AJP - Renal Fuel your research with LabChart
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Am J Physiol Renal Physiol 293: F30-F40, 2007. First published February 27, 2007; doi:10.1152/ajprenal.00023.2007
0363-6127/07 $8.00
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
293/1/F30    most recent
00023.2007v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (4)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Hassoun, H. T.
Right arrow Articles by Rabb, H.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Hassoun, H. T.
Right arrow Articles by Rabb, H.

EDITORIAL FOCUS

Ischemic acute kidney injury induces a distant organ functional and genomic response distinguishable from bilateral nephrectomy

Heitham T. Hassoun,1 Dmitry N. Grigoryev,2 Mihaela L. Lie,1 Manchang Liu,2 Chris Cheadle,2 Rubin M. Tuder,2 and Hamid Rabb2

Departments of 1Surgery and 2Medicine, Johns Hopkins University School of Medicine, Baltimore, Maryland

Submitted 15 January 2007 ; accepted in final form 23 February 2007


   ABSTRACT
TOP
 ABSTRACT
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
Acute kidney injury (AKI) is associated with significant mortality, which increases further when combined with acute lung injury. Experiments in rodents have shown that kidney ischemia-reperfusion injury (IRI) facilitates lung injury and inflammation. To identify potential ischemia-specific lung molecular pathways involved, we conducted global gene expression profiling of lung 6 or 36 h following 1) bilateral kidney IRI, 2) bilateral nephrectomy (BNx), and 3) sham laparotomy in C57BL/6J mice. Bronchoalveolar lavage fluid analysis revealed increased total protein, and lung histology revealed increased cellular inflammation following IRI, but not BNx, compared with sham controls. Total RNA from whole lung was isolated and hybridized to 430MOEA (22,626 genes) GeneChips (n = 3/group), which were analyzed by robust multichip average and significance analysis of microarrays and linked to gene ontology (GO) terms using MAPPFinder. The microarray power analysis predicted that the false discovery rate (q < 1%) and ≥50%-fold change compared with sham would represent significant changes in gene expression. Analysis identified 266 and 455 ischemia-specific, AKI-associated lung genes with increased expression and 615 and 204 with decreased expression at 6 and 36 h, respectively, compared with sham controls. Real-time PCR analysis validated select array changes in lung serum amyloid A3 and endothelin-1. GO analysis revealed significant activation (Z > 1.95) of several proinflammatory and proapoptotic biological processes. Ischemic AKI induces functional and transcriptional changes in the lung distinct from those induced by uremia alone. Further investigation using this lung molecular signature induced by kidney IRI will provide mechanistic insights and new therapies for critically ill patients with AKI.

ischemia-reperfusion; lung injury; microarray

ISCHEMIC ACUTE KIDNEY INJURY (AKI) occurs in various clinical settings including shock, sepsis, transplantation, and vascular surgery. Despite advances in renal replacement therapy, the mortality of patients with AKI has remained high over the past few decades, and renal insufficiency continues to be a sensitive marker for a poor outcome in critically ill patients (2, 22). Epidemiological studies have identified an association between AKI and dysfunction of extrarenal organs (4). Given the persistent AKI-associated poor clinical outcomes, there is a need to study the systemic and remote organ effects that occur in the context of AKI.

Acute lung injury (ALI) is a significant cause of respiratory failure in the intensive care unit and carries a substantial mortality of 30–40% (8). When combined with AKI, mortality approaches 80% (26). 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. It is also 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. Previous studies from our laboratory and others have identified that kidney ischemia-reperfusion injury (IRI) in rodents causes increased pulmonary vascular permeability, lung edema, alveolar hemorrhage, increased leukocyte trafficking, and abnormalities of salt and water transporters, all of which are hallmarks of ALI (10, 19, 20, 28, 30). However, the mechanisms of lung dysfunction in the setting of ischemic AKI remain unclear and are confounded by the fact that uremia alone could potentially affect lung function, as well as induce volume overload.

We hypothesized that ischemic AKI, with effects of ischemia-reperfusion plus accumulation of uremic toxins, would lead to a unique genomic response in the lung. Furthermore, we hypothesized that these lung changes would be distinguishable from those induced by uremia alone. We used an established model of ischemic AKI in the mouse and also used a model of bilateral nephrectomy (BNx) to simulate acute uremia without IRI. We found lung functional and structural changes following IRI, but not BNx, compared with sham controls. We then conducted global gene expression profiling of lung tissues in the different groups. Utilizing a candidate gene approach, we identified ischemia-specific changes in the lung transcriptome which are distinguishable from those produced by uremia (i.e., BNx) and which involve several proinflammatory and proapoptotic pathways. Select genes with major expression changes were validated by real-time PCR. Further detailed bioinformatic analysis highlighted several key target biological processes, including early activation of immune and inflammation genes and later activation of ubiquitin- and apoptosis-related genes. This study identifies potential novel diagnostic and therapeutic targets and provides a platform for further detailed investigation into the molecular mechanisms of ischemic AKI-associated distant organ dysfunction.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
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), 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 IRI, a nontraumatic microvascular clamp was applied across both renal pedicles for 60 min. We used 60 min of ischemia after preliminary experiments demonstrated that 30 min of kidney ischemia did not lead to significant changes in lung genomic signature compared with sham surgery alone. After the allotted ischemia time, the clamps were gently removed, the animals were administered 1 ml of sterile saline intraperitoneally, and the incision was closed in two layers with 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 BNx underwent similar procedures except that both renal pedicles were ligated with 5-0 silk suture and the kidneys removed. 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.

Renal function. Blood samples (~0.5 ml) were obtained from each animal at death and centrifuged for 10 min at 8,000 rpm to obtain serum. Serum creatinine (SCr) levels were measured as a marker of renal function, using a 557A Creatinine Kit (Sigma Diagnostics, St. Louis, MO) and analyzed on a Cobas Mira S Plus automated analyzer (Roche Diagnostics, Indianapolis, IN).

