75, 96). Cardiac surgery-associated AKI is characterized by an abrupt reduction in the glomerular filtration rate of the kidney after cardiac surgery (5, 78) and may contribute to the increased mortality as a small (0.3–0.5 mg/dl) or larger (>0.5 mg/dl) increase in serum creatinine after cardiac surgery correlates with a nearly 3- or 18-fold increase in 30-day mortality, respectively (42). Furthermore, cardiac surgery-associated AKI worsens long-term complications in patients who have additional risk factors such as advanced age, renal insufficiency, diabetes mellitus, systemic hypertension, or systemic lupus erythematosus (78). Unfortunately, there is no any effective therapy available to treat renal IRI as of today.
IRI is induced by ischemia or hypoxia followed by reperfusion or reoxygenation, both of which can induce apoptotic and necrotic cell death (79, 90). Evidence in the literature suggests that renal IRI is multifaceted and includes inflammatory and immune responses, oxidative stress, induction of cell death, and incomplete renal repair (1, 8, 48), but its pathogenesis is still not fully understood. It has been well established that cellular autophagy is activated in the response to pathological stresses (e.g., nutrient depletion or starvation) to rebalance the energy source through a catabolic self-degradative process of misfolded or aggregated proteins and/or damaged organelles [i.e., mitochondria, the endoplasmic reticulum (ER), and proxisomes] (27, 61). Recent studies have shown an increase in autophagic activity in the kidneys after IRI or in hypoxic kidney cells (35, 88, 95, 97). Although different biological functions of autophagy have been reported to associate with either cell survival or death under certain circumstances (40, 41, 44, 45), it has been clearly demonstrated that the activation of autophagy provides a protective mechanism in the kidney for cell survival during IRI (32, 35, 38). However, little is known of the molecular pathways that regulate prosurvival autophagy in the kidneys during IRI.
Clusterin (CLU) is a chaperone-like glycoprotein (3, 30), and its expression is significantly upregulated in the kidneys of animal models of unilateral ureteral obstruction and IRI (77, 83, 94) or in renal biopsies from rejected renal allografts or diseased kidneys (20). Upregulated CLU may play a protective role in the kidney because a deficiency in CLU expression in mice [CLU knockout (KO)] results in the development of age-related glomerulopathy (76), increases the susceptibility to renal IRI, or impairs renal repair after IRI (65, 101). Recently, we reported that CLU overexpression enhances cell survival via autophagy-dependent pathways in prostate tumor cells (99). The objective of the present study was to investigate if the cytoprotection of CLU in the kidney was mediated by the activation of prosurvival autophagy during IRI.
MATERIALS AND METHODS
Animals and cell cultures.
Both WT C57BL/6 (B6) and CLU KO (CLU−/−) strains of mice from the B6 background were received from the breeding colonies in the animal facility of the Jack Bell Research Centre (Vancouver, BC, Canada) (101). Green fluorescent protein (GFP)-light chain (LC)3 transgenic mice from the B6 background were a kind gift from Dr. Noboru Mizushima (Department of Physiology and Cell Biology, Tokyo Medical and Dental University, Tokyo, Japan) (53) and were subsequently backcrossed for 10 generations into B6-CLU−/− mice in our facility to generate GFP-LC3 CLU KO (GFP-LC3+/+CLU−/−) mice as a CLU-null control of GFP-LC3 CLU WT (GFP-LC3+/+CLU+/+) mice, respectively (99). In these GFP-LC3 transgenic mice, GFP-LC3 was ubiquitously expressed under the control of the CAG promoter, and the accumulation of GFP puncta represented the formation of autophagosomes or autophagy activation (53). Genotypes of both CLU KO and GFP-LC3 knockin strains of mice were confirmed by PCR as previously described (49, 99). All animals (male mice, 10–12 wk old) for the experiments were cared in accordance with Canadian Council on Animal Care guidelines under protocols approved by the Animal Use Subcommittee of the University of British Columbia.
The human proximal tubular epithelial cell (TEC) line (HKC-8 cells) was kindly provided by Dr. Daniel L. Sparks (Ottawa, ON, Canada) under permission from Dr. Lorraine Racusen (74). Mouse kidney cell cultures (MKC-1 cells) were generated by immortalizing primary TECs from WT B6 mice with origin-deficient simian virus 40 DNA as previously described (101). Both CLU-expressing TECs and CLU-negative control cells originated from the same CLU-null TECs by stable expression of either human (h)CLU isoform 1 cDNA using pHEX6300 vector (denoted as TEC-CLUhCLU) or empty pHEX6300 vector (denoted as TEC-CLU−/−) (65). Primary murine TECs were isolated from the kidney cortex of mice following a previously described protocol (34). In brief, the renal cortex was collected and minced in HBSS containing penicillin-streptomycin. The tissue fragments were washed twice with HBSS and then digested with 1 mg/ml collagenase type V (Sigma-Aldrich Canada, Oakville, ON, Canada) in HBSS at 37°C for 15 min with intermittent agitation of the tissue. The digested tissue was sieved through a 40-μm Cell Strainer (BD Falcon, BD Biosciences, Mississauga, ON, Canada). After being washed with HBSS and subsequently with complete K1+/+ medium as previously described (19), the sieved cells were seeded and grown in complete K1+/+ medium in a collagen-coated flask. Finally, the phenotypes of primary TECs were confirmed by their expression of E-cadherin and CD13 (alanine aminopeptidase) with FACS analysis. All TEC lines (HKC-8, MKC-1, TEC-CLUhCLU, and TEC-CLU−/−) and primary TECs were grown in complete K1+/+ medium.
