Kidney ischemia-reperfusion injury induces caspase-dependent pulmonary apoptosis

Abstract

Distant organ effects of acute kidney injury (AKI) are a leading cause of morbidity and mortality. While little is known about the underlying mechanisms, limited data suggest a role for inflammation and apoptosis. Utilizing a lung candidate gene discovery approach in a mouse model of ischemic AKI-induced lung dysfunction, we identified prominent lung activation of 66 apoptosis-related genes at 6 and/or 36 h following ischemia, of which 6 genes represent the tumor necrosis factor receptor (TNFR) superfamily, and another 23 genes are associated with the TNFR pathway. Given that pulmonary apoptosis is an important pathogenic mechanism of acute lung injury (ALI), we hypothesized that AKI leads to pulmonary proapoptotic pathways that facilitate lung injury and inflammation. Functional correlation with 1) terminal deoxynucleotidyl transferase-mediated dUTP nick end-labeling and 2) active caspase-3 (aC3) activity, immunoblotting, and immunohistochemistry (IHC) identified kidney IRI-induced pulmonary apoptosis at 24 h, and colocalization studies with CD34 identified predominantly endothelial apoptosis. Mice were treated with the caspase inhibitor Z-VAD-FMK (0.25 mg ip) or vehicle 1 h before and 8 h after sham or kidney IRI, and bronchoalveolar lavage fluid protein was measured at 36 h as a surrogate for lung leak. Caspase inhibition reduced lung microvascular changes after kidney IRI. The pulmonary apoptosis seen in wild-type control mice during AKI was absent in TNFR^−/− mice. Using an initial genomic approach to discovery followed by a mechanistic approach to disease targeting, we demonstrate that pulmonary endothelial apoptosis is a direct mediator of the distant organ dysfunction during experimental AKI.

acute kidney injury (AKI) occurs in 2–5% of hospitalized patients and despite advances in renal replacement therapy, the mortality of patients with AKI has remained unacceptably high (2, 29). Epidemiological studies have identified an association between AKI and dysfunction of extrarenal organs such as the lung (6, 7). Acute lung injury (ALI) is a major cause of respiratory failure in the ICU and carries a substantial mortality of 30–40%, but when combined with AKI, mortality approaches 80% (8, 35). The clinical presentation of AKI-associated ALI, similar to that of other causes of ALI, is characterized by increased pulmonary edema, a defining feature of the syndrome. These lung changes during AKI are generally considered to be caused by volume overload due to impaired fluid clearance by damaged kidneys. However, recent experimental data have demonstrated that AKI directly leads to pulmonary inflammation and abnormalities in salt and water transporters (28, 40). In addition, AKI leads to a systemic increase in serum cytokines and chemokines that can mediate these lung changes, and depending on the timing and characteristics of a secondary insult in the lung, AKI can alter the injury response to mechanical ventilation (11, 23, 25, 37, 47).

While ALI is traditionally characterized by acute inflammation of the air spaces and parenchyma with accumulation of inflammatory cells and loss of barrier function in both the alveolar epithelium and capillary endothelium, increasing evidence suggests that pulmonary cellular apoptosis may also play an important role in the pathophysiology of ALI (13, 14, 22). The pulmonary endothelium performs several homeostatic functions including regulation of solute transport between vascular compartments and surrounding tissues, and it functions as a semipermeable barrier dynamically regulated by the cytoskeleton. Detailed investigations of vascular permeability have underscored the importance of the balance between complex tethering forces involved in cell-cell and cell-extracellular matrix (ECM) interactions, and endothelial apoptosis leads to disruption of these complex interactions and a potential for loss of endothelial barrier function (4, 24).

Given the complexity of interorgan cross talk and the need to discover novel pathways that are functionally important, we have initiated a gene-based discovery approach in a murine model of experimental AKI. In our prior studies, we discovered that kidney ischemia-reperfusion induces functional and transcriptional changes in the mouse lung which are distinct from those induced by uremia alone and have identified potential inflammation-associated mediators that connect local and systemic injury (17, 22). However, in these genome-based studies, analysis identified dysregulation of many genes not traditionally associated with inflammation. In the current study, we focused on apoptosis-related genes and found a marked change in transcription of lung genes associated with apoptosis during both the early and later periods after AKI, particularly of the prominent cell death receptor TNFR1.

Given the evidence of pulmonary apoptosis by gene expression profiling, plus the existing data on apoptosis in acute lung injury, we hypothesized that pulmonary endothelial cell apoptosis mediates the lung dysfunction following AKI. We found that ischemic AKI led to an increase in lung caspases, as well as specific endothelial proapoptotic changes. Blockade with the caspase inhibitor benzyloxycarbonyl-Val-Ala-Asp-fluoromethylketone (Z-VAD-FMK) significantly reduced AKI's effects on lung caspase activation as well as microvascular protein leak. Finally, kidney IRI failed to induce lung apoptosis in TNFR1^−/− mice compared with wild-type controls, suggesting TNFR1-dependent pulmonary endothelial apoptosis as a mechanism for distant organ dysfunction in this model of AKI-induced lung injury.

MATERIALS AND METHODS

Animal Care

All procedures were approved by the Johns Hopkins Animal Care and Use Committee and were consistent with the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals. Male 6- to 8-wk-old mice (C57BL6/J or TNFR1^−/−), weighing ∼25–30 g, were obtained from Jackson Laboratory (Bar Harbor, ME) and housed under pathogen-free conditions according to NIH guidelines at least 5 days before operative procedures. All procedures were performed using strict sterile techniques under general anesthesia with pentobarbital sodium (50 mg/kg ip). Assessment of adequate anesthesia was obtained by paw and tail pinching.