Bronchoalveolar lavage. Bronchoalveolar lavage (BAL) fluid analysis was performed as a surrogate measurement of pulmonary vascular permeability as described by Walters et al. (35). 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 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 (Bio-Rad, Hercules, CA) 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.

Lung histology. For assessment of lung morphological injury, hematoxylin and eosin (H&E) staining was performed as previously described (29). After completion of the allotted treatment period, the right main bronchus was isolated and cross-clamped. Via a tracheotomy, the left lung was filled with 0.5% low-melting agarose in 10% formalin at a constant pressure of 25 cmH2O, allowing for homogenous expansion of lung parenchyma. The inflated lungs were then fixed in 10% formalin for 48 h and embedded in paraffin blocks. Sections of the paraffin-embedded tissues (5 µm) were then obtained and stained with H&E (Fisher Scientific, Pittsburgh, PA).

Statistical analysis. For measurements of SCr and BAL protein, 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. P values <0.05 were considered significant.

Power prediction analysis for minimum array requirements. Sample size determination for class comparisons in microarray experiments was employed as previously described (11). This sample size prediction concept is based on variability of the population, the desired detectable differences in gene expression, the power to detect expression differences, and an acceptable error rate. We generated three gene expression profiles (n = 3) for each experimental condition and used 60-min ischemia/6-h recovery period arrays for the array signal variability calculation. The identified median of SDs of log2 expression values of all probe sets ({sigma} = 0.088) was submitted to the microarray sample size-identifying formula proposed by Wei et al. (36), accepting a 1% false discovery rate (significance level {alpha} = 0.01) and imposing 95% of detecting power (1-beta).

The power.t.test function of the R2.3.1 program (www.r-project.org) was implemented: [power.t.test (n = 3, {Delta} = NULL, SD = 0.088, significance level = 0.01, power = 0.95, type = "2.sample," alternative = "2.sided")]. These calculations identified fold-change log2({Delta})µ = 0.54, which corresponds to a numerical 1.454-fold change. Therefore, we concluded that filtering gene expression data using 1.5-fold change and 1% false discovery rate cutoffs would identify significant transcriptional changes in the setting of three animals per experimental condition.

Transcript profiling with Affymetrix oligonucleotide arrays. Total RNA (7–10 µg) was isolated and reverse transcribed (primers oligo-dT coupled to a T7 RNA polymerase binding site). Double-stranded and biotinylated cRNA was synthesized using T7 polymerase. Unincorporated nucleotides were removed, and the cRNA was quantitated. Typically, the yield of cRNA was between 25 and 80 µg. This cRNA is randomly sheared to ~50 nucleotides and hybridized (16 h) to the Affymetrix GeneChips.

The quality of each step was monitored on an Agilent 2100 Bioanalyzer. External standards were included in each hybridization to control for hybridization efficiency, to test for sensitivity, and to assist in the comparison between data sets. These external standards are cRNA transcribed from bacterial genes (bio b, bio c, bio d, cre, thr, and phe). The first hybridization was against a test chip, containing probes against human, mouse, and yeast rnRNAs, and was designed to determine the quality of the cRNA. The hybridized biotinylated cRNA was detected with phycoerythrin-streptavidin and was quantitated by scanning (Agilent GeneArray Scanner). Following positive analysis of the test chip, the same hybridization mixture was added to the expression chip and processed as described. We employed the Affymetrix chip for mouse studies, MG-430A 2.0 (MOE430A), which provides coverage of the transcribed mouse genome on a single array and is composed of 22,626 probe sets representing transcripts and variants from over 14,500 well-characterized mouse genes. During the initial stages of the study, very strict criteria were established for acceptance of RNA and acceptance of data. Total RNA and IVT-generated target RNA was stockpiled at –80°C.

Sample quality control. The stringent quality control of the purity and integrity of the RNA was assessed using standard criteria for RNA quality. The starting material was analyzed with the Agilent 2100 Bioanalyzer (Agilent Technology) that requires nanogram quantities of RNA. All RNA samples that were used for hybridization exhibited intact 28S and 18S ribosomal RNA on denaturing agarose gel electrophoresis. Absorbance readings (260/280 nm) for both total RNAs and biotin-cRNAs fall in the range of 1.8 to 2.1. Only samples with yields of in vitro transcribed RNA >45 µg were hybridized to the chips. After the GeneChips were scanned, they were inspected for possible image artifacts. The average background and noise (raw Q) of all chips used for expression analysis registered within acceptable ranges. The GeneChip eukaryotic hybridization controls (bioB, bioC, bioD, and cre) demonstrated that all samples have similar hybridization efficiencies. 3'/5' Ratios of internal control genes (actin, GAPDH) for all chips were below 3. When global normalization-to-target intensity of 150 was applied, all chips had about equivalent scaling factors at 3–5. Those chips that failed to pass above standards were excluded from further analysis.

Identification of significant transcriptional changes. The fluorescent intensities of microarray hybridization signals were measured by the Agilent Gene Array Scanner and converted to a digitized matrix (CEL files). These data were then analyzed using the robust multichip average approach corrected for probe GC content (GC-RMA) (18), using the Bioconductor (www.bioconductor.org) affy package (14), and hybridization signal intensity values were generated in log2 format. The probe sets with average signal intensity <4 (16 on a numerical scale) were considered as background noise and excluded from further analysis. Expression values of the detectable 13,593 probe sets were scanned for individual low signals, and probe sets with signal intensities <5 (32 on a numerical scale) were modified by increasing the whole number portion to 5 while the decimal number portion remained intact. Data were then normalized for significance analysis of microarrays (SAM) using SAM-recommended raw wise normalization (13). To increase the statistical power of SAM, the probe sets that identified the same transcripts were clustered as described previously (34). The SAM was conducted using 300 permutations of 5 control and 3 IRI or 3 BNx samples. Based on array power prediction, genes with 1.5-fold change and 1% FDR were considered significantly affected by IRI or BNx.

