Protection of lungs from hyperoxic injury: gene expression analysis of cyclosporin A therapy
E. Matthew,
L. Kutcher and
J. Dedman
Departments of Molecular and Cellular Physiology and Genome Science University of Cincinnati Medical Center, Cincinnati, Ohio 45267-0576
 |
ABSTRACT
|
---|
We have previously shown that cyclosporin A (CsA), an inhibitor of protein phosphatase 2B (calcineurin), attenuates hyperoxia-induced reductions in murine lung compliance. CsA protected against hyperoxia-induced changes in neutrophil infiltration, capillary congestion, edema, and hyaline membrane formation. Gene expression studies were conducted to identify the gene expression patterns underlying the protective effects of CsA during hyperoxic lung injury. After 72 h of simultaneous treatment with >95% oxygen and CsA (50 mg·kg-1·day-1), RNA was isolated from murine lungs. RNA from treated and untreated lungs was reverse transcribed to cDNA, competitively hybridized, and used to probe 8,734 complimentary DNAs on the Incyte mouse GEM 1 array. Several known genes and expressed sequence tags (ESTs) showed increased (GenBank accession numbers: AA125385, AA241295, W87197, syntaxin, and cyclin G) or decreased [AA036517, AA267567, AA217009, W82577, uteroglobin, stromal cell-derived factor 1, and surfactant protein C (SP-C)] expression after hyperoxia. Hyperoxia-stimulated reductions in SP-C gene expression were confirmed through Northern blot analysis. The increase in gene expression of one expressed sequence tag (AA125385) with hyperoxia was reversed by CsA treatment. Sequence data demonstrated that this EST has high homology to murine cyclin B1. Western blot analysis did not demonstrate any changes in distal lung cyclin B1 expression after hyperoxia. Protein expression of cyclin B1 in the distal lung was observed in the endothelial cells, bronchiolar epithelial cells, and both the type I and type II alveolar epithelial cells. Further analysis of cyclin B1 may elucidate the protective actions of CsA in hyperoxic injury.
DNA microarray; calcineurin; hyperoxia; cyclin B1
 |
INTRODUCTION
|
---|
ALTHOUGH HIGH OXYGEN IS NECESSARY for sustaining life following smoke inhalation injury and other forms of acute lung injury, prolonged oxygen therapy will further the damage to the lung. The alveolar epithelium, which represents a major target of oxidant injury (11) consists of flattened (type I) cells responsible for O2/CO2 exchange and cuboidal (type II) cells that secrete surfactant. Type I cells are highly sensitive to oxidative stress and once damaged, must be replaced by type II cells which have undergone differentiation into type I cells. As a result, the repair of the damaged epithelium is dependent on the ability of the type II stem cells to first proliferate, thereby providing additional cells for transition into type I cells (1, 26). Environmental toxins and high oxygen also target the bronchiolar epithelium (7, 13, 23). The progenitor cells of the bronchiolar epithelium, the nonciliated Clara cells, are thought to play a role in repopulation of bronchiolar epithelium during and after injury (32). Similar to its response to other environmental stresses, the cell responds to high oxygen damage by protecting genomic integrity through exit from the cell cycle or regulation of the cell cycle by changes in expression of the key cell cycle proteins. Alternatively, cells sustaining oxidative DNA damage beyond their repair capacity may undergo apoptosis.
High oxygen damage to the alveolar-capillary barrier is seen in a progression of edema from the perivascular and interstitial regions to the intra-alveolar regions where edema fluid, serum proteins, fibrin, and necrotic cells form the thick hyaline membranes, lining the alveoli. The hyaline membranes are barriers for gas exchange and promote alveolar collapse, which can precede respiratory distress and death. We have previously demonstrated that hyperoxia-exposed mice treated with cyclosporin A (CsA) showed reductions in edema, hyaline membranes, capillary congestion, neutrophil infiltration, and changes in lung compliance (22). We concluded that CsA acted through multiple synergistic pathways influencing various cell types to protect the lung from hyperoxic lung injury.