Surgical Procedures

Animals were placed on a heating blanket and underwent a midline laparotomy with isolation of bilateral renal pedicles. For mice assigned to experimental ischemia-reperfusion injury (IRI), a nontraumatic microvascular clamp was applied across both renal pedicles for 60 min. After the allotted ischemia time, the clamps were gently removed, the animals were administered 1 ml of sterile saline intraperitoneally, and the incision closed in two layers with a 4-0 silk suture. The animals were then allowed to recover with free access to food and water. Sham animals underwent the identical procedure without placement of the vascular clamps. The mice assigned to bilateral nephrectomy (BNx) underwent similar procedures except that both renal pedicles were ligated with a 5-0 silk suture and the kidneys were removed. In some experiments, a total of 0.5 mg of Z-VAD-FMK (Promega, Madison, WI) was administered by equal-dose intraperitoneal injections at 1 h before and 8 h after the laparotomy. At 6, 24, or 36 h following the experimental procedure, the mice were euthanized by exsanguination under pentobarbital sodium anesthesia, and tissues were collected for analysis.

Bronchoalveolar Lavage

Bronchoalveolar lavage (BAL) fluid analysis was performed as a surrogate measurement of pulmonary vascular permeability as described by Walters et al. (46). BAL fluid was obtained by slow delivery of 0.75 ml of warm (∼37°C) PBS via a tracheotomy. The fluid was withdrawn by gentle suction, the process was repeated twice, and the solutions were combined and stored on ice. Recovered BAL fluid was centrifuged at 1,500 rpm at 4°C for 10 min. BCA total protein assay (Pierce, Rockford, IL) was performed on BAL supernatants according to the manufacturer's protocol. Optical density readings of samples were converted to micrograms per milliliter using values obtained from standard curves generated with serial dilutions of BSA.

Assessment of Apoptosis

Terminal deoxynucleotidyl transferase-mediated dUTP nick end-labeling (TUNEL) assays were performed with a Fluorescein-FlagEL DNA Fragmentation Detection kit following the manufacturer's instructions (Calbiochem, Gibbstown, NJ). Briefly, after deparaffinization and rehydration, the slides were steamed for antigen retrieval in Vector Antigen Unmascking Solution (Vector Laboratories, Burlingame, CA) for 15 min followed by digestion with 20 μg/ml proteinase K. The slides were then immersed in 1× TdT Equilibration Buffer and incubated at room temperature for 30 min followed by incubation with Fluorescein-FlagEL TdT labeling mixture and TdT enzyme for 1 h at 37°C. For colocalization studies, the slides were blocked with avidin-biotin blocking solution (Vector Laboratories) and incubated with either mouse anti-CD34 monoclonal antibody (Zymed Laboratories, San Francisco, CA) or anti-SPC (Chemicon, Bedford, MA) for 1 h followed by incubation with secondary antibody goat polyclonal to mouse IgG (Abcam, Cambridge, MA) and streptavidin Texas red conjugate (Molecular Probes, Eugene, OR). Staining was completed by a 1-min incubation with 4′6-diamidino-2-phenylindole dihydrochloride (DAPI; Molecular Probes), and coverslips were mounted on slides. Measurements of TUNEL-positive nuclei were performed on 10–15 images/slide captured by an independent observer blinded to the experiment using an Olympus microscope connected to a digital camera, which was attached to an image processor, and images were recorded and saved using Adobe Photoshop 6 (Adobe Systems, San Jose, CA).

Active caspase-3 immunohistochemistry (IHC) was performed utilizing an EnVision visualization system (Dako). Briefly, immunolocalization was performed on paraffin-embedded, formalin-fixed mouse lungs. After paraffin removal in xylene, the sections were rehydrated and submitted to heat-steam treatment for 30 min in a 10 mM citric acid monohydrate solution. The endogenous peroxidase activity was quenched by incubating the specimen for 5 min with Peroxidase Block. The specimens were then incubated with anti-active caspase-3 rabbit primary antibody (Abcam) overnight at 4°C followed by incubation with the labeled polymer for 30 min. Staining was completed by incubation with 3,3′-diaminobenzidine (DAB+) substrate-chromogen, which results in a brown-colored precipitate at the antigen site. Measurements of active caspase-3-positive cells were performed on 10–15 images/slide captured by an independent observer blinded to the experiment and normalized to total cell count by DAPI staining.

Caspase-3 activity was measured by fluorometric assay (BD Biosciences) as described elsewhere (22). Briefly, the right lung was homogenized immediately after isolation in buffer containing 20 mM HEPES, 51.5 mM MgCl₂, 150 mM NaCl, 10% glycerol, 1% Triton X-100, 2 mM EDTA, 2 mM Na₃VO₄ (sodium orthovanadate), 50 mM NaF (sodium fluoride), 1 mM PMSF (phenylmethylsulfonyl fluoride), 1× Protease Inhibitor Cocktail Set I (no. 539131, Calbiochem), 500 μM AEBSF, 150 nM aprotinin, 1 μM E-64, 500 μM EDTA, and 1 μM leupeptin. Lung tissue homogenates were centrifuged at 10,000 g for 15 min, and the supernatants were stored at −80°C until used for the caspase-3 activity assay. Direct protein quantitation was performed using a D_C protein assay kit (Bio-Rad, Hercules, CA). Equal amounts of protein were added to each well in a 96-well plate in 200 μl 1× HEPES buffer. For each reaction, 5 μl reconstituted Ac-DEVD-AMC was added followed by a 1-h incubation at 37°C. The amount of AMC elaborated from Ac-DEVD-AMC was measured using a plate reader with an excitation wavelength of 380 nm and an emission wavelength range of 420–460 nm and expressed as international units per milligram protein.