Hierarchical clustering. Fold-change values for individual IRI and BNx samples were derived by subtraction of average control expression (log2 format) from corresponding experimental expressions (log2 format). One hundred significant genes (top 50 of upregulated and top 50 of downregulated genes) from three comparisons (BNx vs. sham control after 36 h, IRI vs. sham control after 6 and 36 h) were combined and filtered for unique genes. The resulting 160 genes were clustered based on a gene expression pattern similarity (Pearson correlation) using MeV software.

Validation of gene expression data. Quantitative real-time RT-PCR validation of selected candidate genes was conducted as previously described (23, 24). Briefly, transcript levels of selected candidate genes in control and kidney IRI-treated lung tissues were measured (n = 3/condition) in 96-well microtiter plates with an ABI Prism 7700 Sequence Detector Systems (PerkinElmer/Applied Biosystems). TaqMan 18S rRNA Control Reagent was used as an internal control for normalization. Primers and probes were purchased from Applied Biosystems in a 20x mixture. All experimental protocols were based on the manufacturer's recommendation using the TaqMan Gold RT-PCR Core Reagents Kit (PerkinElmer/Applied Biosystems, P/N 402876). Experimental parameters were 48°C for 30 min followed by 40 cycles of 95°C for 15 s and 60°C for 1 min. Relative gene expression was calculated using the 2–{Delta}{Delta}Ct method, which generates fold-change values for mRNA transcript levels expressed in IRI-treated lung samples relative to sham-operated samples as described previously (24).

Computational identification of AKI-associated candidate genes. The gene lists generated for IRI and BNx samples were cross-referenced using Microsoft Access, and genes which were unique to IRI effects were considered IRI specific. These genes were used for Gene Ontology analysis by GeneMAPP 2.1 and MAPPFinder 2.0 (6, 12) for identification of biological processes in which IRI-specific genes were implicated. Briefly, the detectable 13,593 probe sets (see above) were scanned for redundancy, and extra probe sets that targeted the same transcript were eliminated. Remaining probe sets were assigned corresponding IRI-induced fold-change values, whereas IRI nonspecific probe sets were assigned zeros. The resulting dataset was processed by GeneMAPP 2.1 considering all nonzero values as involved in corresponding biological pathways and analyzed by MAPPFinder 2.0. Biological processes with a Z score >2 in which at least five IRI-associated genes were involved and represented at least 10% of a given pathway were considered significantly affected by AKI.


   RESULTS
TOP
 ABSTRACT
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 GRANTS
 REFERENCES
 
Renal function following experimental AKI. We measured SCr (mg/dl) at 6, 24, and 36 h following sham, IRI, or BNx in n ≥ 6 mice/group (Fig. 1) to correlate changes in kidney function with changes in lung histology, BAL, and gene expression. Compared with sham, SCr was significantly increased at 24 h (2.7 ± 0.2 vs. 0.3 ± 0.04, P < 0.001) and 36 h (2.8 ± 0.2 vs. 0.6 ± 0.1, P < 0.001) after IRI, but not at 6 h (1.0 ± 0.3 vs. 0.2 ± 0.04). Similarly, SCr was significantly increased at 24 h (3.1 ± 0.3, P < 0.001) and 36 h (3.3 ± 0.2, P < 0.001) after BNx, but not at 6 h (0.9 ± 0.2). Importantly, similar elevations in serum creatinine occurred with 60-min IRI and BNx.


Figure 1
View larger version (11K):
[in this window]
[in a new window]

  		Fig. 1. Effect of experimental acute kidney injury (AKI) on renal function. Serum creatinine (SCr; mg/dl) was measured at 6, 24, and 36 h following sham, kidney ischemia-reperfusion injury (IRI), or bilateral nephrectomy (BNx) in n = ≥6 mice. SCr was significantly increased at 24 and 36 h following IRI or BNx compared with sham. *P < 0.05 by ANOVA.

 
BAL protein following experimental AKI. To identify AKI-induced changes in lung function, we measured BAL protein levels (µg/ml) as a surrogate for pulmonary microvascular dysfunction at 6 and 36 h following sham, IRI, or BNx in n ≥ 6 mice/group (Fig. 2). Compared with sham, BAL protein was significantly increased at 36 h (357 ± 35 vs. 540 ± 71, P < 0.05) after IRI, but not at 6 h (350 ± 37 vs. 455 ± 68, P = 0.19). In contrast, compared with sham, BAL protein was not increased at either 6 h (316 ± 37, P = 0.70) or 36 h (278 ± 53, P = 0.25) following BNx.


Figure 2
View larger version (11K):
[in this window]
[in a new window]

  		Fig. 2. Effect of experimental AKI on lung permeability. Bronchoalveolar lavage (BAL) protein was measured at 6 or 36 h following sham, kidney IRI, or bilateral nephrectomy (BNx). BAL protein was significantly increased at 36 h following IRI. *P < 0.05 by ANOVA.

 
Lung histology following experimental AKI. To assess potential histological changes in the lung following experimental AKI, H&E-stained sections of the lung were analyzed under light microscopy. Representative sections at 6 or 36 h following sham, IRI, or BNx are depicted in Fig. 3. Kidney IRI induced persistent neutrophil infiltration and focal alveolar hemorrhage in the lung periphery; findings that were not present following sham or BNx.


Figure 3
View larger version (45K):
[in this window]
[in a new window]

  		Fig. 3. Effect of experimental AKI on lung histology. Representative hematoxylin and eosin (H&E)-stained sections (x40 magnification) of lung tissue at 6 (A–C) or 36 (D–F) h following sham, kidney IRI, or BNx. IRI-treated mice demonstrated neutrophil infiltration and focal hemorrhages at 36 h (E), which was present but less severe at 6 h (B) and not present at all following sham (A and D) or BNx (C and F).