Inhibition of the Ca2+/calmodulin-dependent phosphatase, calcineurin, accounts for the major systemic, therapeutic, and toxic effects of the immunosuppressant drugs CsA and FK506 (17). Transcription factors are primary targets of calcineurin. The NFAT ("nuclear factor of activated T cells") family responds to increases in intracellular Ca2+ by translocating from the cytosol to the nucleus. This movement is regulated by calcineurin-specific dephosphorylation (37). Calcineurin has also been demonstrated to be a major Elk-1 phosphatase (40, 44), playing a critical role in Elk-1 regulation (40). Elk-1 dephosphorylation is inhibited by CsA but not by okadaic acid in vitro (40). In addition, MEF2-dependent responses are blocked in cells treated with CsA (20).
Nontranscriptional targets of CsA include the mitochondrial permeability pore. Mitochondria play a central role in apoptotic cell death during oxidative stress (2). The deleterious effects of oxidative damage on mitochondria have been demonstrated to be mediated through the opening of the CsA-sensitive and Ca2+-operated nonspecific pore (42).
Microarrays are valuable tools for evaluation of global gene expression when an effective system for eliminating artifactual data is established. Gene array studies may allow for the identification of gene expression patterns underlying the protective effects of CsA during hyperoxic lung injury and thereby enable the development of specific and effective treatment strategies for lung injury. We have designed an effective system for eliminating noise from the results of cDNA microarray analyses. Previous studies (22) point to possible changes in the expression of genes associated with neutrophil infiltration, surfactant protein dynamics, and the integrity of the alveolar capillary barrier. The hyperoxia-induced gene expression changes observed in our system are consistent with models of injury (22), providing valuable new information with regard to pathways of CsA therapy in hyperoxic injury.
 |
MATERIALS AND METHODS
|
---|
Animal treatment.
Twelve female FVB/n mice weighing 2025 g from Taconic Farms (Germantown, NY) were divided evenly into four groups: 1) controls, 2) CsA treated, 3) hyperoxia exposed, and 4) CsA treated and hyperoxia exposed. Hyperoxia groups were placed in Plexiglas chambers flushed with 100% oxygen and maintained at >95% oxygen for 72 h; oxygen concentrations were measured using a mini-Ox analyzer (MSA, Pittsburgh, PA). The control and CsA groups were maintained in room air. The control group and the hyperoxia group were injected intraperitoneally with vehicle consisting of polyethylated castor oil (Cremophor EL, Sigma), dissolved in alcohol and diluted in saline. The CsA treatment groups were injected once a day with 50 mg·kg-1·day-1 of CsA (Sandoz, East Hanover, NJ) for the 3 days of hyperoxia treatment. With the exception of the mice used for immunohistochemistry (3 days of 10 mg·kg-1·day-1 CsA pretreatment), all injections were performed during the 3 days of high oxygen exposure and required removal of the mice from high oxygen for several minutes. Animals were handled under approved IACUC protocols, provided water and food ad libitum, and kept in alternating 12:12-h light/dark cycles.
Probe labeling and hybridization.
Total RNA was isolated from homogenized lung parenchyma using Tri-Reagent (Molecular Research Center; Tissumizer; Tekmar, Cincinnati, OH) according to manufacturers protocol. RNA labeling and hybridization were performed by University of Cincinnati Microarray Core. Details of the experimental protocol are available at http://microarray.uc.edu. Briefly, 20 µg of total RNA was labeled with Cy3 or Cy5 using 2 µg/µl oligo-dT as primer in a reaction buffer consisting of 200 U/ml Superscript II (GIBCO; Life Technologies, Gaithersburg, GA), 1 M DTT, 40 U/ml RNasin, 100 mM dNTP, and 1 mM of either Cy5-dUTP or Cy3-dUTP. Total RNA from each of the experimental groups was labeled with Cy5, and total RNA from untreated controls was labeled with Cy3. Each Cy5-labeled experimental sample was then combined with a Cy3-labeled untreated sample and purified on a TE-30 column, and the resulting mixture was ethanol precipitated. Probe mixtures were resuspended, incubated for 5 min at room temperature, and vortexed. Hybridization buffer (50% formamide, 10x SSC and 0.2% SDS) was preheated at 45°C and added to the probe mixture.
