p53 target Siva regulates apoptosis in ischemic kidneys

Abstract

The role of p53 in inducing apoptosis following acute kidney injury is well-established; however, the molecular mechanisms remain largely unknown. We report here that the p53 proapoptotic target Siva and its receptor CD27, a member of the tumor necrosis factor receptor family, are upregulated following renal ischemia-reperfusion injury (IRI). Inhibition of Siva using antisense oligonucleotides conferred functional and morphological protection, and it prevented apoptosis postrenal IRI in mice. Renal IRI in CD27-deficient mice displayed functional protection and partial inhibition of apoptosis, suggesting an incomplete role for CD27 in Siva-mediated apoptosis. To further elucidate mechanisms by which Siva elicits apoptosis, in vitro studies were performed. In Siva-transfected LLC-PK₁ cells, Siva is persistently expressed in the nucleus at 3 h onwards and its translocation to mitochondria and the plasma membrane occurred at 6 h. Moreover, Siva overexpression induced mitochondrial permeability, cytochrome c release, caspase-8 and -9 activation, translocation of apoptosis-inducing factor (AIF) to the nucleus, and apoptosis. Inhibition of Siva in ischemic kidneys prevented mitochondrial release of cytochrome c and AIF. These data indicate that Siva function is pivotal in regulating apoptosis in the pathology of renal IRI. Targeting Siva may offer a potential therapeutic strategy for renal IRI.

acute kidney injury (AKI) (1) is a devastating clinical syndrome with high mortality rate. Ischemia-reperfusion injury (IRI), generally accepted as the major cause of AKI, results from compromised perfusion of renal tissues and the ensuing hypoxia-induced necrotic and/or apoptotic cell death to renal cells. Thus, one potentially useful strategy for treating AKI is to target specific signaling events that lead to postischemic cell death. However, the exact mechanisms by which renal tubular cells undergo either necrotic or apoptotic form of cell death have not been well-defined. Apoptosis, an important consequence of renal IRI, contributes immediately to the cell death and subsequently occurs during tubular regeneration to prevent excessive cellular proliferation in the process of restoring tubular architecture (24, 33). Several factors have been identified in the mediation of apoptosis following renal IRI, and to understand the molecular mechanism of apoptosis, we previously performed a differential screening following IRI in rats (32). We found that the p53 proapoptotic target Siva and its associated receptor molecule CD27 are upregulated in kidneys following IRI (32). Nevertheless, the role and the mechanisms by which Siva activation induces injury and cell death post-IRI in the kidney have not been defined.

Siva is a proapoptotic protein that contains a death domain homology region, a box-B-like ring finger, and a zinc finger-like domain (35). Siva binds the cytoplasmic domain of CD27, a member of the tumor necrosis factor receptor (TNFR) superfamily (12, 35). In addition to its interaction with TNFR superfamily members, Siva has also been shown to interact with anti-apoptotic Bcl-X_L and sensitize MCF7 breast cancer cells to UV-induced apoptosis (48). Similarly, recent reports indicate that Siva can bind to lysophosphatidc acid recptor 2 (25, 40a), the adaptor protein TRAF2 in Jurkat T leukemia cells to modulate NF-κB activity (14), pyrin in possibly human neutrophils, monocytes, and synovial cells (3) and with slimmer in skeletal myoblasts to delay apoptosis (8). Of two Siva splice variants (Siva-1 and Siva-2), the predominant form, Siva-1, retains the exon2 and it is believed to be critical for apoptotic activity (49). However, a recent study suggests that both splice forms can induce apoptosis (36).

The signaling pathways triggered by Siva-1 in T-cell lines include activation of caspases and initiation of mitochondrial events. It has been suggested that Siva-1 mediates activation of caspase-8 by associating with its cognate receptor, CD27 (31, 35). Caspase-8 then activates the effector caspases either directly by cleaving them or indirectly through a Bid-mediated mitochondrial pathway (17, 29). In type II Jurkat cells, Siva-induced apoptosis resulted in proteolytic activation of Bid and release of cytochrome c (7, 48). However, Siva-1 can directly activate caspase-3 in type I SKW6.4 lymphoid cells without the mitochondrial involvement (36). Interestingly, in both of these cell lines, inhibition of caspase led to only a moderate inhibition of apoptosis, suggesting that Siva can also induce apoptosis by caspase-independent mechanisms (36). Thus, these discrepancies in the choice of downstream pathways underscore the possibility of cell- or tissue-specific execution of apoptosis by Siva-1.

We previously reported that the expression of Siva (Siva-1, unless otherwise specified) was upregulated in the damaged epithelium of the S₃ segment of the proximal tubule at 12 and 24 h following IRI, and also in cells of papillary proliferation during regenerative phase (32). In addition, CD27, the plasma membrane receptor of Siva, was also correspondingly localized in injured epithelium of S₃ segment and in cells of papillary proliferation following IRI, suggesting an interaction of Siva with CD27 in the mediation of apoptosis (32). Nonetheless, the mechanism of Siva-mediated apoptosis in the injured renal tubular epithelium and its functional significance following IRI have not been studied. We reasoned that targeting Siva function that mediates the p53 apoptotic response in the setting of renal ischemia might provide a useful strategy for preventing apoptosis in IRI. In this study, we report that Siva antisense administration protects from experimental renal IRI, and CD27 gene ablation partially protects mice from IRI. Furthermore, by employing both in vitro and in vivo experimental models, our data indicate that Siva mediates apoptosis via extrinsic and intrinsic pathways by activating its plasma membrane receptor CD27 and mitochondrial release of cytochrome c and apoptosis-inducing factor (AIF).

MATERIALS AND METHODS

Animal and surgical procedures.