 
Gene expression changes in the lung following experimental AKI. Global gene expression profiling is a robust tool for identification of diagnostic and mechanism-related candidate genes. To evaluate the lung genomic response to kidney IRI or BNx, we conducted global gene expression profiling of whole lung tissue obtained 6 or 36 h after sham, kidney IRI, or BNx. Total RNA from the lung was isolated from 3–5 separate animals for each group and hybridized to 20 Affymetrix Mouse MOE430A GeneChips. GC-RMA analysis of the 22,626 genes present on each chip identified ~60% genes as expressed in lung tissues on each array.

At 6 h, SAM of expression profiles identified 266 lung genes with increased expression and 615 genes with decreased expression after kidney IRI compared with sham, while the BNx effects on lung transcription were undetectable. At 36 h, analysis identified 600 lung genes with increased expression after kidney IRI compared with sham and 327 genes with decreased expression. Analysis of genomic responses to BNx after 36 h also demonstrated significant expressional changes of 519 genes of which 226 genes were upregulated and 293 downregulated. A detailed list of the significantly affected lung genes following kidney IRI or BNx can be found in the Supplemental Tables 1–3 (all supplemenmtary material can be found in the online version of this study), and GC-RMA array values are provided under accession no. GSE6730 at the National Center for Biotechnology Information (NCBI) Gene Expression Omnibus database at www.ncbi.nlm.nih.gov/projrcts/geo.


View this table:
[in this window]
[in a new window]

  		Table 1. Top differentially expressed lung genes between IRI and BNx

 

View this table:
[in this window]
[in a new window]

  		Table 3. Biological processes activated in the lung during ischemic AKI

 
To investigate potential differences in the lung genomic response to kidney IRI and BNx, cluster analysis of the most significantly affected genes identified groups of genes with similar as well as discordant expression patterns between the two treatment groups at both 6 and 36 h (Fig. 4). The detailed analysis of the most differentially expressed genes between IRI- and BNx-treated animals are presented in Table 1. The two [serum amyloid A3 (Saa3) and endothelin-1 (Edn1)] genes with the most polar expression at 6 h were further validated with real-time PCR (Fig. 5).


Figure 4
View larger version (50K):
[in this window]
[in a new window]

  		Fig. 4. Hierarchical clustering of lung gene expression following kidney IRI or BNx. As described in MATERIALS AND METHODS, 160 lung genes significantly affected by kidney IRI or BNx were combined and clustered using MeV software. Each column represents an experimental condition of the corresponding lung sample. Hierarchical clustering using Pearson correlation identified 5 major clusters (blue triangles), of which 2 clusters demonstrated clear differences in gene expression between IRI and BNx. Genes from these clusters, which represent the most polar expression patterns, are highlighted with yellow rectangles and are listed on the right. Red indicates upregulation, and green indicates downregulation of gene expression relative to corresponding controls, with color intensity corresponding to the fold-change in amplitude (fold-change scale is shown on the left).

 

Figure 5
View larger version (9K):
[in this window]
[in a new window]

  		Fig. 5. Expression of serum amyloid A3 (Saa3) and endothelin-1 (Edn1) genes in lung by Affymetrix microarray and real-time PCR. The relative message abundance of Saa3 and Edn1 was detected by hybridization of total mouse lung RNA obtained from tissues collected 6 or 36 h following kidney IRI or BNx to the MOE430A GeneChip (filled bars) or by real-time RT-PCR (open bars). The relative fold-change identified by both assays was calculated for each IRI sample and is represented as means ± SD. *Significant changes in gene expression between sham (n = 5) and treatment (n = 3) groups by either unpaired t-test (RT-PCR) or significance analysis of microarrays (SAM; microarray), P < 0.05 or q < 1%, respectively.

 
Validation of gene expression profiling findings. To validate the results of our gene expression profiling studies, quantitative real-time PCR measurements of selected individual genes was conducted using the TaqMan assay as described in METHODS. The overall agreement between gene expression by microarray analysis and by the TaqMan assay was good, and the fold-changes for the selected genes (Edn1 and Saa3) are demonstrated in Fig. 5. These genes were chosen for validation based on most representative significantly increased and decreased gene expressions following IRI (Table 1).

Ischemia-specific gene expression changes in the lung following experimental AKI. Once significant differences in lung genomic responses between IRI and BNx were established, we focused on elucidation of potential mechanistic pathways unique to kidney IRI-induced lung structural and functional changes. We conducted subtractive analysis of global transcriptional changes of whole lung tissue detected at 6 or 36 h after kidney IRI or BNx. Genes that were significantly affected by BNx were subtracted from those affected by IRI, thus identifying specific IRI-related genes. Given that BNx genomic effects on the lung after 6 h were undetectable, all genomic changes identified 6 h following bilateral kidney ischemia were therefore considered ischemia related. At 36 h, analysis identified 600 lung genes with increased expression after kidney IRI compared with sham, of which 145 (24%) were also increased after BNx, and 327 genes with decreased expression, of which 123 (38%) were also decreased after BNx. Therefore, 455 upregulated and 204 downregulated ischemia-specific lung candidate genes were identified at 36 h postischemia (Fig. 6). Ischemia-specific, at both 6- and 36-h time points, lung candidate genes with the greatest fold-changes (average FC >4) in expression following kidney IRI are listed in Table 2.


Figure 6
View larger version (13K):
[in this window]
[in a new window]

  		Fig. 6. Detection of up- and downregulated IRI-specific lung candidate genes. Venn diagram depicts gene expression profiles of lung tissues (n = 3–5/group), which identified 266 and 455 IRI-specific genes with increased expression at 6 and 36 h, respectively (A and B) and 615 and 204 IRI-specific genes with decreased expression (C and D).