TNFR1 and Caspase-3 Western Blot Analyses

Aliquots from tissue homogenates prepared as described above were assayed for protein measurement using the Bradford protein assay. Equal amounts of protein (50 μg) were then loaded in each well of 12% Tris-glycine gels (Bio-Rad). After electrophoresis for 90 min at 125 V of constant voltage, the gel was blotted onto a nitrocellulose membrane by electrophoretic transfer at 70 V of constant voltage for 1–2 h. The membrane was then washed, blocked with 5% blocking solution, and probed with cleaved caspase-3 antibody (Cell Signaling, Danvers, MA), which detects both 19- and 17-kDa lengths of the cleaved caspase-3 or TNFR1 antibody (Santa Cruz Biotechnology, Santa Cruz, CA). The immunoreactive bands were visualized using a secondary antibody conjugated to horseradish peroxidase and a chemiluminescent detection system (Pierce, Rockford, IL). Nitrocellulose blots were visualized using a Bio-Rad Gel doc 1000 (Bio-Rad) and analyzed by National Institutes of Health Image J 1.61 software.

Microarray Methods

Transcript profiling with Affymetrix oligonucleotide arrays.

The gene expression profiling was conducted at the Lowe Family Genomics Core using Affymetrix GeneChip MG-430A 2.0 (MOE430A 22,626 transcripts) as previously described (17, 22). Briefly, power prediction analysis for minimum array requirements demonstrated that three chips per group would generate statistically sound results. The quality of the entire hybridization procedure was monitored on an Agilent 2100 Bioanalyzer. The stringent quality control of the purity and integrity of the RNA was assessed using standard criteria for RNA quality, and only samples with yields of in vitro transcribed RNA above 45 μg were hybridized to the chips. After the GeneChips were scanned and the quality of fluorescent intensities of microarray hybridization signals was assessed, we proceeded with the bioinformatics analysis.

Identification of AKI-associated genes and biological processes.

Digitized matrices of fluorescent intensities (CEL files) were processed and normalized as described previously (20). The digitized hybridization signals were analyzed on the probe level, and probes that produced detectable hybridization signal in <5% of hybridizations were considered nonfunctional and flagged for masking. The subsequent expression data analysis was conducted applying electronic masking of nonfunctional probes, and matrices of raw expression data were generated (19). The significance of hybridization signals was tested by GCOS 1.4, and “Present” (P < 0.05) and “Absent” (P > 0.065) transcripts were identified. The chip background and brightness were computed using high quartile and whole sets of “Absent” hybridization signals, respectively. The expression data were stratified by experimental conditions (n = 3), and hybridization of each transcript was evaluated for each cluster. The transcripts that were identified as “Present” by the GCOS 1.4 algorithm and produced a signal at least twice as high as that of a background in at least 75% of hybridizations in any given group of mice were considered expressed. The signal intensity values of the resulting 15,332 transcripts from each chip were increased by corresponding to a given chip's background value (background adjustment) and divided by a chip's brightness coefficient (normalization). Significance Analysis of Microarrays (SAM 2.20) of processed expression values was conducted using full 720 permutations of 3 sham control and 3 IRI samples.

Based on array power prediction and applying the relax filtering concept for detecting relevant gene ontologies, genes with a 1.5-fold change and 5% false discovery rate were selected for identification of AKI-associated biological processes (25). Gene ontology (GO) analysis was conducted as described previously (17, 22). Briefly, the detectable 15,332 probe sets were scanned for redundancy and extra probe sets that targeted the same transcript were not further considered. The remaining 14,228 probe sets were assigned corresponding IRI-induced fold-change (FC) and false discovery rate (FDR) values, where undetectable (Absent) probe sets were assigned FC = 1 and FDR = 100%. The resulting data set was processed by GenMAPP 2.1 considering all genes with FC >1.5 and FDR <5% as involved in corresponding biological pathways and analyzed by MAPPFinder 2.0. GO terms for each mouse probe set were enriched by cross-referencing with their human orthologs provided by NetAffx (http://www.affymetrix.com/analysis/index.affx) using Microsoft Access. Resulting data sets were then dynamically linked to the human GenMAPP database Hs-Std_20070817.gdb using Affymetrix probe set identifiers as a linkage key. Biological processes with a Z score >2 in which at least 1% of all GO-linked gene were involved in the first GO node and 5% of all GO-linked gene were involved in the nested GO node were considered significantly affected by AKI. The nonspecific gene ontologies (i.e., signal transduction) that are represented by >1,000 genes and common terms (i.e., protein kinase cascade or phosphorylation) were not considered. The gene composition of identified bioprocesses was extracted from the MAPPFinder output using the GO-Quant tool (http://depts.washington.edu/irarc/Go-Quant/index.html). Finally, the genomic network analysis of apoptosis-related genes (gene symbols from Table 2 were submitted for this analysis), which identifies known interactions between candidate genes, was conducted using the Ingenuity Pathways Knowledge Base tool (http://www.ingenuity.com) as described previously (20).

Statistical Analysis

Data are expressed as means ± SE and were analyzed with one-way ANOVA. Individual group means were then compared with a Tukey multiple comparison test, and P values <0.05 were considered significant.

RESULTS

Kidney IRI Induces a Proapoptotic Transcriptome in the Mouse Lung

We have previously identified that kidney IRI induces a unique genomic response in the mouse lung that is distinguishable from that induced by uremia alone (i.e., bilateral nephrectomy) (22). In the current study, we performed a specific analysis of apoptosis-related genes activated in the mouse lung during ischemic AKI. Total RNA from the lung was isolated from three to five separate animals per group obtained at 6 or 36 h after sham, kidney IRI, or BNx, and hybridized to Affymetrix Mouse MOE430A GeneChips. Identified transcriptional changes were dynamically linked to corresponding Gene Ontology (GO) terms, and IRI-associated biological pathways were calculated. This analysis identified 518 and 267 genes that were uniquely involved in IRI-induced lung transcriptional changes within multiple bioprocesses at 6 and 36 h, respectively. Apoptosis- and inflammation-related genes were among pathways represented by the highest number of IRI-affected genes across time points. GO analysis also identified early (i.e., 6 h) activation of genes involved in the “innate immume response” and “cell adhesion,” and later activation of genes related to the “response to stress” and “cell motility” bioprocesses (Table 1).

Table 1.