CD27-knockout (KO) mice (B6/129P2) and wild-type (WT; B6/129SF2/J; Jackson Laboratories) mice were cared before and during the experimental procedures in accordance with the policies of the Institutional Animal Care and Use Committee (IACUC), University of Nebraska Medical Center (UNMC) and the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Littermate controls were used for all antisense administration studies. All protocols had received prior approval from the UNMC-IACUC. WT male mice (∼20 g; Jackson Laboratories) were anesthetized by intraperitoneal administration of a cocktail containing ketamine (200 mg) and xylazine (16 mg) per kilogram of body weight. Ischemic injury was induced by bilateral renal pedicle clamping using microaneurysm clamps as we previously described (52). After 37 min of occlusion, the clamps were removed, and reflow was verified visually. Sham-operated control animals underwent the same surgical procedure, except for the occlusion of the renal arteries. At different time points (12, 24, and 48 h), postinjury mice were anesthetized using pentobarbital sodium for blood collection. At the time of death, kidneys of the animals were perfused with sterile phosphate-buffered saline to remove the blood from organs. The kidneys were quickly excised and half of the kidneys were snap-frozen in liquid nitrogen for RNA and protein isolation and the other half was fixed in Bouins or formalin solution for morphological studies and immunohistochemistry.

Oligonucleotide screening and administration.

Oligonucleotides corresponding to regions within the coding regions of Siva-1 were synthesized and purified in the UNMC core facility. Oligonucleotides were phosphorothioate-modified chimeric oligonucleotides composed of five 2′-O-(2-methoxy) ethyl modifications on both the 5′ and 3′ ends and at least 10 oligodeoxynucleotides in the center to support RNase H activity (23). Oligonucleotide sequences were as follows: ACCGCCTTCCCATCCACAGAT (AS-Siva01), GCTCCTACCACACCAGACTTC (AS-Siva02), and AGAACCAAGCAGCTCCTTT (AS-Siva03). The sequence of the scrambled (scr) oligonucleotide was CTACTCAAACCCCATCGGCCT. The effective oligonucleotides were selected based on in vitro screening after transfection into mouse cortical tubular epithelial cell line with 300 nM antisense oligonucleotide. Twenty-four hours after transfection, cells were harvested and total RNA was isolated. Quantitative real-time PCR was performed for determining the efficacy of oligonucleotides, and the selected Siva antisense oligonucleotides (AS-Siva01, AS-Siva02, and AS-Siva03) were administered intraperitoneally at a dose of 10 mg·kg body wt⁻¹·day⁻¹ for 2 days before IRI in two divided doses (4 injections in total) and renal functional assessments were measured 24 h following IRI.

Morphological studies.

WT mice treated with AS-Siva or saline were subjected to IRI. Kidneys were harvested and fixed in either Bouin's or formalin fixative, paraffin-embedded, and 5-μm sections were cut. The tissue sections were stained with hematoxylin and eosin (H&E) and periodic acid-Schiff.

Histological changes were evaluated by quantitative measurements of epithelial necrosis, tubular dilation, and tubular casts in 5–10 high-power fields per section and these data were quantified as previously (6). Tissue scoring was performed in a blinded manner in adjacent fields in the corticomedullary junction by grading the above pathologic changes according to the following scoring system: grade 0 - no pathologic changes; grade 1 - pathologic changes in 1–25% of the area; grade 2 - pathologic changes in 25–50% of the area; grade 3 - pathologic changes in 50–75% of the area; and grade 4 - pathologic changes in 75–100% of the area.

Renal functions.

The blood collection from mice was performed under ketamine and xylazine anesthetics. The retro-orbital sinus was selected as the source of venous blood. Using a heparinized glass capillary, (Drummond Microcaps) medial canthus of the eye was approached and with gentle pressure retro-orbital sinus was punctured and 75–85 μl of the blood were collected using a glass capillary. Postcollection hemorrhage, if any, was controlled by application of optimal direct pressure over eyelids with a clean tissue paper, and appropriate postanesthetic care was provided to mice. Collected blood samples were centrifuged in heparinized microcentrifuge tubes preloaded with gel separator (Capiject, Terumo). Plasma urea concentration was measured by colorimetric assay using QuantiChrom^M Urea Assay Kit (BioAssay Systems) according to the manufacturer's instructions. The assay uses substrates that detect urea directly, and the reactions were carried out in 96-well plates. The readings were read in a plate reader (Infinite 200 plate reader, Tecan), and results were statistically analyzed.

Immunohistochemistry.

Localization of Siva was performed by immunofluorescent staining on kidney tissue sections as previously described (32).

Antibodies.

The following primary antibodies were used in the study: 1) Siva, rabbit polyclonal antibody from Santa Cruz Biotechnology, 2) AIF, rabbit polyclonal antibody from Cell Signaling Technology, 3) cytochrome c, mouse monoclonal antibody from Cell Signaling Technology, and 4) GAPDH antibody from Novus. Appropriate fluorescent-conjugated species-specific secondary antibodies (Alexa Fluor Dyes) were from Invitrogen.

Real-time PCR analysis.

Total RNA was extracted from renal tissues or tubular epithelial cells with TRIzol reagent (Invitrogen) and RNA quality was assessed by gel electrophoresis and Bioanalyzer analysis. First-strand cDNA synthesis was performed using the RT² First Strand Kit (SuperArray). qPCR was performed in ABI 7500 Real-Time PCR System (Applied Biosystems) and analyzed using SDS version 2.3 (Applied Biosystems). Twenty microliter reactions were performed using RT² Real-Time SYBR Green/ROX PCR Master Mix (SuperArray Bioscience). The mouse β-actin was used as an internal control. Superarray Primer IDs were as follows: Siva, PPM03413A and β-actin, PPM02945A. All primers used spanned an intron/exon border.

Apoptotic assay in renal tissues.