 

View this table:
[in this window]
[in a new window]

  		Table 2. Ischemia-specific candidate genes with greatest fold-changes

 
Identification of ischemia-specific biological processes involved in ischemic AKI-induced lung dysfunction. To identify relevant biological pathways potentially involved in kidney IRI-induced lung injury, we performed gene ontology analysis of the 266 upregulated ischemia-specific candidate genes identified at 6 h and the 455 upregulated ischemia-specific candidate genes identified at 36 h using MAPPFinder. GO analysis identified predominant activation of immune and inflammatory responses at the early stages of ischemic AKI and ubiquitin cycle and apoptosis pathways at the later time point (Table 3).

To validate the potential biological relevance of candidate genes involved in these bioprocesses, a qualitative analysis of these IRI-specific transcripts was conducted using PubMatrix, the automated literature search engine (http://pubmatrix.grc.nia.nih.gov). The 29 activated genes from the immune and inflammatory bioprocesses in the 6-h GO table and the 39 activated genes from the ubiquitin cycle and apoptosis bioprocesses from the 36-h GO table were matched against the term "acute lung injury" in the PubMed database. The most commonly cited genes related to these terms, as well as fold-changes at 6 or 36 h, are listed in Table 4.


View this table:
[in this window]
[in a new window]

  		Table 4. Activated genes representing GO bioprocesses

 

   DISCUSSION
TOP
 ABSTRACT
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
GRANTS
 REFERENCES
 
AKI is a syndrome of multiple etiologies characterized by a decline in renal function occurring over a short period of time and is associated with unacceptably high morbidity and mortality (33). Numerous recent studies have demonstrated that outcomes of AKI are dependent on the severity of comorbid conditions (2, 4). Isolated AKI has a much better prognosis than AKI associated with multiple organ failure. AKI frequently coexists with ALI in critically ill patients, but despite the clear clinical association between these two conditions, little is known about direct pathophysiological kidney-lung interactions during clinical settings of ischemic AKI. To investigate potential mechanisms involved in ischemic AKI-induced lung dysfunction, we conducted global gene expression profiling of mouse lung tissue obtained 6 or 36 h following sham, kidney IRI, or BNx. In addition, to correlate any potential lung transcriptomic changes with mouse structural and functional kidney and lung injury, we assessed time-dependent changes in serum creatinine, BAL protein levels, and lung morphology under these same experimental conditions. After elucidating a molecular signature of ischemic AKI-induced lung changes, we compared these to BNx and found differences associated with changes in barrier function and histology, despite similar elevations in serum creatinine in the two groups. We selected two of the major genes changing in the lung during AKI and validated our array findings with real-time PCR. We then used a bioinformatic-driven approach to elucidate clinically significant biological processes changing in the lung during ischemic AKI.

Serum creatinine began to rise within 6 h of experimental manipulation, which was significant at 24 and 36 h compared with sham. There was a similar degree of rise in creatinine in the animals with 60-min IRI or BNx compared with sham laparotomy controls, despite visual verification of renal reperfusion following clamp removal at the time of surgery and despite previous findings from our laboratory that 60 min of ischemia does not cause frank organ necrosis on H&E. In addition, this finding of a similar rise in SCr between treatment groups was recently corroborated by Hoke et al. (17). However, only mice treated with kidney IRI developed an increased BAL protein (surrogate for increased lung microvascular permeability) and persistent morphological evidence of lung injury and inflammation. Subsequent gene expression-profiling studies identified early lung transcriptional changes following kidney IRI, but not BNx, and GO analysis revealed involvement of several proinflammatory and proapoptotic biological pathways. Taken together, these findings suggest that while our distinct models of experimental AKI are similar in terms of the timing and severity of induced renal failure, the development of distant lung dysfunction occurred only under conditions involving concomitant organ ischemia-reperfusion.

The observation that visceral organ (i.e., intestinal, hepatic) ischemia-reperfusion induces remote organ injury and inflammation has been well documented under both clinical and experimental conditions (5, 16, 27). The mechanisms of gut reperfusion-induced lung injury, while not completely elucidated, center on 1) infiltration and/or activation of lung neutrophils by circulating soluble factors (i.e., cytokines) induced by damaged cells/tissues (9, 16) and 2) infiltration and/or activation of lung neutrophils by factors derived from the mesenteric lymph (7, 25, 38). However, the effects of kidney IRI on subsequent lung dysfunction have not been studied extensively, and clearly, the potential molecular mechanisms involved remain unknown. Unlike other visceral organs such as the gut and liver, the kidney is generally not considered an immune organ capable of generating a robust proinflammatory response and subsequent multiple organ failure. However, previous studies from our laboratory 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 (20). 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-{kappa}B and AP-1 in the lung. This was followed by expression of lung TNF-{alpha} and ICAM-1 and accumulation of pulmonary neutrophils by 4 h postinjury. The findings of our current study corroborate those findings in mice of early activation of inflammatory genes, but they also reveal persistent lung neutrophil infiltration and evidence of microvascular injury following kidney IRI.

To validate the quantitative and qualitative results of our gene expression-profiling studies, we performed separate measurements of the most significantly activated early inflammatory gene (Saa3) as well as the gene with the lowest early expression level (Edn1) using quantitative real-time PCR. The overall agreement in gene expression between microarray analysis and by the TaqMan assay was very consistent at the early 6-h time point, but less so at the later 36-h time point. For instance, at 36 h, Saa3 gene expression following BNx was significantly increased compared with sham when measured by RT-PCR (4.4 FC), but not when measured by microarray (2.2 FC). In addition, Edn1 gene expression at 36 h following IRI was significantly increased compared with sham when measured by microarray (4.3 FC), but not when measured by RT-PCR (1.3 FC). These findings highlight the need for further validation of candidate genes, both quantitatively and qualitatively (i.e.. protein expression levels), that emerge from discovery tools such as gene expression-profiling with microarrays.