Biological processes activated in the lung during ischemic AKI at 6 and 36 h

GO ID	Biological Process Name	Bioprocess-Associated Genes		Z Score
		Changed/detected	% Changed
6 h
6955	Immune response	57/428	13.32	7.389
6915	Apoptosis	51/595	8.57	3.511
7155	Cell adhesion	41/491	8.35	2.949
6954	Inflammatory response	35/232	15.09	6.588
36 h
6950	Response to stress	35/738	4.74	3.345
6915	Apoptosis	27/597	4.52	2.648
6928	Cell motility	21/347	6.05	3.752
6954	Inflammatory response	14/232	6.03	3.033

AKI, acute kidney injury; GO, gene ontology.

Apoptosis-specific transcriptional changes induced by ischemic AKI.

Details of the GO analysis are listed in Table 1; kidney IRI activated 51 (8.6%) apoptosis-related lung genes at 6 h (Z score = 3.5) and 27 (4.5%) genes at 36 h (Z score = 2.6). Twelve genes were activated across time points, and therefore kidney IRI activated a total of 66 unique apoptosis-related lung genes which are listed in Table 2. To identify potential lung targets of the distant effectors from injured kidneys, we conducted a genomic network analysis (i.e., Ingenuity Pathway Analysis) based on these 66 activated lung transcripts and focused on membrane receptors. This analysis identified activation of genes from three major transmembrane receptors, including those from the Fas and TNFR family, which form a node functionally linked to another node composed of caspase-related genes (Fig. 1).

Table 2.

Transcriptional changes in lung apoptosis genes 6 and 36 h after kidney IRI

Gene Title	Gene Symbol	6-h IRI		36-h IRI
		FC	FDR	FC	FDR
Significantly upregulated at 6-h IRI
B-cell leukemia/lymphoma 3	BclIII	6.37	<0.1	1.66	17.52
CD14 antigen	Cd14	4.08	<0.1	1.70	21.35
CCAAT/enhancer binding protein (C/EBP), β	Cebpb	3.73	<0.1	1.56	37.24
G protein-coupled receptor 65	Gpr65	3.33	0.569	1.12	55.69
Activin A receptor, type 1	Acvr1	3.04	0.905	−1.79	35.30
PYD and CARD domain-containing	Pycard	3.03	<0.1	1.16	54.53
Baculoviral IAP repeat-containing 5	Birc5	2.86	2.908	1.52	49.21
Cell division cycle 2 homolog A (S. Pombe)	Cdc2a	2.64	2.908	−2.97	0.970
LPS-induced TN factor	Litaf	2.57	<0.1	1.70	14.36
Ring finger and FYVE-like domain-containing protein	Rffl	2.51	3.147	1.35	52.07
Lymphotoxin B receptor	Ltbr	2.50	0.905	1.80	19.25
Microtubule-associated serine/threonine kinase 1	Mast1	2.44	2.126	1.05	55.69
Thioredoxin-like 1	Txnl1	2.44	3.147	−1.50	14.36
Myelocytomatosis oncogene	Myc	2.43	3.147	−1.01	48.89
Tumor necrosis factor receptor superfamily, member 1b	Tnfrsf1b	2.41	0.905	1.95	8.465
Phosphoinositide-3-kinase, catalytic, γ polypeptide	Pik3cg	2.30	3.147	1.82	19.25
MARCKS-like 1	Marcksl1	2.18	4.359	−1.09	42.38
Myeloid cell leukemia sequence 1	Mcl1	2.13	1.656	1.29	49.78
BPY2-interacting protein 1	Bpy2ip1	2.11	1.647	−1.25	30.96
BH3-interacting domain death agonist	Bid	2.11	4.707	1.68	21.35
Receptor-interacting serine/threonine kinase 3	Ripk3	2.05	2.908	1.57	35.30
Fas (TNF receptor superfamily member)	Fas	1.94	3.790	2.36	15.89
Prokineticin 2	Prok2	1.91	3.147	−1.09	48.89
Interleukin-1β	Il1b	1.90	3.425	1.57	42.38
Hypoxia-inducible factor 1, α-subunit	Hif1a	1.89	2.929	1.07	55.69
Caspase-11, apoptosis-related cysteine peptidase	Casp11	1.86	2.929	1.13	54.24
Serine/threonine kinase 17b (apoptosis-inducing)	Stk17b	1.85	2.929	−1.39	19.25
BCL2-antagonist/killer 1	Bak1	1.82	3.147	1.45	32.92
Programmed cell death 6-interacting protein	Pdcd6ip	1.81	4.359	1.20	55.69
Rab6-interacting protein 2	Rab6ip2	1.79	4.359	−1.14	42.38
Transforming growth factor-β-regulated gene 4	Tbrg4	1.79	3.790	−1.06	55.69
CD38 antigen	Cd38	1.78	4.707	1.34	42.38
Baculoviral IAP repeat-containing 3	Birc3	1.78	3.425	2.20	5.220
Caspase-1	Casp1	1.78	3.147	−1.32	32.92
B-cell leukemia/lymphoma 6	Bcl6	1.78	3.147	−1.59	14.36
Baculoviral IAP repeat-containing 1e	Birc1e	1.67	3.790	1.00	55.69
Tumor necrosis factor receptor superfamily, member 1a	Tnfrsf1a	1.65	3.790	1.26	49.21
Mitochondrial ribosomal protein S30	Mrps30	1.60	4.707	1.10	55.69
Bridging integrator 3	Bin3	1.52	4.707	1.32	48.57
Significantly upregulated at both 6- and 36-h IRI
Suppressor of cytokine signaling 3	Socs3	7.39	<0.1	2.64	2.867
Growth arrest and DNA-damage-inducible 45 γ	Gadd45 g	2.71	2.126	2.37	3.611
Cyclin-dependent kinase inhibitor 1A (P21)	Cdkn1a	2.55	2.126	6.93	<0.1
Ras association (ralgds/AF-6) domain family 5	Rassf5	2.49	0.905	1.26	49.787
Matrix metallopeptidase 9	Mmp9	2.18	1.647	3.58	0.957
Son cell proliferation protein	Son	2.09	4.359	2.14	21.359
Proviral integration site 1	Pim1	2.05	2.908	2.51	4.485
Angiopoietin-like 4	Angptl4	2.04	2.929	9.54	<0.1
Feminization 1 homolog b (C. elegans)	Fem1b	2.03	4.707	4.80	<0.1
Integrin-β2	Itgb2	1.95	3.147	3.96	4.485
Serum/glucocorticoid-regulated kinase	Sgk	1.90	2.929	3.57	<0.1
NF-κB inhibitor, α	Nfkbia	1.73	3.147	3.89	<0.1
Significantly upregulated at 36-h IRI
Transformation related protein 53 inducible nuclear protein 1	Trp53inp1	−1.03	53.63	4.26	0.367
Death effector domain-containing DNA binding protein 2	Dedd2	1.56	5.462	3.98	1.251
Xeroderma pigmentosum, complementation group A	Xpa	−1.05	56.09	3.54	<0.1
Sphingosine kinase 1	Sphk1	2.14	7.107	3.38	0.482
Paternally expressed 3	Peg3	1.35	21.15	3.10	1.251
Tumor necrosis factor receptor superfamily, member 12a	Tnfrsf12a	−1.39	29.21	3.02	4.098
BCL2-like 11 (apoptosis facilitator)	Bcl2l11	−1.08	55.26	2.79	1.505
Insulin-like growth factor binding protein 3	Igfbp3	−1.29	50.10	2.71	1.780
Tumor differentially expressed 1	Tde1	1.16	31.94	2.70	2.176
Transmembrane protein 23	Tmem23	1.34	23.84	2.67	3.611
TSC22 domain family 3	Tsc22d3	1.37	14.79	2.59	3.611
STE20-like kinase (yeast)	Slk	1.50	10.08	2.48	2.867
CASP8 and FADD-like apoptosis regulator	Cflar	−1.38	26.33	2.47	4.098
Dynamin 2	Dnm2	1.35	14.79	2.28	3.611
Tnf receptor-associated factor 6	Traf6	1.55	6.349	2.12	4.098