Apoptotic cells in renal tissues were labeled using the Fluorescein-FragEL DNA Fragmentation Detection Kit (EMD Biosciences) as per the manufacturer's instructions. Briefly, renal tissues were fixed with 10% formaldehyde and 5-μM-thick sections were made. Tissue sections were deparaffined with xylene followed by rehydration. Permeablization was performed with incubation of proteinase K for 20 min at room temperature. Following brief equilibration, tissue sections were incubated with terminal deoxynucleotidyl transferase (TdT) enzyme reaction mixture at 37°C for 1.5 h. TdT enzyme reaction was terminated and tissue sections were counterstained with DAPI. Tissue sections were then mounted for evaluation under Leica fluorescent microscope with appropriate filters

Cell culture and hypoxic injury.

Cell culture and hypoxic injury were performed as previously described (41). Briefly, hypoxic culture conditions (oxygen level <2%) were achieved with a pouch system providing a CO₂-enriched anaerobic environment (BBL GasPak Pouch System, Becton Dickinson) as previously reported (50). Hypoxic injury was induced by incubating 80–90% confluent monolayers of LLC-PK₁ for 4 h in Krebs-Ringer bicarbonate buffer (in mM: 115 NaCl, 1 KH₂PO₄, 4 KCl, 1 MgSO₄, 1.25 CaCl₂, and 27 NaHCO₃) at 37°C. After hypoxia, cells were allowed to recover for different time intervals by reoxygenation in normal growth media.

Immunofluorescence.

Cells cultured on glass coverslips were fixed with methanol-acetone (1:1) for 10 min at −20°C or with 4% formaldehyde for 15 min at room temperature. Permeablization and blocking were carried out with TBST (20 mM Tris·HCl, pH 8.0, 150 mM NaCl, 0.05% Tween 20) containing 10% calf serum for 10 min at room temperature followed by incubation with primary antibodies for 1 h at 37°C. The following primary antibodies were used in the study: 1) Siva, rabbit polyclonal antibody from Santa Cruz Biotechnology, 2) AIF, rabbit polyclonal antibody from Cell Signaling Technology and Santa Cruz Biotechnology, 3) cytochrome c, mouse monoclonal antibody from Cell Signaling Technology, and 4) GAPDH antibody from Novus. Subsequently, cells were incubated with fluorescent-conjugated secondary antibodies for 1 h at room temperature and mounted for detection under fluorescent microscopy. Appropriate fluorescent-conjugated species-specific secondary antibodies (Alexa Fluor Dyes, Molecular Probes) were from Invitrogen. Labeled sections were examined under the Leica DMR fluorescent microscope or Zeiss LSM 510 Meta confocal laser-scanning confocal microscope (UNMC Core facility). The digital images were processed using Adobe Photoshop software for final layout.

Apoptotic assay in renal epithelial cells.

Quantification of cells undergoing apoptotic cell death by flow cytometry analysis was performed as we previously described using the Telford method (41).

Quantification of AIF translocation.

Kidneys that underwent IRI or sham operation or LLC-PK₁ cells were immunostained for detecting the expression of AIF. The expression of AIF demonstrating a typical punctate expression pattern was considered to be localizing to the mitochondria while its expression in the nucleus colocalizing with Hoechst stain was considered to be in the nucleus. The number of cells with punctate appearance of AIF vs. nuclear localization was counted in 10 high-magnification fields (×400).

Caspase activity.

Caspase activation was assayed by measuring enzyme activity using luminogenic caspase-specific substrates from Promega as per the manufacturer's instructions (Caspase-Glo Assay, Promega). Briefly, cells were cultured to semiconfluency and transiently transfected with GFP-Siva or GFP-only control vectors. Following 1 h of hypoxic injury, cell suspensions were made in normal medium and mixed with either Caspase-Glo 8 Reagent or Caspase-Glo 9 Reagent. Following 30 min of incubation at room temperature, peak luminescent signals were read in a luminometer. Enzyme activity was expressed in relative light units/seconds. Results were statistically analyzed using Student's t-test for two groups and ANOVA with a post hoc test across all groups.

Mitochondrial membrane permeability assay was performed using the JC-1 dye method (Molecular Probes) as we previously described (41).

Statistics.

The Student's t-test was used to compare the means of two experimental groups. One-way ANOVA with post hoc test was used to compare the means of more than two groups. A P value of <0.05 was considered statistically significant.

RESULTS

Siva protein expression is induced in a p53-dependent manner postrenal ischemia.

We previously showed the upregulation of Siva mRNA at 12 and 24 h following renal IRI in proximal tubular cells (PTC) that were undergoing apoptosis (32). Here, we demonstrate that Siva protein expression was highly induced in WT mice 1 day postinjury in the damaged proximal tubule epithelial cells by immunofluorescent microscopy (Fig. 1B) and its subcellular localization by confocal imaging (Fig. 1C), and complete downregulation of its expression in the p53-KO mice (Fig. 1, E and F). Western blot analysis (Fig. 1G) followed by quantification of the data (Fig. 1I) demonstrates that the expression level of Siva protein in ischemic mouse kidneys was induced at 12 h, 1 and 2 days postinjury, time points at which maximal apoptosis was previously detected in ischemic kidneys (33) and the expression level was normalized by 3 days postinjury. No expression of Siva was observed in ischemic p53-KO or in sham (SH)-operated mice kidneys (Fig. 1, A, D-F). The dependence of Siva expression on p53 activation was further confirmed by Western blot analysis. The data presented in Fig. 1H demonstrate that at 1 day postinjury, Siva expression was induced in WT but was completely absent in p53-KO mouse kidneys. Collectively, the temporal and spatial expression pattern of Siva mRNA and protein in ischemic kidneys at the sites of apoptosis in a p53-dependent manner indicate that Siva is a downstream p53-apoptotic effector in the setting of renal ischemia.

Fig. 1.