Reagents and antibodies.
Both annexin V conjugated with phycoerythrin (annexin V-PE) and 7-amino-actinomycin D (7-AAD) were purchased from BD Biosciences. Bafilomycin A1 (Baf-A1) was from Sigma-Aldrich Canada. Spautin-1 was from Cellagen Technology (San Diego, CA).
The following primary antibodies were used in this study: goat polyclonal anti-CLU-α (C-18, Santa Cruz Biotechnology, Santa Cruz, CA), rabbit polyclonal anti-LC3B (no. 2775, Cell Signaling Technology, New England Biolabs, Whitby, ON, Canada), rabbit polyclonal anti-inositol-requiring element (IRE)1 [phosphorylated (p-)Ser724, Biorbyt, Cambridge, UK], rabbit polyclonal anti-IRE1-α (NB100-2324, Novus Biologicals, Littleton, CO), mouse monoclonal anti-β-actin (clone AC-40, Sigma-Aldrich Canada), and anti-GAPDH (clone GA1R, Epitope Biotech, Vancouver, BC, Canada). Secondary antibodies were as follows: horseradish peroxidase (HRP)-conjugated anti-goat IgG (sc-2020), anti-mouse IgG HRP (sc-2314), and anti-rabbit IgG HRP (sc-2313) (all from Santa Cruz Biotechnology).
Renal IRI model.
Renal IRI was induced in the left kidney. In brief, mice were anesthetized with a combination of ketamine (100 mg/kg) and xylazine (10 mg/kg) with isoflurane as needed. The left kidneys were exposed through a flank incision followed by the induction of ischemia in these kidneys by clamping renal pedicles at the body temperature of 32°C for 45 min. After the clamps were released, reperfusion of the kidneys was confirmed visually. To examine the increase in autophagic activity (GFP prunta) in the left kidneys of GFP-LC3+/+CLU−/− and GFP-LC3+/+CLU+/+ mice after reperfusion, nonischemic right kidneys in the same mice were kept as contralateral controls.
Induction of hypoxia in vitro.
Hypoxia in cultured kidney cells in complete K1+/+ medium was induced by incubation in a humidified hypoxic chamber (Coy Laboratory Products, Grass Lake, MI) pumped with a gas mixture of 1% O2-5% CO2-94% N2 at 37°C while identical (twin culture) cell cultures were grown a humidified CO2 incubator in an atmosphere containing 5% CO2-95% air (∼20% O2) at 37°C as normoxia controls.
Analysis of GFP-LC3 puncta formation.
GFP-LC3 mice are commonly used to monitor the induction of autophagic activity or autophagosome numbers, indicated by the increased number of GFP-LC3 puncta in tissue sections (54, 55). For the assessment of IRI-induced autophagy in the kidney, renal tissues after 24 h of ischemia-reperfusion were collected from both injured and contralateral kidneys, fixed with paraformaldehyde fixation solution (4% paraformaldehyde dissolved in 0.1 M Na-phosphate buffer, pH 7.4) for 4 h, and incubated overnight with 15% sucrose-PBS and 30% sucrose-PBS for 4 h. The tissues were then embedded in Tissue-Tek OCT medium and stored at −80°C. Sections (5–7 μm) were prepared with a cryostat and air dried at room temperature for 30 min. To examine hypoxia-induced autophagy in cultured kidney cells, kidney cells on glass coverslips were exposed to hypoxia or normoxia for 24 h followed by fixing in paraformaldhyde solution for 15 min and a rinse with PBS. Nuclei in both tissue sections and cultured cells were stained with 4′,6-diamidino-2-phenylindole (Cell Signaling Technology). GFP-LC3 puncta formation (green) and 4′,6-diamidino-2-phenylindole-stained nuclei (blue) in tissue sections or cultured cells were visualized using a Zeiss LSM 780 confocal microscope (Carl Zeiss, Thornwood, NY). The number of bright GFP-LC3 puncta was counted in each microscopic view, randomly selected in the region of renal cortex, and presented as an average of at least of 25 nonoverlapping fields (under ×1m000 magnification) per 1,500 μm2 for each kidney. For GFP-LC3 puncta quantification in cultured cells, cells displaying >15 bright fluorescent GFP-LC3 puncta were counted as positive, and 100–200 cells were analyzed from each sample in each experiment.