IRI, ischemia-reperfusion injury; FC, fold-change; FDR, false discovery rate.

Fig. 1.

Ingenuity pathway analysis of apoptosis-related lung candidate genes. The Ingenuity modified network number 1 built from apoptosis-related candidate genes (listed in Table 2) is presented, and only genes with at least 2 relationships are depicted. The relationship between genes is presented as follows: solid lines (binding), solid arrows (direct activation), or broken arrows (indirect activation). Arrows point to the element on which an action is performed. Annotation of relationship: A, activation; E, expression; L, proteolysis; LO, localization; MB, group/complex membership; P, phosphorylation/dephosphorylation; PP, protein-protein binding; RB, regulation of binding; TR, translocation. CASP, caspase; TNFR, TNF receptor.

Kidney IRI Induces Pulmonary Endothelial Apoptosis

To correlate these transcriptional events in the lung during ischemic AKI with functional and structural endpoints, multiple techniques were used to assess the potential extent, specificity, and localization of pulmonary cellular apoptosis during experimental AKI. Our prior studies had identified that kidney IRI induces lung microvascular changes at 36 h, but not at the earlier 6-h time point, following ischemia (22). Therefore, TUNEL assays were performed on lung tissue at 6 or 24 h following sham, kidney IRI, or BNx to assess for potential antecedent pulmonary cellular apoptosis as a potential mechanism for kidney IRI-induced lung leak. TUNEL-positive nuclei were identified in lung alveolar spaces at 24 h following kidney IRI, but not at 6 h (Fig. 2). Furthermore, TUNEL-positive cells were not present following sham or BNx at either time point. To corroborate the TUNEL findings and to quantify lung apoptosis, we measured activation of caspase-3 in mouse lung tissues at 6 or 24 h following sham, kidney IRI, or BNx. First, we performed active caspase-3 IHC and the numbers of positive cells were calculated as outlined in materials and methods. Compared with sham controls, there was a significant increase in the number of active caspase-3 positively-stained cells (1.0 ± 0.3 vs. 4.0 ± 0.6, P < 0.01) in mouse lung at 24 h after kidney IRI, but not following BNx (1.0 ± 0.3 vs. 1.9 ± 0.3, P = 0.06) or at the earlier 6-h time point (Fig. 3).

Fig. 2.

Effect of ischemic acute kidney injury (AKI) on pulmonary apoptosis by terminal deoxynucleotidyl transferase-mediated dUTP nick end-labeling (TUNEL). A: representative lung immunofluorescence micrographs obtained 24 h following bilateral 60-min kidney ischemia (ischemia-reperfusion injury; IRI), bilateral nephrectomy (BNx), or sham. Costaining with TUNEL (green) and DAPI (blue) identified alveolar cells undergoing apoptosis (arrow), and this was most prominent in lungs following kidney IRI. B: measurements of TUNEL-positive cells were performed on 10–15 images/slide and expressed as cells/high-power field. IRI-treated mice demonstrated increased TUNEL-positive cells compared with sham or BNx; n = ≥5/group. *P < 0.05.

Fig. 3.

Lung caspase-3 activation during ischemic AKI. A: representative lung caspase-3 IHC micrographs (×40) of mouse lung at 24 h following sham, IRI, or BNx. B: measurements of active caspase-3-positive cells were performed on 10–15 images/slide and expressed as cells per high-power field. IRI-treated mice demonstrated increased caspase-3-positive cells at 24 h, but not at the earlier 6-h time point, compared with sham or BNx; n = ≥5/group. *P < 0.05. C: representative Western immunoblot of active (cleaved) caspase-3 protein expression in lung tissue homogenates obtained 24 h following sham, from 60-min bilateral kidney IRI, or from bilateral BNx demonstrated increase caspase-3 expression during kidney IRI. D: caspase-3 activity measured by fluorometric assay and expressed in normalized international units (IU) per milligram lung tissue protein demonstrated increase lung caspase-3 activity 24 h following IRI, but not at the earlier 6-h time point, compared with sham or BNx; n = ≥5/group. *P < 0.05.