Siva is induced in apoptotic cells in a p53-dependent manner. Expression of Siva protein in sham-operated (SH; A) or ischemia-induced wild-type (WT) kidneys shows that Siva is induced at 1 day postreperfusion in some of the injured proximal tubule cells (PTC; B). Confocal imaging demonstrates Siva expression in the cytoplasmic region in most cells, while it is localized to the nuclei in a few cells (C; arrowheads). Expression of Siva was absent in p53-knockout (KO) kidneys at 1 day postreperfusion (E: ×200 and F: ×400) and in SH kidneys derived from both WT and p53-KO animals (A and D, respectively). G: Western blot analysis of expression of Siva in ischemic kidneys at various time points postischemic injury compared with that in SH animal kidneys. H: Western blot analysis of expression of Siva in ischemic kidneys as a function of p53. I: GAPDH was used as a loading control. G and H: quantification of the temporal expression of Siva showed that its expression levels at 12, 24, and 48 h postinjury are significantly increased (*P < 0.05; n = 3) compared with that in SH animals. IRI, ischemia-reperfusion injury.

Siva inhibition offers functional protection from IRI.

To determine whether downregulation of Siva offers functional protection from IRI, we administered a cocktail of three Siva-antisense (AS-Siva) oligonucleotides at a dosage of 10 mg·kg⁻¹·day⁻¹ in adult mice during 48 h before IRI. Administration of AS-Siva oligonucleotides effectively inhibited the expression of Siva mRNA by 87.96 ± 3.1% (n = 4; P < 0.05) and protein levels by 82.3 ± 4.5% (n = 4; P < 0.05) in the ischemic kidneys compared with vehicle administration at 24 h postinjury (see supplementary data; Fig. S1A and B; the online version of this article contains supplemental data). At 1 day post-IRI, blood urea nitrogen (BUN) values were determined. In addition, the BUN levels were significantly lower in AS-Siva-pretreated mice compared with that in vehicle-treated and scr oligonucleotide-treated mice following IRI (n = 6; P < 0.01; Fig. 2A). No significant difference existed in BUN levels between vehicle-treated and scr-treated sham-operated mice (data not shown) or that in vehicle-treated and scr-treated mice post-IRI. These results indicate that downregulation of Siva offered functional protection against IRI in mice.

Fig. 2.

A: pretreatment with Siva antisense oligonucleotides protects from IRI. Blood urea nitrogen (BUN) levels were assessed at 24 h post-IRI in vehicle-treated and AS-Siva-pretreated mice and in SH mice. *P < 0.01 between SH group and IRI group. #P < 0.01 between IRI group and AS-Siva group (n = 4). B: effect of Siva inhibition on histological parameters at 1 day postinjury. Top: hematoxylin and eosin (H&E)-stained representative images derived from the outer medullary region of SH, vehicle-treated IRI-induced, and AS-Siva-treated ischemic kidneys. Bottom: morphological scoring (performed as described in materials and methods) demonstrates that AS-Siva-treated ischemic kidneys are significantly (*P < 0.05; n = 4) protected compared with vehicle-treated IRI-induced kidneys. C: Siva inhibition prevented apoptosis in ischemic kidneys. Left: renal tissues derived from WT vehicle-treated group and AS-Siva-treated group at 24 h post-IRI and from SH animals were assayed using the TUNEL method. Right: number of cells positive for TUNEL staining was counted in 5–10 high-magnification fields under a fluorescent microscope. At 24 h, the number of apoptotic cells was significantly lower in AS-Siva-treated group (*P < 0.01; n = 5; 1-way ANOVA) compared with IRI vehicle-treated or scrambled (scr)-treated groups.

Morphological protection following Siva inhibition in ischemic mice kidneys.

To examine whether Siva inhibition leads to histopathological improvement following IRI, we performed histological scoring in kidneys of AS-Siva-treated and vehicle-treated mice 1 day post-IRI. Low-magnification photographs (×100) of H&E-stained kidney sections originating from sham-operated, AS-Siva-treated, and vehicle-treated mice at 1 day post-IRI are shown in Fig. 2B, top. Ischemic kidneys from vehicle-treated mice showed widespread necrosis, tubular casts, and sloughed cells in the outermedullary proximal straight tubule segments, whereas these features were dramatically reduced in AS-Siva-treated mice.

Histological scoring was conducted on kidneys derived from the two IRI groups and from sham-operated mice. Semiquantitative data demonstrate that tubular damage was significantly (P < 0.05; n = 4) reduced in kidneys of AS-Siva-treated mice compared with that in vehicle-treated mice (Fig. 2B, graph), demonstrating that Siva inhibition attenuated tissue injury. The histological damage was negligible in sham-operated animals.

Siva inhibition attenuated apoptotic cell death in ischemic kidneys.

TUNEL assay was performed to determine the effects of inhibition of Siva expression on the number of cells undergoing apoptosis postinjury in kidneys obtained from AS-Siva-treated and vehicle-treated animals. AS-Siva treatment prevented apoptosis compared with that in vehicle-treated post-IRI kidneys. Quantification of the number of TUNEL-positive apoptotic cells per high-magnification field demonstrates that Siva inhibition completely prevented (n = 5; P < 0.001) apoptosis post-IRI (Fig. 2, Cb). The number of apoptotic cells in sham-operated and AS-Siva-treated tissues did not significantly differ (n = 4; P = 0.47). Collectively, these results demonstrate that expression of Siva is physiologically important and that Siva protein functions are a key mediator of p53-mediated apoptotic cell death in renal ischemia.

Functional protection attenuated caspase-8 activation and apoptosis in CD27-deficient mice following renal IRI.