Semiquantitative analysis of renal injury.
Histological assessment of tubular injury in kidney sections was performed in a blinded fashion. Formalin-fixed and paraffin-embedded sections (5 μm thickness, longitudinal) were stained with periodic acid-Schiff reagents. The percentage of damaged tubules (combined both necrosis and vacuolization) in total tubules was counted in each view, randomly selected in regions of the renal cortex under ×100 magnification, and presented as an average of at least of 20 nonoverlapping fields for each kidney.
Analysis of cell apoptosis.
Apoptosis in cultured kidney cells was measured using FACS analysis following the manufacturer's protocol (BD Biosciences) as previously described (101), in which annexin V-PE stained early apoptosis and 7-AAD stained late apoptosis or necrosis. In a FACS graph, nonapoptotic (viable) cells were in the bottom left quadrant, necrotic cells were in the top left quadrant (7-AAD positive only), late apoptotic cells were in the top right quadrant (both annexin V and 7-AAD positive), and early apoptotic cells were in the bottom right quadrant (annexin V positive only). Briefly, monolayers of kidney cells were released by a brief incubation with trypsin-EDTA solution (Sigma-Aldrich Canada) and then incubated with annexin V-PE in 1× binding buffer for 15 min followed by 7-AAD in the dark. The intensity of fluorescence of apoptotic and necrotic cells was measured by flow cytometry and analyzed compared with background controls using CELLQUEST software (BD Biosciences).
Western blot analysis.
Western blot analysis of protein expression was routinely performed following a standard protocol in the laboratory. In brief, whole cellular protein extract was prepared by a brief sonication of pieces of the renal cortex or cell pellets in RIPA buffer [50 mM Tris·HCl (pH 8.0), 150 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate, and 0.1% SDS] containing protease inhibitor cocktail (Roche, Mannheim, Germany) and fractioned by 7% or 15% SDS-PAGE according to the molecular size of the target protein. After being transferred to nitrocellulose membranes (Bio-Rad Laboratories, Hercules, CA), the target proteins on the blot were identified using relevant primary antibodies along with HRP-conjugated secondary antibodies as described in Reagents and antibodies section and visualized by an enhanced chemiluminescence assay (Thermo Fisher Scientific, Rockford, IL). Blots were reprobed with anti-β-actin IgG (Sigma-Aldrich Canada) or anti-GAPDH antibody (Epitope Biotech) to confirm equal protein loading in each sample. Expression levels of the target proteins were measured using densitometry and presented as a ratio units of the target protein to β-actin or GAPDH on the same blots.
The expression of 84 unfolded protein response (UPR)-associated genes in cultured TECs was quantitatively examined using PCR array kits following the manufacturer's instructions (SABiosciences-QIAGEN, Valencia, CA). Each group (CLU-expressing TEC-CLUhCLU vs. CLU-null TEC-CLU−/−) included four different samples (TEC cultures grown at four separate times) that were randomly selected for the determination of the gene expression profile using PCR arrays. Total RNA from each sample was directly extracted and purified from the monolayer after 24 h of either hypoxia or normoxia using the RNeasy Microarray Tissue Mini kit (QIAGEN) and was converted to cDNA using a RT2 First Strand Kit (QIAGEN). The expression of selected genes was amplified by real-time PCR using RT2 Profile PCR arrays (QIAGEN). Data were analyzed using Web-based PCR Array Data Analysis Software (www.SABiosciences.com/pcrarraydataanalysis.php.). The functional gene grouping was listed based on the reference provided by the manufacturer's website (http://www.sabiosciences.com/rt_pcr_product/HTML/PAMM-089A.html).
ANOVA or t-tests (two-tailed distribution) with Prism GraphPad software (GraphPad Software, La Jolla, CA) were used as appropriate for comparisons between groups. Data were collected from each individual experiment or each mouse for statistical analysis. P values of ≤0.05 were considered significant.
CLU is required for autophagy activation and reduced renal tissue injury with IRI.