To confirm and quantify increased pulmonary cellular apoptosis, active caspase-3 tissue protein levels were measured by Western immunoblotting, which confirmed increased expression in lung 24 h following kidney IRI but not after BNx. Additionally, active caspase-3 enzyme activity, measured in lung tissue homogenates obtained 6 or 24 h following surgery, increased after kidney IRI compared with sham (100 ± 6 vs. 216 ± 15, P < 0.01) but not following BNx (100 ± 6 vs. 92 ± 10) or at the earlier 6-h time point (Fig. 3). To identify the primary cellular components of ischemic AKI-induced lung apoptosis, we performed TUNEL immunofluorescence and parallel colocalization staining with SPC for type-2 alveolar epithelial cells (Fig. 4) and CD34 for endothelial cells (Fig. 5). Kidney IRI induced primarily lung endothelial cell apoptosis as indicated by predominant costaining of TUNEL- and CD34-positive cells, but also some type-2 epithelial cells. Taken together, these data indicate that ischemic AKI induces distant organ apoptosis as evidenced by 1) activation of early proapoptotic transcriptional events and 2) subsequent structural confirmation by TUNEL-positive nuclei and expression of downstream executioner caspases in the pulmonary endothelium.

Fig. 4.

Type II lung epithelial cell apoptosis during ischemic AKI. A: lung immunofluorescence micrographs (×40) obtained 24 h following 60-min ischemia of bilateral kidney stained with DAPI (A; blue), TUNEL (B; green), SPC (C; red), or composite SPC-TUNEL-DAPI (D). The majority of TUNEL-and SPC-positive cells failed to colocalize, suggesting limited type II epithelial cell apoptosis under these conditions.

Fig. 5.

Lung endothelial cell apoptosis during ischemic AKI. Lung immunofluorescence micrographs (×40) obtained 24 h following 60-min ischemia of bilateral kidney stained with DAPI (A; blue), TUNEL (B; green), CD34 (C; red), or composite CD34-TUNEL-DAPI (D). TUNEL and CD34 colocalized (yellow arrow), suggesting lung endothelial cell apoptosis in this model of ischemic AKI.

Effect of Caspase Inhibition on Lung Injury During Experimental AKI

To establish a potential direct role for pulmonary endothelial apoptosis in the previously observed ischemic AKI-induced lung microvascular injury, we measured BAL protein levels (μg/ml) in mice 36 h following sham or kidney IRI, with or without administration of the caspase inhibitor Z-VAD-FMK (0.5 mg ip). Control mice that underwent IRI demonstrated increased BAL protein compared with sham, while IRI mice treated with Z-VAD-FMK had BAL protein levels not different from sham controls (Fig. 6). In addition, mice treated with Z-VAD-FMK demonstrated decreased caspase-3 activity in lung tissue homogenates, suggesting effective drug delivery in vivo (Fig. 6). These results indicate that pulmonary endothelial apoptosis is an important regulatory event in kidney IRI-induce lung microvascular injury.

Fig. 6.

Effect of Z-VAD-FMK on ischemic AKI-induced lung injury and apoptosis. A: bronchoalveolar lavage (BAL) protein was measured at 36 h following sham or 60-min bilateral kidney ischemia (IRI) in saline (control)- or Z-VAD-FMK-treated mice. BAL protein was significantly increased following IRI in control, but not Z-VAD-FMK-treated mice. B: caspase-3 activity was measured in lung tissue homogenates obtained at 36 h following sham or 60-min bilateral kidney ischemia (IRI) in saline (control)- or Z-VAD-FMK-treated mice, expressed in IU/mg lung tissue protein. Caspase-3 activity was significantly increased following IRI in control, but not Z-VAD-FMK-treated mice; n = ≥5/group. *P < 0.05.

TNFR1-Dependent Pulmonary Apoptosis

The TNFR superfamily of cell surface receptor proteins play a critical role in many cell regulations including death ligand-mediated apoptosis. To study potential mechanisms of the observed ischemic AKI-induced lung apoptosis, we investigated a potential role for the TNFR1 receptor. Analysis of our ischemia-specific and apoptosis-related lung candidate genes following kidney IRI revealed that a large fraction (29/78, 37%) of these genes either belong to the TNFR superfamily (i.e.. Tnfrsf1a, Fas) or are related to the TNFR pathway based on Pubmatrix analysis (Table 3). To validate our gene expression profiling findings of activation of the Tnfrsf1 gene, we measured TNFR1 protein expression by Western blotting in mouse lung tissue homogenates following sham or kidney IRI. TNFR1 protein expression was increased in the lung at 24 h, but not at the earlier 6-h time point (data not shown) following kidney IRI, compared with sham (Fig. 7).

Table 3.