The mechanism by which Siva induces apoptosis in ischemic kidneys is not defined. Siva has been shown to bind to CD27 at the cytoplasmic tail and initiates extrinsic apoptotic cascades (35). We previously reported the expression of CD27 on the plasma membrane of rat renal PTC in rat kidney following renal ischemia (32). A similar temporal and spatial expression pattern for CD27 is observed in ischemic mouse kidneys (data not shown). To determine the functional relevance of CD27 expression in ischemic renal injury, renal functions were assessed by measuring serum BUN levels (Fig. 3A) at 1 day postinjury in WT and CD27-KO mice (kindly provided by Prof. Jannie Borst, Netherlands Cancer Institute). These data demonstrate that CD27 gene ablation significantly protected mice from ischemic renal injury; however, the percentage of change in BUN levels from the WT-injured animals was less than that observed in AS-Siva-treated animals, suggesting an incomplete role for CD27. Quantification of apoptotic cell death also showed that CD27 gene ablation resulted in inhibition of apoptosis by ∼56% (Fig. 3B), a value far less than that achieved by Siva inhibition. The number of apoptotic cells in sham-operated WT and CD27 kidneys did not significantly differ (n = 4; P = 0.38). The temporal expression of CD27 and the functional relevance of its expression demonstrate that CD27 is a downstream effector of p53-Siva-directed apoptotic pathways. However, this result also suggests that CD27 may only be a part of the p53-Siva-triggered theme of apoptotic execution and parallel pathways may be involved in the p53-Siva-mediated apoptosis.

Fig. 3.

Effect of CD27 gene ablation on BUN and apoptosis postrenal ischemia. A: ischemic injury was induced in CD27-deficient and WT mice. The levels of BUN at 1 day postinjury were significantly reduced (#P < 0.05, Student's t-test; n = 4) at 1-day time point in CD27-deficient mice compared with the WT mice. B: to determine whether there is a variation in the number of cells undergoing apoptosis in WT mice compared with CD27−/− mice, TUNEL assay was performed in conjunction with nuclear morphology analysis. Renal tissues derived from both WT and CD27−/− mice at 24 h postischemic injury were assayed and the number of cells positive for TUNEL staining was counted in 10–20 high-magnification fields under a fluorescent microscope. Data from WT and CD27−/− mice were used to quantify the extent of apoptosis. At 24 h, the number of apoptotic cells was significantly (*P < 0.05; n = 4) lower in CD27-deficient kidneys compared with that in WT. The number of apoptotic cells in SH WT (0.58 ± 0.13) and CD27-KO (0.53 ± 0.1) kidneys did not significantly differ (n = 4; P = 0.38).

Overexpression of Siva in renal epithelial cells induces apoptosis.

To determine whether Siva induces apoptosis in renal tubular epithelial cells and to further elucidate the mechanisms by which Siva induces apoptosis, green fluorescent protein (GFP) or GFP-Siva vector was transiently transfected and examined for apoptotic morphology. Twelve hours following transfection, GFP-Siva-transfected cells showed distinct spatiotemporal expression of Siva and underwent apoptosis as evidenced by cellular condensation and nuclear fragmentation (Fig. 4, B and C). Untransfected cells (not shown) or GFP-transfected cells (Fig. 4A) did not show any distinct subcellular expression of GFP and displayed less number of apoptotic cells (Fig. 4A). The percentage of cells undergoing apoptosis was estimated by the Telford method, followed by flow cytometry. Compared with GFP-expressing cells (13.29 ± 2.5%; n = 4; Fig. 4E) and untransfected cells (3.02 ± 0.49%; n = 4; Fig. 4D), GFP-Siva-expressing cells underwent significantly greater percentage of apoptosis (22.33 ± 1.17%; n = 4; P < 0.0001; Fig. 4, F and H) at 12 h and (41.45 ± 3.08%; n = 4; P < 0.0001; Fig. 4, G and H) at 24 h, indicating that overexpression of Siva per se is sufficient to induce apoptosis in renal cells.

Fig. 4.

Overexpression of Siva induces apoptosis in LLC-PK₁ cells. Confocal microscopy images (×400 magnification) showing the expression of green fluorescent protein (GFP; A), GFP-Siva (B and C). C: cellular condensation and nuclear fragmentation in GFP-Siva-transfected cells at ×630 magnification. D-G: flow cytometry for apoptosis: untransfected cells (D), control GFP cells (E), GFP-Siva-expressing cells at 12 h (F), GFP-Siva-expressing cells at 12 and 24 h (G), and statistical analysis (H).

Siva localizes to the nucleus, mitochondria, and plasma membrane.

To elucidate the events involved in the Siva-mediated apoptosis, GFP or GFP-Siva vector was transiently transfected and the spatiotemporal localization of Siva in LLC-PK₁ cells was examined. After a brief presence in the cytoplasm, GFP-Siva consistently localized to the nucleus as early as 3 h following transfection (Fig. 5A). At 6 h, GFP-Siva typically accumulated in a beaded fashion along the periphery of the nucleus (Fig. 5B). Subsequently, Siva localized in the plasma membrane and mitochondria by 12 h following transfection (Fig. 5, C, F, and H). To identify the localization of Siva in mitochondria, we labeled mitochondria with MitoTracker Deep Red 633 (Invitrogen; Fig. 5G), which upon excitation emits in far-red range of the light spectrum and provides better signal separation. Following overexpression, Siva colocalized with MitoTracker at 12 h in LLC-PK₁ cells indicating the translocation of Siva to mitochondria (Fig. 5H). At the same time point, localization of Siva was also detected on the plasma membrane.

Fig. 5.

Spatiotemporal expression of Siva following overexpression in LLC-PK₁ cells. A: nuclear localization of GFP-Siva 3 h after overexpression. B: centrifugal aggregation of GFP-Siva along the periphery of the nucleus at 6 h after transfection (×1,000 magnification). C: expression of GFP-Siva in the cytoplasm and on the plasma membrane at 12 h after transfection. E-H: colocalization of GFP-Siva with mitochondria-specific marker Mito Tracker dye and plasma membrane localization of GFP-Siva. No distinct spatiotemporal expression of GFP was observed (D).