Several studies have demonstrated either the upregulation of CLU expression (65, 101) or the activation of autophagy (32, 35, 88) in tubular cells of the kidneys with IRI. To investigate if autophagy played a role in CLU cytoprotection during renal IRI (101), autophagic activity in WT kidneys compared with CLU-null kidneys was examined after IRI induction. As previously reported, the number of LC3-GFP prunta was markedly increased in tubules of the kidneys after 24 h of reperfusion (denoted as IRI, 86.4 ± 14.6) compared with that in contralateral kidneys (denoted as control, 25.5 ± 4.8) in WT LC3-GFP transgenic mice (n = 4, Fig. 1A, left; control vs. IRI in Fig. 1B, P = 0.0002 by two-tailed t-test), indicating the activation of autophagy by IRI in these kidneys. In contrast, in CLU KO LC3-GFP mice, IRI did not induce such significant autophagic activity in the kidneys, with no apparent difference in LC3-GFP prunta between the kidneys with IRI (20.8 ± 3.8) and contralateral controls (12.6 ± 6.5, n = 4, Fig. 1A, right; control vs. IRI in Fig. 1B, P = 0.0722 by two-tailed t-test). Similar to our previous study (101), WT kidneys displayed less tubular injury (11.3 ± 2.5%) than their CLU KO counterparts (32.6 ± 9.8%, n = 4, P = 0.0057 by two-tailed t-test) after the induction of IRI in LC3-GFP transgenic mice (Fig. 1C), linking autophagic activity with resistance to IRI in WT kidneys. To further quantify autophagic activity and its role in the kidneys during IRI, LC3 expression was examined by Western blot analysis in kidney samples from our previous study (101), in which nontransgenic WT mice were shown to be more resistant to renal IRI compared with CLU KO mice after 48 h of reperfusion. As shown in Fig. 1D, the levels of both LC3-I and LC3-II in protein extracts of the renal cortex from WT kidneys were significantly higher than those in CLU-null kidneys. Taken together, these data suggest that the autophagy activated by IRI in the kidneys is largely dependent on the expression of CLU and is associated with resistance to tissue injury.
CLU is required for activating autophagy and promoting cell survival in cultured kidney cells under hypoxia.
IRI is initiated by profound tissue hypoxic injury (22), and autophagy is activated in cultured renal TECs by exposure to a hypoxic environment (an atmosphere containing 1% O2) (35, 88). As shown in Fig. 2A, LC3B-II expression in human proximal tubular HKC-8 cells was increased after 24 h of incubation under a hypoxic atmosphere (1% of O2) compared with that in cultures at 95% of atmospheric O2 levels (∼20% of O2), suggesting the induction of autophagic activity in human proximal TECs by hypoxia in vitro. In additon, the activation of autophagy in hypoxic tubular cells was positively correlated with the upregulation of CLU expression (Fig. 2A). The relationship of hypoxia-induced autophagy with CLU expression was further verified in primary renal cortical epithelial cells isolated from WT or CLU KO mice. As shown in Fig. 2B, compared with control cells at normoxia, LC3B-II expression was markedly upregulated by hypoxia in WT cells, whereas there was a slightly increase in LC3B-II expression in hypoxic CLU KO cells. The hypoxia-induced autophagy in WT kidney cells was further confirmed in primary renal cortical epithelial cells isolated from LC3-GFP transgenic WT or CLU KO mice (Fig. 2C), in which the number of LC3-GFP prunta was significantly increased in hypoxic WT kidney cells (86.3 ± 21.4%) compared with cells in normoxia (21.5 ± 6.2%, n = 3, P = 0.0073, two-tailed t-test), which was not seen in CLU-null kidney cells (11.4 ± 5.2% vs. 15.4 ± 3.2%, n = 3, P = 0.3199; Fig. 2D).
CLU expression increases cell survival in cultured kidney cells exposed to a cytokine mixture (interferon-γ and TNF-α) (101). To verify if CLU mediates kidney cell survival under hypoxia, cell viability or death was examined in hypoxic CLU-expressing kidney cells compared with CLU-null control cells. As shown in Fig. 3, cell viability in cultured primary renal cortical epithelial cells isolated from WT or CLU KO mice was significantly decreased after 24 h of exposure to hypoxia, as indicated using WT cells (86.9 ± 5.95% under normoxia, n = 5, vs. 67.93 ± 4.17% under hypoxia, n = 9, P < 0.0001 by two-tailed t-test) or CLU KO cells (78.13 ± 5.31% under normoxia, n = 7, vs. 55.77 ± 8.2% under hypoxia, n = 9, P < 0.0001 by two-tailed t-test). Further statistical analyses showed that there were more viable WT cells than CLU KO cells under both normoxic and hypoxic conditions, as indicated by a two-tailed t-test (normoxia: P = 0.0226 and hypoxia: P = 0.0009) or two-way ANOVA (WT vs. KO cells, P < 0.0001). These results were further confirmed using human CLU-expressing CLU-null kidney cells (TEC-CLUhCLU) compared with CLU-null TEC-CLU−/− cells that had been previously established (65). As shown Fig. 4, CLU-expressing TEC-CLUhCLU cells were more resistant to hypoxia-induced cell death, as indicated by a viable cell population of 75.96 ± 3.0% in TEC-CLUhCLU cell cultures compared with 62.68 ± 1.81% in TEC-CLU−/− control cells (n = 3, P = 0.0028 by two-tailed t-test). Taken together, these data clearly indicate that CLU expression is required for to activate autophagy and increase cell survival in cultured kidney cells under hypoxia, which are consistent with less tissue injury and a higher degree of autophagy activation in WT kidneys compared with those in CLU-null kidneys (Fig. 1).