Activated lung apoptosis genes representing the TNFR pathway

Gene Title	Gene Symbol	Fold-Change		PubMatrix
		6 h	36 h	TNFR	ALI
Related to TNFR family
Tnf receptor-associated factor 6	Traf6	1.55*	2.12	251	1
Tumor necrosis factor receptor superfamily, member 1b	Tnfrsf1b	2.41	1.95*	46	0
Tumor necrosis factor receptor superfamily, member 12a	Tnfrsf12a	NS	3.02	3	0
Lymphotoxin β receptor (TNFRS, member 3)	Ltbr	2.50	NS	165	0
Fas (TNF receptor superfamily member)	Fas	1.94	NS	8215	49
Tumor necrosis factor receptor superfamily, member 1a	Tnfrsf1a	1.65	NS	149	1
Related to TNFR pathway
Matrix metallopeptidase 9	Mmp9	2.18	8.58	6	5
Cyclin-dependent kinase inhibitor 1A (P21)	Cdkn1a	2.55	6.93	111	0
Integrin-β2	Itgb2	1.95	3.96	30	37
Son cell proliferation protein	Son	2.09	3.08	19	4
Suppressor of cytokine signaling 3	Socs3	7.39	2.64	14	3
Proviral integration site 1	Pim1	2.05	2.51	5	0
Baculoviral IAP repeat-containing 3	Birc3	1.78	2.20*	9	0
BCL2-like 11 (apoptosis facilitator)	Bcl2l11	NS	2.79	1	0
Insulin-like growth factor binding protein 3	Igfbp3	NS	2.71	13	1
CASP8 and FADD-like apoptosis regulator	Cflar	NS	2.47	343	0
CD14 antigen	Cd14	4.08	NS	202	25
Baculoviral IAP repeat-containing 5	Birc5	2.86	NS	55	0
Myelocytomatosis oncogene	Myc	2.43	NS	182	2
Phosphoinositide-3-kinase, catalytic, γ polypeptide	Pik3cg	2.30	NS	187	12
Myeloid cell leukemia sequence 1	Mcl1	2.13	NS	25	3
BH3-interacting domain death agonist	Bid	2.11	NS	299	2
Receptor-interacting serine/threonine kinase 3	Ripk3	2.05	NS	4	0
Hypoxia-inducible factor 1, α-subunit	Hif1a	1.89	NS	6	6
Caspase-11, apoptosis-related cysteine peptidase	Casp11	1.86	NS	1	0
BCL2-antagonist/killer 1	Bak1	1.82	NS	42	0
CD38 antigen	Cd38	1.78	NS	153	0
Caspase-1	Casp1	1.78	NS	193	5
B-cell leukemia/lymphoma 6	Bcl6	1.77	NS	14	0

All fold-change values correspond to significant FDR <5, unless otherwise stated. TNFR, TNF receptor; ALI, acute lung injury; NS, not significant (FDR >10).

^*Marginal significance (5<FDR<10).

Fig. 7.

Effect of ischemic AKI on lung TNFR1 expression. A: representative Western blot of lung tissue homogenates obtained 24 h following sham or 60-min bilateral kidney ischemia (IRI). TNFR1 (p55) protein expression was increased in IRI-treated mice compared with sham. B: densitometry (normalized to β-actin) confirmed increased TNFR1 in IRI-treated mice; n = ≥3/group. *P < 0.05.

Immunofluorescence microscopy confirmed both IRI-induced increased TNFR1 expression and colocalization with active caspase-3-positive (i.e., apoptotic) endothelial cells in the lung (Fig. 8). To verify the functional relevance of ischemic AKI-induced TNFR1 lung expression, we next measured caspase-3 activity in TNFR1^−/− mice at 36 h following sham or kidney ischemia. Kidney IRI increased lung caspase-3 activity in wild-type, but not TNFR1^−/− mice compared with sham controls, and this was validated by IHC (Fig. 9). In aggregate, these results indicate that kidney IRI induces lung apoptosis primarily via TNFR1-dependent pathways and that these proapoptotic effects are instrumental in the deleterious distant organ effects during ischemic AKI.

Fig. 8.

TNFR1 colocalizes with apoptotic lung endothelial cells during experimental AKI. A: lung immunofluorescence micrographs (×40) obtained 24 h following sham or 60-min bilateral kidney ischemia (IRI) stained with TNFR1 (A and B; red), active caspase-3 (C and D; green), DAPI (E and F; blue), or composite TNFR1-caspase-3-DAPI (G and H). IRI-treated mouse lungs demonstrated increased TNFR1-positive cells, and the majority of caspase-3-positive lung cells colocalized with TNFR1 expression (yellow arrowhead), suggesting TNFR1-associated lung endothelial apoptosis during ischemic AKI.

Fig. 9.

TNFR1-dependent lung apoptosis during experimental AKI. A: lung caspase-3 immunohistochemistry of mouse lung at 24 h following sham or 60-min bilateral kidney ischemia (IRI) in C57BL6/J wild-type (WT) or TNFR^−/− mice. IRI increased lung caspase-3 (brown stain)-positive cells in WT, but not TNFR^−/− mice, compared with sham. B: caspase-3 activity (expressed as relative fluorescent units) was increased following IRI in WT, but not TNFR^−/− mice compared with sham; n = ≥5/group. *P < 0.05.

DISCUSSION

AKI frequently coexists with ALI in critically ill patients, but despite the clear clinical association between these two conditions, little is known about the direct pathophysiological kidney-lung interactions during clinical settings of ischemic AKI. This area has recently become an intense area of interest with the disappointing data that, despite the advent of dialysis, the mortality from native kidney AKI has remained at ∼50%. Thus there is an acute need to explore other approaches to reducing mortality. In the current manuscript, we used a genome-based discovery approach followed by a targeted approach to demonstrate potential mechanisms. In a mouse model of ischemic AKI-induced distant lung dysfunction, we found that 1) AKI induced a marked upregulation of apoptosis genes, 2) AKI induced pulmonary endothelial apoptosis demonstrated by caspase-3 upregulation and TUNEL staining, 3) administration of the antiapoptotic agent Z-VAD-FMK decreased lung apoptosis and endothelial barrier dysfunction, and 4) targeted deletion of the cell death receptor TNFR1 decreased pulmonary endothelial apoptosis and improved barrier dysfunction during AKI.

Initial studies examining the mechanisms underlying AKI-induced distant lung dysfunction have shown that kidney IRI in rats caused increased pulmonary vascular permeability to albumin, which was associated with lung edema and alveolar hemorrhage that peaked at 48 h after initial surgery (28). Interestingly, macrophage inhibition with CNI-1493, but not neutrophil depletion, prevented the increase in pulmonary vascular permeability, suggesting that resident macrophages may be the primary mediators of lung vascular leak under these conditions. Subsequent studies by Deng et al. (10) found that kidney IRI in mice causes an early inflammatory response characterized by rapid activation of the transcription factors NF-κB and AP-1 in the lung. This was followed by expression of lung TNF-α and ICAM-1 and accumulation of pulmonary neutrophils by 4 h postinjury. This involvement of inflammatory cytokines in kidney IRI-induced distant organ dysfunction has been corroborated by two recent studies that have highlighted the critical roles of IL-6 and IL-10 in modulating the distant organ lung effects during experimental AKI (23, 27).