Siva induces mitochondrial membrane permeability in renal epithelial cells.

Mitochondrial outer membrane permeabilization plays a decisive role in triggering different cell death cascades by mediating the release of proapoptotic molecules, including cytochrome c and AIF (22). Nevertheless, the mitochondrial release of cytochrome c can occur independent of mitochondrial membrane permeability (45). Thus, to determine whether Siva can induce mitochondrial membrane polarization, we assessed the mitochondrial membrane potential by flow cytometry using JC1 assay. JC-1 is a cationic dye, which accumulates in mitochondria in a membrane potential-dependent manner and forms red fluorescent J-aggregates, which emit fluorescence at ∼590 nm [Fig. 6, A–C, second quadrant (Q2)]. Loss of mitochondrial membrane polarization leads to a decrease in red fluorescent J-aggregates and an increase in green fluorescent monomers, which emit fluorescence at ∼529 nm [Fig. 6, A-C, fourth quadrant (Q4)]. The fourth quadrant represents cells experiencing loss of mitochondrial membrane depolarization. Compared with non-GFP control vector-transfected cells (18.04 ± 2.4%; n = 4; Fig. 6, A and D), Siva-overexpressing cells showed significantly higher percentage of mitochondrial depolarization (37.02 ± 4.3%; n = 4; *P < 0.01; Fig. 6, B and D), indicating that Siva mediates mitochondrial membrane depolarization following its translocation to mitochondria. Carbonyl cyanide m-chlorophenyl-hydrazone, a potent mitochondrial membrane potential disrupter, was used as a positive control and it induced mitochondrial membrane permeability (MMP) in 84.69% of vehicle-treated control cells (Fig. 6C).

Fig. 6.

Overexpression of Siva mediates mitochondrial membrane depolarization. Results of JC-1 assay used to determine mitochondrial membrane depolarization 8 h following Siva transfection. A: control vector-transfected cells. B: Siva-overexpressing cells. C: cells treated with carbonyl cyanide m-chlorophenyl-hydrazone (CCCP) as a positive control. D: graphical representation of statistical analysis. *P < 0.01.

Siva induced cytochrome c release and mitochondrial-nuclear translocation of AIF in vitro.

AIF, a mitochondrial intermembrane protein, has been shown to mediate caspase-independent apoptosis following mitochondrial membrane permeabilization. Although the mitochondrial presence of Siva and activation of caspases have been previously demonstrated, the role of AIF in the Siva-mediated apoptosis has not been previously studied. At 12 h following transient transfection of a control vector, immunofluorescence double labeling with anti-AIF and anti-cytochrome c antibodies followed by confocal imaging revealed a punctate pattern of colocalization in the cytoplasm indicating the mitochondrial localization (Fig. 7A, b-d). However, a diffuse nuclear localization of AIF was detected at 12 h following transient transfection of pCMV-FLAG-Siva in LLC-PK₁ cells, and following translocation to the nucleus, AIF initiated nuclear condensation (Fig. 7A, g, h, k, and l), suggesting the involvement caspase-independent apoptotic mechanisms. Moreover, number of cells with nuclear translocation of AIF in Siva-overexpressing cells (10.41 ± 0.6%; n = 3; Fig. 7Ab) was significantly more than in control vector-transfected cells (2.41 ± 1.43%; n = 3; P < 0.01). In addition, cytoplasmic release of cytochrome c from mitochondria was also observed in cells showing nuclear localization of AIF (Fig. 7Al), suggesting that both caspase-dependent and caspase-independent apoptotic cascades are simultaneously activated in Siva-mediated apoptosis.

Fig. 7.

A: mitochondrial-nuclear translocation of apoptosis-inducing factor (AIF) following Siva overexpression. a-d: Representative images of control empty vector-transfected cells. e-l: Representative images of pCMV-FLAG-Siva-transfected cells. Nuclear translocation of AIF is indicated by white arrows (g). B: compared with control vector-transfected cells, more number of pCMV-FLAG-Siva-transfected cells showed nuclear translocation of AIF (2.41 ± 1.43 vs. 10.41 ± 0.6%; n = 3; *P < 0.01). Cells with nuclear translocation of AIF were counted in 5 different microscopic fields per sample at ×400 magnification. Siva inhibition prevents mitochondrial release of cytochrome c and nuclear translocation of AIF in 1-day postischemic kidneys. Mice treated with AS-Siva or untreated were subjected to IRI and the kidneys were harvested at 1 day post-IRI. The expression of cytochrome c and AIF was immunodetected and analyzed by confocal microscopy (×400) in the kidney sections. Representative images from the outer medullary region of the kidneys are shown. In the majority of cells in samples from untreated mice, cytochrome c translocated from the mitochondria to the cytoplasmic region as demonstrated by the diffused expression pattern compared with the punctate pattern observed in samples from AS-Siva-treated mice. Nuclear translocation of AIF occurred in some cells of samples from untreated mice while it was minimal in AS-Siva-treated kidneys.

Inhibition of Siva attenuated cytochrome c and AIF release in ischemic kidneys.

To determine whether the prevention of apoptosis following Siva inhibition is mediated by mitochondrial events involving cytochrome c and AIF post-IRI, kidneys derived from AS-Siva-treated and vehicle-treated ischemic mice were immunostained for cytochrome c and AIF with their respective antibodies. Inhibition of Siva using antisense oligonucleotides post-IRI attenuated cytochrome c release from mitochondria as demonstrated by preservation of positive punctate pattern typical of mitochondrial localization in 1 day postischemic kidneys (Fig. 7BA). Conversely, a diffuse pattern suggesting mitochondrial release of cytochrome c in kidneys of vehicle-treated animals was observed (Fig. 7BB). Moreover, immunostaining of the tissues using AIF antibody demonstrated nuclear localization of AIF in a significantly large number of cells in kidney sections of vehicle-treated mice (Fig. 7BD) compared with that of AS-Siva-treated mice (Fig. 7BC). The number of cells with nuclear localization of AIF in ×40 high-power fields was counted and is significantly decreased (P < 0.01; n = 3) in AS-Siva-treated cells (2.41 ± 1.4%) compared with vehicle-treated mice kidneys (10.4 ± 0.6%).