CLU-mediated autophagy protects kidney cells from hypoxia-induced cell death.
Autophagic activity links to both types of cell death (apoptosis and necrosis), and it either promotes cell survival or results in cell death in response to an adverse environment and stress (13, 67). As shown above, a higher level of autophagic activity was found in WT kidneys with less IRI (Fig, 1) and in CLU-expressing kidney cells with less cell death under hypoxia (Figs. 2–4). To reveal the role of autophagy in kidney cell survival in response to hypoxia, cell survival was examined in hypoxic kidney cells in the presence of autophagy inhibitors (Baf-A1 and spautin-1). Both human and mouse kidney cell cultures (HKC-8 and MKC-1) were treated with Baf-A1 (0, 5, or 10 nM) under either hypoxia or normoxia for 24 h. As shown by Western blot analysis, LC3B-II expression (a functional marker of autophagy) was upregulated by hypoxia in HKC-8 cells and suppressed by the presence of 10 nM Baf-A1 (Fig. 5A). Using FACS analysis, hypoxia induced cell death in both HKC-8 and MKC-1 cell cultures as expected (Fig. 5B), and the addition of Baf-A1 further reduced cell survival in hypoxic HKC-8 cells from 88.2 ± 1.5% in the absence of the inhibitor to 79.8 ± 4.2% in the presence of 10 nM Baf-A1 (n = 3, P = 0.0245 by one-way ANOVA), whereas the inhibitor did not attenuate the viability of cell cultures under normoxia (Fig. 5C). Similar results were seen in cultured MKC-1 cells (Fig. 5D). These data were further confirmed in cultured TEC-CLUhCLU compared with TEC-CLU−/− cells, which were treated with a different autophagy inhibitor, spautin-1. Again, Western blot analysis showed that hypoxia upregulated the expression of LC3B-II (as well LC3B-I) only in TEC-CLUhCLU cells, which was also inhibited by spautin-1 (Fig. 6A). As shown by FACS analysis in Fig. 6B, this inhibitor reduced cell survival in TEC-CLUhCLU cells under normaxia and more profoundly under hypoxia (Fig. 6C), which was similar to Baf-A1-treated MKC-1 cells (Fig. 5). In contrast, the autophagy inhibitor did not show significant impact on the cell death of CLU-null TEC-CLU−/− cells under either normoxia or hypoxia (Fig. 6D). These data suggest that the cytoprotection conferred by CLU in kidney cells is mainly mediated by autophagy activation.
CLU-mediated autophagy correlates with an increase in the UPR in hypoxic kidney cells.
Acute ischemia induces ER stress and subsequently activates the UPR in the kidney (25, 57). The UPR, autophagy, and proteasomal degradation together degrade and recycle excess or damaged proteins to restore cellular homeostasis and/or inactivate the cell death (i.e., apoptosis) pathway (7, 11). To confirm if CLU-dependent prosurvival autophagy was associated with the UPR, the activation of UPR signaling pathways was examined in hypoxic TEC-CLUhCLU cells compared with TEC-CLU−/− cells. By statistical comparison of the expression of 84 UPR-related genes in hypoxic TEC-CLUhCLU cells with TEC-CLU−/− cells using PCR array analysis, 57 genes were significantly upregulated, whereas 8 genes [glucosidase-α neutral C, UDP-glucose glycoprotein glucosyltransferase 2, ERO1-like-β, heat shock (70 kDa) protein 2, HtrA serine peptidase 4, heat shock protein family B (small) member 9, ER to nucleus signaling 2, and cAMP responsive element-binding protein 3-like 3] were downregulated, in hypoxic TEC-CLUhCLU cells (Table 1). The functional gene grouping showed that the mostly affected functional pathways of the UPR by CLU comprised “regulation of translation,” “heat shock protein,” and “ER protein folding quality control” followed by “protein folding,” “unfold protein binding,” “protein disulfide isomerization,” and “antiapoptosis” (Table 2). The CLU-dependent activation of UPR in hypoxic kidney cells was further confirmed by Western blot analysis of ER stress for IRE1, one of the principal UPR receptors. As shown in Fig. 7, the expression of IRE1 protein was upregulated by hypoxia in both TEC-CLUhCLU and TEC-CLU−/− cells compared with those under normoxia. However, a high level of p-IRE1 was seen only in hypoxic TEC-CLUhCLU cells (not TEC-CLU−/− cells; Fig. 7), suggesting that this UPR receptor was more active in hypoxic TEC-CLUhCLU cells than in CLU-null TEC-CLU−/− cells. Taken together, these data indicate that CLU-mediated autophagy is associated with more profound activation of the UPR, which may be related to the enhancement of protein translation and heat shock protein expression for protein folding and ER protein folding quality control, resulting in the prevention of cell apoptosis.