There is increasing evidence that pulmonary cell apoptosis may also play a direct role in the pathophysiology of ALI (24, 38, 32, 33). Apoptosis is a highly regulated mechanism of cell death in which unwanted, infected, or damaged cells are eliminated by activation of specific death-signaling pathways. However, untimely or excessive cellular apoptosis has been implicated in a variety of human diseases, and both enhanced pulmonary endothelial/epithelial cell apoptosis and delayed inflammatory cell apoptosis have been associated with ALI (1, 24, 32, 33) The pulmonary endothelium performs several homeostatic functions including regulation of solute transport between vascular compartments and surrounding tissues, and it functions as a semipermeable barrier dynamically regulated by the cytoskeleton. Imbalances in endothelial barrier function are recognized as critical determinants of the morbidity and mortality associated with disease states characterized by inflammation and increased vascular permeability, and detailed investigations have underscored the importance of the balance between complex tethering forces involved in cell-cell and cell-ECM interactions (4, 13, 16). Endothelial apoptosis leads to disruption of these complex interactions and a potential for loss of endothelial barrier function (4, 9, 39).

While it seems intuitive that increased pulmonary endothelial apoptosis would contribute to the derangements in microvascular permeability, the hallmark of ALI, there have been few studies investigating a direct role for pulmonary cellular apoptosis in the pathophysiology of this condition. Utilizing a candidate gene discovery approach along with disease-oriented bioinformatic analysis of the kidney IRI-induced lung molecular signature, we identified activation of apoptosis-related genes during ischemic AKI. We confirmed this genomic data with physiological end points of lung apoptosis in our model, including an increase in active caspase-3- and TUNEL-positive cells which primarily colocalized with CD34, suggesting endothelial apoptosis. To establish a mechanistic role for lung endothelial apoptosis in our model of extrapulmonary ALI, inhibition of caspase-3 with Z-VAD-FMK significantly reduced AKI's effects on lung apoptosis and injury. Evaluation of the activated apoptosis-specific lung genes identified transcripts related to the TNFR superfamily, and translation of lung TNFR1 was confirmed by Western blotting and immunofluorescence. Finally, kidney IRI failed to induce lung apoptosis in TNFR1^−/− mice compared with wild-type controls, suggesting TNFR1-dependent pulmonary endothelial apoptosis as a mechanism for distant organ dysfunction in this model of AKI-induced lung injury.

Cells generally undergo apoptosis through distinct pathways that can be activated by environmental factors such as hypoxia, oxidants, or physical stress (i.e., mitochondrial intrinsic pathway), or specific receptor-ligand interactions such as Fas/Fas ligand or TNF/TNFR (death receptor-mediated extrinsic pathway). Apoptosis of endothelial cells has been shown to occur under a variety of experimental and clinical conditions including hypoxia, radiation, LPS, ATP, and cytokine administration (3, 9, 15, 42). In addition, BAL fluid from patients with ARDS is cytotoxic to cultured human microvascular endothelial cells, and this response appears to be mediated by TNF-α (21). Our analysis of the mouse lung transcriptome during ischemic AKI indeed identified activation of several genes either belonging to the TNFR (i.e., Tnfrsf1a, Fas) or associated with the TNFR pathway (i.e., Bid, BCl-2). It is not surprising that this pathway plays a fundamental role in kidney IRI-induced distant organ proapoptotic effects since prior studies have shown that kidney ischemia-reperfusion induces both local and circulating TNF-α, and TNFR1 mediates most of the known functions of TNF (5, 12, 44). A recent clinical study in patients with ALI from the ARDS network found that elevated levels of soluble TNFR1 was associated with the development of AKI (30). Meldrum et al. (36) have shown that unilateral rat renal ischemia-reperfusion leads to an increase in both contralateral kidney and transient systemic TNF-α concentrations, and Serteser (43) has shown that renal ischemia-reperfusion induces distant hepatic TNF- α and infiltrating neutrophils in mice. In addition, Kelly (25) has shown that bilateral kidney IRI in mice leads to a sustained increase in systemic and cardiac TNF-α and this was associated with cardiac cellular apoptosis. Our current manuscript is the first to identify that ischemic AKI induces lung endothelial apoptosis and that TNFR1 is a mediator of these distant organ effects. Whether systemic TNF-α, other soluble factors, resident macrophages, or circulating leukocytes mediate this response remains to be determined. Indeed, we have recently measured lung TNF-α levels and found an early increase (613 ± 119 vs. 459 ± 59 pg/ml) in kidney IRI-treated mice at 3 h reperfusion compared with sham, but this did not reach statistical significance, likely secondary to a type 2 error (unpublished observations). Lung TNF-α levels were equal at 24 h (436 ± 94 vs. 499 ± 75).

In summary, utilizing an established mouse model of experimental ischemic AKI and an initial genomic approach to discovery followed by a mechanistic approach to disease targeting, we have identified that kidney IRI activates proapoptotic factors that cause lung endothelial apoptosis, which mediates subsequent microvascular dysfunction. In addition, TNFR1 activation appears to play a fundamental role in orchestrating these distant organ effects. Further mechanistic studies highlighting the AKI-induced imbalances between complex tethering forces involved in endothelial cell-cell and cell-ECM interactions and subsequent lung microvascular dysfunction, as well as the specific role of the TNFR1 pathway in these pathophysiological interactions, may lead to novel therapeutic targets for the deadly distant organ effects of acute kidney injury.

GRANTS

This work was supported by National Institutes of Health Grants P50 HL-073944, P50-HL-084945, K08 HL-089181, R01 DK-54770, and R01-HL-66554.