Saa3, an isoform of SaA, is an acute-phase reactant induced in both liver and extrahepatic sites in response to proinflammatory stimuli, and its increased expression in the lung has been observed in animal models of LPS-induced lung injury (32, 37). A direct pathophysiological role in facilitation of ALI under conditions of ischemic AKI remains to be determined but, as demonstrated by its greater than 40-fold increase in early lung gene expression, Saa3 is an attractive candidate gene to be investigated as a potential mediator of AKI-induced ALI in future studies. We found an early significant decrease in Edn1 lung expression following kidney IRI, and therefore found it to be an ideal candidate gene to validate our ability to accurately measure genes whose expression is deactivated following treatment. In addition, Edn-1 is a potent vasoconstrictor and its expression is associated with pulmonary endothelial injury and microvascular leak (15, 21). The deactivation of gene expression at 6 h may represent a response to an increase in early protein abundance or activity, and given that it is also the gene with most links to "acute lung injury" term in the PubMed database, makes Edn1 an attractive candidate gene to be investigated as a potential mediator of AKI-induced ALI in future studies.

To identify relevant biological pathways potentially involved in kidney IRI-induced lung injury, we performed gene ontology analysis of the ischemia-specific candidate genes. GenMAPP is a computer application designed to link gene expression genomic data to biological pathways and groupings of genes. Integrated with GenMAPP are MAPPFinder tools that perform a global analysis of gene expression in the context of hundreds of pathways and thousands of gene ontology terms and facilitate identification of global biological trends in gene expression data. Our results identified early activation of several pathways including genes involved in inflammatory and immune bioprocesses and later activation of proapoptotic pathways. These findings provide a platform for future detailed studies investigating a likely role for the innate immune response in initiating distant organ inflammation following kidney ischemia and may indicate pulmonary endothelial and/or epithelial apoptosis as effector mechanisms of lung microvascular barrier disruption in the setting of ischemic AKI.

Confounding variables in our understanding of ischemic AKI-associated distant organ dysfunction are the relative contributions of uremia, impaired fluid and toxin clearance, volume overload, and increased capillary hydrostatic pressure vs. the effects of renal cellular hypoxia and reoxygenation. The current findings support our hypothesis that the lung genomic changes induced by kidney IRI are distinct from those induced by accumulation of fluid and uremic toxins alone and reveal that global uremia, in the absence of ischemia-reperfusion, is not sufficient to induce distant lung injury. While our study is the first, to our knowledge, comprehensive comparison of global gene expression changes in the lung following kidney IRI or isolated uremia (BNx), other investigators have very recently reported the effects of kidney IRI vs. BNx on lung histological injury and neutrophil accumulation as well as serum cytokine levels (17). This study, in contradistinction to our findings in 8- to 10-wk-old C57BL/6 mice, discovered similar lung injury between animals treated with kidney IRI and those treated with BNx, compared with sham, with findings of septal edema, hemorrhage, and inflammation. Both groups displayed increased serum cytokines compared with sham, albeit distinct profiles. In comparison of the two studies, it is important to note that the duration of bilateral renal ischemia was significantly greater in our study (60 vs. 22 min). Furthermore, we have previously observed that, despite an increase in proinflammatory molecules such as chemokine KC and IL-6 in the lung after 30 min of kidney ischemia, there is no significant changes in lung histology, wet-to-dry ratio, increase in microvascular permeability, or microarray changes compared with sham surgery alone (31). This was the rationale for our focus on the 60-min ischemia time for the current studies. Another important potential variable is the effect of volume expansion on lung functional and genomic changes. As we and others have previously published, rodents lose weight after kidney IRI, most likely from the impaired ability of the kidney to retain salt and water, while nephrectomized animals do not have this acute weight loss despite losing both kidneys. Despite the increase in net body fluid in the nephrectomized animals compared with the IRI animals, their lung functional and genomic changes were minimal compared with sham, thus excluding the effects of fluid retention on the lung changes seen in ischemic AKI.

Although it is well established that AKI is associated with high mortality and distant organ dysfunction, contemporary thinking is that AKI, in fact, induces lung changes which predispose patients to a worse outcome. For instance, renal IRI in mice prone to sickle cell disease markedly worsens pulmonary inflammation compared with the lung changes induced by AKI in wild-type mice (28). However, ischemic preconditioning, both local and distant, is a well-established phenomenon where an initial organ injury protects the organ from a secondary insult. We have found that ischemic AKI in mice leads to alterations in splenocytes so that they are less capable of orchestrating a secondary ischemic organ inflammation and injury on transfer (3). Similarly, kidney-infiltrating lymphocytes 24 h after IRI have reduced ability, compared with sham AKI lymphocytes, to augment secondary ischemic injury (1). Recently, bilateral nephrectomy, but not kidney IRI, was shown to partially attenuate hydrochloric acid/injurious ventilation induced lung injury via decreased neutrophil-induced injury (39). Our findings of a marked difference in lung genomic response in BNx vs. IRI, despite similar increases in serum creatinine, could explain these findings.

In summary, using an established model of ischemic AKI in the mouse and a model of bilateral nephrectomy to simulate acute uremia without IRI, we found lung structural and functional changes following IRI, but not BNx, compared with sham controls. Global gene expression profiling of lung tissues identified ischemia-specific changes in the lung transcriptome which are distinguishable from those produced by uremia alone and which involve several proinflammatory and proapoptotic pathways. Further bioinformatic analysis highlighted several key target biological processes including early activation of immune and inflammation genes and later activation of ubiquitin- and apoptosis-related genes. This study identifies potential novel diagnostic and therapeutic targets and provides a platform for further detailed investigation into the molecular mechanisms of ischemic AKI-associated distant organ dysfunction.


   GRANTS
TOP
 ABSTRACT
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
REFERENCES
 
This work was supported by National Institutes of Health Grants P50-HL-073944 and R01-DK-54770.