AKI due to IRI is a common pathological factor leading to unacceptably high morbidity and mortality and to the development of chronic kidney disease or the transition from preexisting chronic kidney disease to end-stage renal disease (85); however, the pathogenesis of renal IRI is not fully understood. Recently, many studies have demonstrated that autophagy is a cytoprotective mechanism in the kidneys against IRI (32, 35, 38), but the molecular mechanisms regulating this prosurvival autophagy remain largely unknown. The present study demonstrates that CLU expression is required for the activation of autophagy during IRI in the kidneys or by hypoxia in cultured kidney cells, and CLU-dependent autophagy is at least one of the pathways mediating the cytoprotection of CLU, as indicated by less IRI in WT kidneys or more surviving cells in cultures of CLU-expressing kidney cells under hypoxia compared with their CLU-null controls. The activation of CLU-dependent, prosurvival autophagy in hypoxic kidney cells also positively correlates with induction of the UPR, but how CLU regulates the UPR in kidney cells in response to hypoxia needs further investigation.
In the research field of autophagy, a consensus has not yet been reached of a single assay for the determination of autophagic activation (4, 55). It has been found that during autophagy, a cytosolic form of LC3 (LC3-1) is conjugated to phosphatidylethanolamine to form LC3-phosphatidylethanolamine conjugate (LC3-II), which then is recruited to be a part of the lipid membrane at the phagophore and autophagosome (52, 63, 89). Based on the conversion of LC3-I to LC3-II, the ratio of LC3-I to LC3-II has been used to determine the changes of autophagy (36). However, due to differential affinities or sensitivities of antibodies for LC3-I compared with LC3-II as well as different expression levels of LC3-I and LC3-II, depending on cell lines and tissues, this approach gave numerous false-positive or false-negative results (4). In fact, LC3-II (but not LC3-I) is the only specific functional marker for autophagic activation, as evidenced by that LC3-II at the autophagosome is required for both selecting cargo for degradation (e.g., an interaction of LC3-II with p62/sequestosome 1 targets-associated protein aggregates for turnover) and promoting membrane tethering and fusion (62), and, as of today, LC3-II is the only protein specifically localizing to autophagic structures throughout the process from phagophore to lysosomal degradation (63). Additionally, LC3-II itself is subjected to autophagic degradation at the lysosome. Thus, a consensus has emerged that the overall levels of LC3-II are normalized to a loading control, such as β-actin or α-tubulin, and the cellular level of LC3-II, rather than the LC3-I conversion, should be a cellular readout of autophagy level in all different experimental systems (4, 37). In the present study, levels of both LC3-1 and LC3-II are higher in WT kidneys than those in CLU-null controls after IRI (Fig. 1D) and perhaps in cultured WT kidney cells versus CLU-null controls after hypoxia (Fig. 2B). Furthermore, ectopic expression of human CLU in CLU-null mouse kidney cells also largely increases the expression of both LC3-I and LC3-II after exposure to hypoxia that can be suppressed by the presence of spautin-1 (Fig. 6). All these data indicate that CLU plays a role in LC3 expression, and using the ratio of LC3 II to LC3 I to compare autophagic activation between WT and CLU KO samples may not be correct. Therefore, we use the level of LC3-II to total cellular protein (β-actin or GAPDH) as a reliable marker of autophagic activation in this study. Indeed, cellular LC3-II levels were positively correlated with GFP-LC3 puncta both in vivo and in vitro in this study. Furthermore, similar to the decreased level of LC3 in CLU KO cells, a low level of LC3 expression (including LC3-I) by knockdown of LC3 gene expression inactivates autophagy in different studies in the literature (56, 72).
Many studies in the literature have confirmed that autophagy is activated in the kidneys after IRI or in hypoxic kidney cells (35, 88, 95, 97), and inhibition of autophagy by chloroquine and 3-methyladenine worsens renal IRI (34). The kinetics of autophagic activation in the kidneys with IRI has been well examined in a strain of transgenic mice expressing both markers of autolysosomes and early autophagic vacuoles, showing that both early autophagic vacuoles and autolysomes reach the maximal at day 1 after IRI surgery and that early autophagic vacuoles return to control levels at day 3, whereas autolysosomes remain unchanged (46). In addition, proliferative activity decreases in cells containing autolysosomes (46). This study suggests autophagy initiation at day 1 and autophagosome clearance during renal recovery at day 3 and that autophagic cells are less likely to involve in renal repair after IRI (46). Similar to these studies, the present study also shows an increase in autophagic activation, as indicated by high levels of both LC3-GFP puncta and cellular LC3-II in WT kidneys (Fig. 1). However, interestingly, compared with those of WT controls, there is only a minimum level of autophagic activation in CLU KO kidneys (Fig. 1). Furthermore, similar results have been demonstrated in our previous study (99), showing that rapamycin elevates LC3-GFP puncta in the hearts of WT LC3-GFP mice but not CLU KO LC3-GFP mice and that chloroquine treatment only attenuates LC3-GFP puncta in WT mice. Taken together, there is a positive correlation of IRI with autophagic activation in the kidneys, and a deficiency in CLU expression impairs this autophagic activation, which, however, may need to be further confirmed with chloroquine or 3-methyladenine treatment.