Siva activates caspases-8 and -9 in renal epithelial cells.

Siva has been shown to interact with CD27, a cell surface receptor member of TNFR family, to activate caspases in T lymphocytes (36). Moreover, we previously showed that CD27 is coexpressed in injured renal proximal tubular epithelial cells (32). In this study, we demonstrated that CD27 gene ablation blocked caspase-8 activation and partially prevented apoptosis in ischemic kidneys. In addition, 12 h following transfection in LLC-PK₁ cells, GFP-Siva localized on the plasma membrane (Fig. 5), suggesting its interactions with CD27. GFP-Siva also localized in mitochondria and mediated the release of AIF and cytochrome c at 12 h (Fig. 7A) and inhibition of Siva prevented mitochondrial release of AIF and cytochrome c in ischemic kidneys (Fig. 7B). Thus, following GFP-Siva or GFP overexpression, we examined the activation of initiator caspases, caspase-8 and -9, using a cell-based sensitive assay, which uses luminogenic caspase substrates. Caspase-8 activity was significantly increased by 2.27-fold in Siva-overexpressing cells compared with GFP-overexpressing cells (125,900 ± 4,921 vs. 55,430 ± 6,128; n = 6; P < 0.0001). Similarly, caspase-9 activation was significantly increased by 1.94-fold in Siva-overexpressing cells compared with GFP-overexpressing cells (907,800 ± 63,460 vs. 467,300 ± 40,980; n = 6; P < 0.001). The data are expressed in relative light units. These results indicate that Siva mediates apoptosis through both extrinsic and intrinsic apoptosis pathways.

DISCUSSION

IRI results in the induction of apoptotic cell death in the proximal tubular epithelium (39, 40, 44). Although several factors including lethal cytokines and receptors (FasL, Fas) (15, 28), Toll like receptor-4 (47, 51), apoptotic regulators (Bcl2, Bcl-xL, Bax) (5, 37), and caspase activation (18) have been implicated in apoptotic cell death in AKI, a correlation between the effect of many of these gene deficiencies and renal function in the setting of ischemic AKI has not been well-established. Recent studies demonstrate that activation of the transcription factor p53 can promote apoptosis in renal ischemia (9, 19).

The subset of p53 target genes that are up- or downregulated in response to renal ischemic injury has not been defined. To identify the regulatory targets by which p53 induces renal cell death, we employed a differential screening strategy and identified two of the p53 apoptotic effector genes, Siva (32) and PERP (2) are enhanced in a p53-dependent manner. The data presented in this study demonstrate that Siva is a pivotal player in the regulation of apoptosis in renal IRI.

We previously reported that the expression of the proapoptotic molecule Siva, and its cognate receptor CD27, is highly upregulated and coexpressed in the ischemic rat renal tissues (32), indicating that Siva and CD27 play a critical role in apoptosis following IRI. Nonetheless, the functional significance and the subcellular mechanism of Siva and CD27 after IRI in the kidney are unknown. To understand the physiological role of Siva and CD27 in AKI, we used two models that target Siva or CD27 in ischemic renal injury. Downregulation of Siva by administration of Siva antisense cocktail (AS-Siva) during the 48 h before IRI has significantly protected renal functions possibly by inhibiting the incidence of apoptotic cell death, minimizing sublethal injury, and minimizing the tubular lumen obstruction caused by cells undergoing cell death (44). Downregulation of Siva has been shown to inhibit the T-cell receptor activation-induced cell death in T-cells (13), promote cell survival following lysophosphatidic acid 2 receptor activation in NIH 3T3 cells and in cisplatin-injured carcinoma cells, underscoring the importance of Siva in the mediation of apoptosis (4, 25). However, to our knowledge, a role for Siva to be directly involved in a pathological condition such as IRI has not been reported previously in any organ. Our results from AS-Siva pretreatment clearly indicate that inhibition of Siva offered functional protection and prevented apoptosis in renal IRI.

Our data demonstrate that the expression of Siva is p53-dependent in injured kidney concurring with previous results observed in in vitro cultures of irradiated mouse embryonic fibroblasts, campothecin-injured cerebellar granule neurons (17), and in cisplatin-injured colorectal carcinoma cells (4). The effect of Siva inhibition on renal functions and apoptosis commensurates with the results of p53 inhibition in mice (20). These data further suggest that the expression of Siva can recapitulate the apoptotic response of p53 in the renal IRI setting and Siva activation may be sufficient to carry out the p53 downstream apoptotic program.

Although Siva is well-established as an apoptosis-inducing molecule, the exact mechanisms by which Siva elicits apoptosis are not well-defined. Siva shares no homology with other proteins, including any significant homology to Bcl2 family members. However, Siva is shown to bind to TNF receptors, CD27 and GITR, and induce apoptosis (35, 42). Using a membrane fractionation approach, a recent report demonstrated that Siva could associate with the plasma membrane during apoptosis (17). Overexpression studies using epitope-tagged Siva demonstrated that Siva can localize to the cytoplasm diffusely, mitochondria, and unidentified substructures in the nucleus (36). Interaction studies previously demonstrated that Siva can bind to a variety of molecules including Bcl2 and Bclxl. It is proposed that Siva may counteract the prosurvival functions of Bcl2 or Bclxl, which may result in Bax oligomerization and mitochondrial pore formation (4).