   FOOTNOTES
 

Address for reprint requests and other correspondence: H. T. Hassoun, Johns Hopkins Univ., 600 N. Wolfe St./Harvey 611, Baltimore, MD 21287 (e-mail: hhassou1{at}jhmi.edu)

The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


   REFERENCES
TOP
 ABSTRACT
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 

  1. Ascon DB, Lopez-Briones S, Liu M, Ascon M, Savransky V, Colvin RB, Solosky MJ, Rabb H. Phenotypic and functional characterization of kidney-infiltrating lymphocytes in renal ischemia reperfusion injury. J Immunol 177: 3380–3387, 2006.[Abstract/Free Full Text]
  2. Bagshaw SM, Laupland KB, Doig CJ, Mortis G, Fick GH, Mucenski M, Godinez-Luna T, Svenson LW, Rosenal T. Prognosis for long-term survival and renal recovery in critically ill patients with severe acute renal failure: a population-based study. Crit Care 9: R700–R709, 2005.[CrossRef][ISI][Medline]
  3. Burne-Taney MJ, Liu M, Baldwin WM, Racusen L, Rabb H. Decreased capacity of immune cells to cause tissue injury mediates kidney ischemic preconditioning. J Immunol 176: 7015–7020, 2006.[Abstract/Free Full Text]
  4. Chen YC, Fang JT, Tien YC, Chang MY, Huang CC. Organ system failures predict prognosis in critically ill patients with acute renal failure requiring dialysis. Chang Gung Med J 23: 8–13, 2000.[Medline]
  5. Colletti LM, Cortis A, Lukacs N, Kunkel SL, Green M, Strieter RM. Tumor necrosis factor up-regulates intercellular adhesion molecule 1, which is important in the neutrophil-dependent lung and liver injury associated with hepatic ischemia and reperfusion in the rat. Shock 3: 182–191, 1998.
  6. Dahlquist KD, Salomonis N, Vranizan K, Lawlor SC, Conklin BR. GenMAPP, a new tool for viewing and analyzing microarray data on biological pathways. Nat Genet 31: 19–20, 2002.[CrossRef][ISI][Medline]
  7. Davidson MT, Deitch EA, Lu Q, Osband A, Feketeova E, Nemeth ZH, Hasko G, Xu DZ. A study of the biologic activity of trauma-hemorrhagic shock mesenteric lymph over time and the relative role of cytokines. Surgery 136: 32–41, 2004.[CrossRef][ISI][Medline]
  8. Davidson TA, Caldwell ES, Curtis JR, Hudson LD, Steinberg KP. Reduced quality of life in survivors of acute respiratory distress syndrome compared with critically ill control patients. JAMA 281: 354–360, 1999.[Abstract/Free Full Text]
  9. Deitch EA, Xu D, Franko L, Ayala A, Chaudry IH. Evidence favoring the role of the gut as a cytokine-generating organ in rats subjected to hemorrhagic shock. Shock 2: 141–145, 1994.[Medline]
  10. Deng J, Hu X, Yuen PS, Star RA. Alpha-melanocyte-stimulating hormone inhibits lung injury after renal ischemia/reperfusion. Am J Respir Crit Care Med 169: 749–756, 2004.[Abstract/Free Full Text]
  11. Dobbin K, Simon R. Sample size determination in microarray experiments for class comparison and prognostic classification. Biostatistics 6: 27–38, 2005.[Abstract]
  12. Doniger SW, Salomonis N, Dahlquist KD, Vranizan K, Lawlor SC, Conklin BR. MAPPFinder: using Gene Ontology and GenMAPP to create a global gene-expression profile from microarray data. Genome Biol 4: R7, 2003.[CrossRef][Medline]
  13. Grigoryev DN, Ma SF, Irizarry RA, Ye SQ, Quackenbush J, Garcia JG. Orthologous gene-expression profiling in multi-species models: search for candidate genes. Genome Biol 5: R34, 2004.[CrossRef][Medline]
  14. Grigoryev DN, Ma SF, Simon BA, Irizarry RA, Ye SQ, Garcia JG. In vitro identification and in silico utilization of interspecies sequence similarities using GeneChip technology. BMC Genomics 6: 62, 2005.[CrossRef][Medline]
  15. Habre W, Petak F, Ruchonnet-Metrailler I, Donati Y, Tolsa JF, Lele E, Albu G, Beghetti M, Barazzone-Argiroffo C. The role of endothelin-1 in hyperoxia-induced lung injury in mice. Respir Res 27: 7:45, 2006.
  16. Hassoun HT, Kone BC, Mercer DW, Moody FG, Weisbrodt NW, Moore FA. Postinjury multiple organ failure: the role of the gut. Shock 15: 1–10, 2001.[ISI][Medline]
  17. Hoke TS, Douglas IS, Klein CL, He Z, Fang W, Thurman JM, Tao Y, Dursun B, Voelkel NF, Edelstein CL, Faubel S. Acute renal failure after bilateral nephrectomy is associated with cytokine-mediated pulmonary injury. J Am Soc Nephrol 18: 155–164, 2007.[Abstract/Free Full Text]
  18. Irizarry R, Gautier L, Cope L. An R Package for Analyses of Affymetrix Oligonucleotide Arrays. New York: Springer, 2003.
  19. Kelly KJ. Distant effects of experimental renal ischemia/reperfusion injury. J Am Soc Nephrol 14: 1549–1558, 2003.[Abstract/Free Full Text]
  20. Kramer AA, Postler G, Salhab KF, Mendez C, Carey LC, Rabb H. Renal ischemia/reperfusion leads to macrophage-mediated increase in pulmonary vascular permeability. Kidney Int 55: 2362–2367, 1999.[CrossRef][ISI][Medline]
  21. Kuklin VN, Kirov MY, Evgenov OV, Sovershaev MA, Sjoberg J, Kirova SS, Bjertnaes LJ. Novel endothelin receptor antagonist attenuates endotoxin-induced lung injury in sheep. Crit Care Med 32: 766–773, 2004.[CrossRef][ISI][Medline]
  22. Levy EM, Viscoli CM, Horwitz RI. The effect of acute renal failure on mortality. A cohort analysis. JAMA 275: 1489–1494, 1996.[Abstract]
  23. Li J, Grigoryev DN, Ye SQ, Thorne L, Schwartz AR, Smith PL, O'Donnell CP, Polotsky VY. Chronic intermittent hypoxia upregulates genes of lipid biosynthesis in obese mice. J Appl Physiol 99: 1643–1648, 2005.[Abstract/Free Full Text]
  24. Ma SF, Grigoryev DN, Taylor AD, Nonas S, Sammani S, Ye SQ, Garcia JG. Bioinformatic identification of novel early stress response genes in rodent models of lung injury. Am J Physiol Lung Cell Mol Physiol 289: L468–L477, 2005.[Abstract/Free Full Text]
  25. Magnotti LJ, Upperman JS, Xu DZ, Lu Q, Deitch EA. Gut-derived mesenteric lymph but not portal blood increases endothelial cell permeability and promotes lung injury after hemorrhagic shock. Ann Surg 228: 518–527, 1998.[CrossRef][ISI][Medline]
  26. Mehta RL, Pascual MT, Gruta CG, Zhuang S, Chertow GM. Refining predictive models in critically ill patients with acute renal failure. J Am Soc Nephrol 13: 1350–1357, 2002.[Abstract/Free Full Text]
  27. Moore EE, Moore FA, Franciose RJ, Kim FJ, Biffl WL, Banerjee A. The postischemic gut serves as a priming bed for circulating neutrophils that provoke multiple organ failure. J Trauma 37: 881–887, 1994.[ISI][Medline]
  28. Nath KA, Grande JP, Croatt AJ, Frank E, Caplice NM, Hebbel RP, Katusic ZS. Transgenic sickle mice are markedly sensitive to renal ischemia-reperfusion injury. Am J Pathol 166: 963–972, 2005.[Abstract/Free Full Text]
  29. Peng X, Abdilnour RE, Sammani S, Ma SF, Han EJ, Hasan EJ, Tuder R, Garcia JG, Hassoun PM. Inducible nitric oxide synthase contributes to ventilator-induced lung injury. Am J Respir Crit Care Med 172: 470–479, 2005.[Abstract/Free Full Text]
  30. Rabb H, Wang Z, Nemoto T, Hotchkiss J, Yokota N, Soleimani M. Acute renal failure leads to dysregulation of lung salt and water channels. Kidney Int 63: 600–606, 2003.[CrossRef][ISI][Medline]
  31. Savransky V, Molls R, Bevans S, King L, Tuder R, Rabb H. Rapid upregulation of lung keratinocyte chemokine (KC) in ischemic acute renal failure (Abstract). J Am Soc Nephro 16: 624A, 2005.
  32. Sun HC, Qian XM, Nie SN, Wu XH. Serial analysis of gene expression in mice with lipopolysaccharide-induced acute lung injury. Chin J Traumatol 8: 67–73, 2005.[Medline]
  33. Thadhani R, Pascual M, Bonventre JV. Acute renal failure. N Engl J Med 334: 1448–1460, 1996.[Free Full Text]
  34. Tusher V, Tibshirani R, Chu G. Significance analysis of microarrays applied to the ionizing radiation response. Proc Natl Acad Sci USA 98: 5116–5121, 2001.[Abstract/Free Full Text]
  35. Walters DM, Wills-Karp M, Mitzner W. Assessment of cellular profile and lung fuction with repeated bronchoalveolar lavage in individual mice. Physiol Genomics 2: 29–36, 2000.[Abstract/Free Full Text]
  36. Wei C, Li J, Bumgarner RE. Sample size for detecting differentially expressed genes in microarray experiments. BMC Genomics 5: 87, 2004.[CrossRef][Medline]
  37. Wilson TC, Bachurski CJ, Ikegami M, Jobe AH, Kallapur SG. Pulmonary and systemic induction of SAA3 after ventilation and endotoxin in preterm lambs. Pediatr Res 58: 1204–1209, 2005.[CrossRef][ISI][Medline]
  38. Zallen G, Moore EE, Tamura DY, Johnson JL, Biffl WL, Silliman CC. Posthemorrhagic shock mesenteric lymph primes circulating neutrophils and provokes lung injury. J Surg Res 1583: 83–88, 1999.
  39. Zarbock A, Schmolke M, Spieker T, Jurk K, Van Arken H, Singbartl K. Acute uremia but not renal inflammation attenuates aseptic acute lung injury: a critical role for uremic neutrophils. J Am Soc Nephrol 17: 3124–3131, 2006.[Abstract/Free Full Text]