The present study confirms that hypoxia- or IRI-stimulated autophagy is associated with increased cell survival in the kidneys or cultured kidney cells, and incubation with autophagy inhibitors results in an increase in cell death, suggesting a prosurvival function of autophagy in kidney cells under hypoxia. However, the molecular pathways that activate prosurvival autophagy in the kidney in response to hypoxia or IRI are not fully understood. It is known that AMP-activated kinase (AMPK) is a sensor of cellular energy homeostasis (69), and its activity is required for autophagy activation in response to starvation (50) and hypoxia (16). Indeed, AMPK is strongly activated in the kidneys within 1 min of ischemia and remains so until 30 min (59), and agonist enhancement of AMPK activity improves autophagy and reduces renal IRI (15, 43), suggesting that AMPK may be a key regulator of autophagy in the kidneys with IRI. Furthermore, it is well documented that many biochemical events occur in renal tubular cells during the ischemia or hypoxia, which include a rapid degradation of ATP to ADP and AMP and an increase in cytosolic Ca2+ and ROS (17). Interestingly, the literature reports that hypoxia-activated AMPK in different experimental systems has been found due to elevated mitochondrial ROS and AMP through the activation loop of the upstream kinase liver kinase B1 (23, 28) and/or increasing cytosolic Ca2+ via Ca2+/calmodulin-dependent kinase kinase-β (28, 60). However, a recent study has shown that a significant reduction of AMPK activity in the kidneys of AMPK-β1 KO mice does not affect renal IRI at all compared with WT controls (58), indicating that the role of AMPK in the activation of prosurvival autophagy, in particular CLU dependent autophagy, in kidney cells under hypoxia or ischemia needs further investigation.
One of the interesting findings in this study is that autophagy was only detected in CLU-expressing kidney cells, even in those with ectopic expression of CLU, suggesting a role of CLU in the activation of autophagy in the kidneys with IRI or hypoxic kidney cells. If CLU takes part in hypoxia-induced autophagy, autophagy-related proteins together with CLU are likely upregulated by hypoxia in WT kidney cells. Indeed, hypoxia activates many transcription factors, including hypoxia-inducible factor (HIF)-1 (12). Activation of HIF-1 has been found in kidneys subjected to 45 min of renal ischemia and in cultured renal epithelial LLC-PK cells under hypoxia within 30 min and persisted for 4 h (21). HIF-1 regulates gene expression through its interaction with a hypoxia response element in the target gene promoter (84). A recent study has identified three putative hypoxia response element sites (A/GCGTG) in the CLU gene promoter and demonstrated the binding of HIF-1 to these sites (71), which may result in the upregulation of CLU expression in the kidneys in the response to IRI (65, 94, 101) and in HKC-8 cells (Fig. 2A) or others (73) under hypoxia. On the other hand, HIF-1 activity is also required for hypoxia-induced autophagy, in which HIF-1 induces the expression of Bcl-2/adenovirus E1B 19 kDa-interacting protein 3 (BNIP3) and BNIP3L, which play important roles in the induction of autophagy by disrupting the interaction of beclin 1 with Bcl-2 via their BH3 domain (6, 33, 100). The second substantial evidence to support the role of CLU in autophagy in hypoxic kidney cells comes from recent studies that have shown that CLU is colocalized with autophagy proteins (LC3-II and p62) in renal TECs from patients with nephropathic cystinosis (82) and with LC3 within the autophagosome in prostate cancer cells to enhance cell survival (99). However, how CLU regulates autophagy in hypoxic kidney cells is not fully understood as of yet.
Although the biology of autophagy has received increasing interest recently, the understanding of its pathways and functions still remain in its infancy. Beside nonselective macroautophagy, there are several selective autophagies, including chaperone-mediated autophagy, pexophagy, and mitophagy, that contribute to the degradation of intracellular components in lysosomes in mammalian cells (98); however, the regulatory pathways of each type of these autophagies remain poorly defined. It has been reported that the HIF-1-BNIP3 pathway involves autophagy in the kidneys with IRI or hypoxic kidney cells (33), and HIF-1-BNIP3-mediated autophagy selectively targets the mitochondria (mitophagy) (33). The prosurvival role of mitophagy is also found in human umbilical vein endothelial cells and cardiac myocytes under hypoxia (87). Hypoxia or reduced O2 availability primarily impacts mitochondrial oxidation and results in mitochondrial damage in renal TECs (24, 93), and the damaged mitochondria could release intermembrane proteins or other internal molecules that may trigger autophagy through an unknown mechanism (98). CLU is a chaperone protein containing several large molten globule domains, amphipathic regions, and coiled-coil α-helixes (3, 14), which allow CLU protein to bind various substrate proteins or lipids with high affinity and low specificity (3, 30, 70). Thus, it is possible that CLU may act as an adaptor protein that facilitates mitophagy by binding to mitochondrial intermembrane components, which, however, needs further investigation. Moreover, there is a low level of autophagic activity in CLU-null kidneys with IRI (Fig. 1) as well as in hypoxic CLU-deficient kidney cells (Fig. 2), implying that there is probably CLU-independent autophagy in kidney cells in response to hypoxia. Indeed, in addition to HIF-1-BNIP3 mitophagy, p53-sestrin-2 also induces autophagy in hypoxic kidney cells (33) and chaperone-mediated autophagy in hypoxic neuronal cells for cell survival (18). Whether or not these types of autophagy occur in the absence of CLU remains elusive.