To further elucidate the mechanisms and the importance of subcellular localization of Siva in the induction of apoptosis in renal epithelial cells, we transiently transfected GFP-fused Siva in the porcine-derived proximal tubular cell line, LLC-PK₁. We acknowledge that the interpretation of the results from LLC-PK₁ cells may not directly correlate to the events that may occur in mouse kidneys under similar conditions. Siva displayed a distinct spatiotemporal pattern as early as 3 h following transfection of GFP-Siva in LLC-PK₁ cells. GFP-Siva consistently translocated to the nucleus at 3 h after a weak and brief expression in the cytoplasm; however, no significant apoptotic changes have been observed at this stage suggesting a critical preapoptotic role of Siva in the nucleus. Thus, it is not clear whether Siva plays an earlier anti-survival or proapoptotic role at the nuclear level. At 6 h after transfection, GFP-Siva consistently forms a beaded appearance along the periphery of the nucleus; however, the significance of this centrifugal aggregation of GFP-Siva in the nucleus is currently unknown (36).

At 12 h following transfection, GFP-Siva appeared on the plasma membrane suggesting interactions with its cognate receptor CD27 and possibly triggering the caspase cascade that involves caspases-8 and -3. Thus, to determine whether the absence of CD27 protects renal tissues from IRI, we assessed renal function and apoptosis in CD27^−/− mice compared with its control mice. CD27 deficiency protected mice from deterioration of renal function and partially inhibited apoptosis, suggesting that CD27 contributes to IRI-induced cell death. However, in CD27^−/− mice, the percentage of apoptosis was downregulated by only 44%. This suggests that CD27 may play only a partial role in eliciting the p53-Siva-triggered downstream apoptotic pathways and parallel pathways may be activated by p53-Siva apoptotic stimuli to fully execute the apoptotic program. Our overexpression studies also showed that GFP-Siva localized in the cytoplasm in a punctate pattern suggesting mitochondrial localization. Furthermore, Siva colocalized with the mitochondrial labeling dye, Mitotracker Deep Red (Molecular Probes), indicating the mitochondrial localization of Siva. Transfection of pCMV-FLAG-Siva displayed a significant loss of Δψ_m, indicating that Siva mediates MMP upon translocation to mitochondria.

MMP results in the release of proapoptotic molecules such as cytochrome c and AIF and cessation of the bioenergetic functions of mitochondria leading to cell death (22). Nevertheless, depending on the model of apoptosis, the release of cytochrome c can occur earlier or later to the release of AIF and has been interpreted as a differential release mechanism (30, 38). Moreover, the release of cytochrome c can also be independent of MMP (46). Thus, examination of the release of cytochrome c following Siva overexpression is important in renal epithelial cells. At 9 h after GFP-Siva overexpression, both cytochrome c and AIF were diffusely present in the cytoplasm, indicating the release of mitochondrial proapoptotic molecules following the membrane depolarization and permeabilization. In normal cells, both AIF and cytochrome c reside in the intermembrane space of the mitochondria and following induction of MMP, AIF and cytochrome c are released into the cytoplasm in different injury models (21, 27, 43). Upon release from the mitochondria to the cytoplasm, AIF translocates to the nucleus and initiates nucleosomal fragmentation independent of the involvement of caspases (10, 43). Here, we show for the first time that Siva mediates the mitochondrial-nuclear translocation of AIF following overexpression implying that Siva mediates the caspase-independent apoptotic pathway. Conversely, cytochrome c, following the mitochondrial release, interacts with the WD40 domains of apoptotic protease-activating factor 1 (Apaf-1) to activate procaspase-9 to active caspase-9, which subsequently initiates the caspase-dependent cascade of apoptosis (54). The functional relevance of Siva in cytochrome c release and AIF nuclear translocation in inducing apoptosis in ischemic kidneys was assessed by downregulating Siva expression post-IRI. Our results indicate that Siva inhibition attenuated both cytochrome c release and AIF nuclear translocation, suggesting Siva-mediated apoptotic pathways involve these postmitochondrial cell death effectors.

In summary, Siva antisense oligonucleotides protect renal functions from IRI by reducing cell death in proximal tubular epithelium. In renal tubular epithelial cells, Siva upregulation induces apoptotic cell death via caspase-8 activation through interactions with its cognate receptor CD27 and by mitochondrial translocation resulting in MMP. Subsequently, Siva mediates caspase-9 activation via the release of cytochrome c and caspase-independent apoptosis via AIF release from mitochondria. Inhibition of Siva prevented AIF and cytochrome c translocation further confirming that Siva plays a key role in their release. We conclude that in renal tubular epithelial cells, Siva mediates apoptosis via both caspase-dependent and caspase-independent apoptotic pathways (Fig. 8). Combination therapies that inhibit both caspase-dependent and -independent pathways may achieve improved efficacy as elegantly demonstrated in murine models of neonatal hypoxia/ischemia (53) or retinal detachment (16). Thus, targeting Siva may prevent activation of both of these pathways and offer a newer and more specific therapeutic strategy for the management of AKI.

Fig. 8.

Hypothetical schema of the proposed mechanisms by which Siva induces apoptosis in ischemic kidneys: p53 mediates apoptosis in renal PTC by activating the proapoptotic transcriptional target Siva. Activation of Siva simultaneously induces caspase-dependent and -independent pathways. Translocation of Siva to the plasma membrane induces caspase-8 activation possibly through the TNF receptor CD27. Mitochondrial translocation of Siva induces mitochondrial permeability and releases proapoptotic molecule cytochrome c leading to caspase-9 activation. Siva-mediated mitochondrial release of AIF may result in its translocation to the nucleus to induce DNA fragmentation and caspase-independent cell death.

GRANTS

This work was supported by a Nebraska Kidney Association research grant to B. J. Padanilam and UNMC graduate fellowship to K. Singaravelu. The work was also partially supported by National Institutes of Health Grant DK083291to B. J. Padanilam.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).