This article has been cited by other articles:


Home page
 J. Am. Soc. Nephrol.Home page
M. Liu, Y. Liang, S. Chigurupati, J. D. Lathia, M. Pletnikov, Z. Sun, M. Crow, C. A. Ross, M. P. Mattson, and H. Rabb
Acute Kidney Injury Leads to Inflammation and Functional Changes in the Brain
J. Am. Soc. Nephrol., July 1, 2008; 19(7): 1360 - 1370.
[Abstract] [Full Text] [PDF]

Home page
 CJASNHome page
J. L. Koyner and P. T. Murray
Mechanical Ventilation and Lung-Kidney Interactions
Clin. J. Am. Soc. Nephrol., March 1, 2008; 3(2): 562 - 570.
[Full Text] [PDF]

Home page
 J. Am. Soc. Nephrol.Home page
D. N. Grigoryev, M. Liu, H. T. Hassoun, C. Cheadle, K. C. Barnes, and H. Rabb
The Local and Systemic Inflammatory Transcriptome after Acute Kidney Injury
J. Am. Soc. Nephrol., March 1, 2008; 19(3): 547 - 558.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
293/1/F30    most recent
00023.2007v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (4)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Hassoun, H. T.
Right arrow Articles by Rabb, H.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Hassoun, H. T.
Right arrow Articles by Rabb, H.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Visit Other APS Journals Online
Copyright © 2007 by the American Physiological Society.