In addition to autophagy, ER stress is also induced by IRI in the kidneys, as indicated by the accumulation of aberrant unfolded and misfolded proteins in the ER lumen, which subsequently activates the UPR pathway (31). The primary purpose of the UPR is to enhance protein-folding capacity by activating the transcription of UPR target genes, such as Ca2+-dependent molecular chaperones, including glucose-regulated protein (GRP)78, GRP94, and calreticulin (66), and to eliminate unfolded or misfolded proteins within the ER (11). In the absence of IRE1, double-stranded RNA-activated protein kinase-like ER kinase (PERK) and activating transcription factor (ATF)6 are normally held in an inactive state in the ER membrane by binding to GRP78, but under ER stress or as unfolded or misfolded proteins accumulate in the ER lumen, GRP78 releases these transmembrane ER proteins (IRE1, PERK, and ATG6), which consequently initiate UPR signaling (31). By activating the UPR, unfolded or misfolded proteins are eliminated through an ER-associated degradation pathway (ERAD) (26, 51), in which the soluble targeted proteins are retrotranslocated into the cytosol, ubiquitinated, and degraded by the proteasome in ERAD I (9, 64) or insoluble misfolded protein aggregates are degraded by autolysosomes/lysosomes by autophagy in ERAD II (26). The present study shows that the cytoprotection of CLU in hypoxic kidney cells is associated with an increase in the activity of IRE1 (Fig. 7) and the expression of a panel of UPR-targeted genes (Table 1), suggesting that CLU may play an important role in the elimination of unfold or misfolded proteins within the ER. However, the mechanisms by which CLU facilitates the UPR to restore the functions of the ER for cell survival are not well understood. CLU is associated with GRP78 in prostate cancer cells under ER stress (47) and is retrotranslocated into the cytosol by a mechanism similar to ERAD I (47, 68) and colocalized with the mitochondria (47). Furthermore, GRP78 stabilizes CLU protein and its hypoglycosylated form and is required for stress-induced upregulation of CLU (47) or vice versa (92), resulting in the inhibition of cell apoptosis. It seems that there is a requirement of CLU for the prosurvival UPR in hypoxic kidney cells via the enhancement of GRP67 expression, which may mediate prosurvival mitophagy via the ERAD II pathway.
In conclusion, the incidence of AKI due to transient renal hypoxia is rising but effective treatments and preventative approaches are currently lacking. Hypoxia- or IRI-induced kidney injury significantly contributes to delayed graft function of kidney transplants as well as to the development of kidney disease after ischemic AKI. Both clinical and experimental studies have demonstrated the cytoprotection of CLU in the kidney (65); however, the protective properties or actions of CLU in the kidney are not fully understood. This preliminary study, for the first time, demonstrates that the protection of CLU in hypoxic kidney cells is associated with prosurvival autophagy and the UPR, indicating that CLU-dependent autophagy and/or the UPR may be a potential target for the development of new therapeutic strategies for preventing and/or enhancing recovery after ischemic AKI in patients. However, before this, an increased mechanistic understanding of the protective nature of CLU is needed, in particular, further investigation of how these CLU-dependent pathways are coordinated in the response to hypoxia to increase cell survival.
H. A. Alnasser received the King Abdullah Scholarship from Saudi Arabia. The authors were supported by a grant from the Kidney Foundation of Canada.
No conflicts of interest, financial or otherwise, are declared by the author(s).
Author contributions: H.A.A., Q.G., and F.Z. performed experiments; H.A.A., Q.G., and C.D. analyzed data; H.A.A., F.Z., M.E.G., C.Y.N., and C.D. interpreted results of experiments; H.A.A., Q.G., and C.D. prepared figures; H.A.A. and C.D. drafted manuscript; H.A.A., F.Z., M.E.G., C.Y.N., and C.D. edited and revised manuscript; H.A.A., Q.G., F.Z., M.E.G., C.Y.N., and C.D. approved final version of manuscript; M.E.G., C.Y.N., and C.D. conception and design of research.
- Copyright © 2016 the American Physiological Society