Mouse GEM 1 arrays (Incyte Pharmaceuticals, Palo Alto, CA) were preincubated with hybridization buffer to block nonspecific binding. The probe mixture was applied to the array and covered with a 22-mm2 glass coverslip and placed in a sealed, humidified chamber and incubated at 42°C for 1620 h. Slides were then washed in consecutive solutions of decreasing ionic strength (1x to 0.1x SSC) and scanned at 10-mm resolution with an Axon GenePix 4000A scanner (Axon Instruments, Union City, CA), which simultaneously scans Cy3 and Cy5 channels.
Biological and technical evaluations of array analysis.
To evaluate the technical consistency of the University of Cincinnati Microarray Core Facility microarray system and establish baseline expression changes, RNA isolated from a single mouse was aliquoted. Equal aliquots were labeled with Cy3 or Cy5, mixed, and allowed to compete on the array. Fluorescence intensity values (x- and y-axis; Fig. 1) generated from hybridization to individual DNA spots are representative of the levels of gene expression, and the resultant intensity ratios (open squares, Fig. 1) are derived from the comparisons in gene expression between two samples. After eliminating poor quality spots, and normalizing to correct for differences in dye labeling, relative fluorescence intensity values were obtained. When all of the values were analyzed, 148 (1.69%) genes varied by twofold or greater. When values of 10 or greater were compared, these false positives were reduced to 106 (1.2%). When values of 100 or greater were compared, these false positives were further reduced to 3 of the 8,734 (0.03%) cDNAs spotted on the array. This finding demonstrated that intensity values below 100 are less reliable.

View larger version (18K):
[in this window]
[in a new window]
|
Fig. 1. Evaluation of microarray analysis: scatter plots of gene expression ratios after competitive hybridization. A: comparison of the same lung sample to itself. B: comparison of lung samples from two different mice. C: comparison of a lung sample to a heart sample. The Cy3 and Cy5 fluorescence intensity values are shown on the x-axis and y-axis, respectively. Each open box represents the fluorescence signal intensity ratio of a single gene among the 8,734 on the Incyte GEM 1 array. Intensity values generated from hybridization to individual DNA spots are representative of the levels of gene expression. The resultant intensity ratios are derived from the comparisons in gene expression between two samples. Upregulated genes are represented as red, genes that show no change are presented as yellow, and genes that are downregulated (appear as green on the array slide) are represented as blue.
|
|
Comparisons were made of tissue (two different organs), biological (two different animals, the same organ), and technical (the same animal, the same organ), differences in gene expression through competitive hybridization. After eliminating defective spots (see below), normalizing (to correct for difference in dye labeling), and eliminating genes with fluorescent signal intensities below 100, comparison of gene expression of the lungs of two untreated mice (biological variability) yielded 32 (0.37%) genes that were either overexpressed or underexpressed (Fig. 1, middle). The differences in gene expression between the heart of an untreated mouse and the lung of an untreated mouse were compared (tissue to tissue variability) (Fig. 1, bottom), and 981 (11.2%) genes were observed to be either overexpressed (twofold increase) or underexpressed (twofold decrease).
Data analysis.
Each of the spots on the array were analyzed and given a rating both by the computer software and the University of Cincinnati Microarray Core personnel. The types of spots that are marked for elimination from further analysis include those which demonstrate high local background, poor morphology, saturation, scratches, or have coalesced with neighboring spots. Beginning with background subtraction, GeneSpring software (Silicon Genetics, Redwood City, CA) was used to exclude any known genes or expressed sequence tags (ESTs) that did not show a mean fluorescence of 100 or more from further analysis. Both normalization to the control channel and slide-by-slide normalizations were performed to correct for differences in labeling between the two fluorescent dyes. Dividing all values on a slide by the median fluorescence value for the slide (slide-by-slide normalizations) enables comparisons between slides. The design of the experiment (Fig. 2) generates normalized values (expression ratios).

View larger version (41K):
[in this window]
[in a new window]
|
Fig. 2. Methods: RNA is reverse transcribed to fluorescent-labeled (Cy3 or Cy5) cDNA. Competitive hybridization of Cy3-labeled cDNA (untreated) with Cy5-labeled (experimental) cDNA to the cDNA on the microarray slide allows for determination of relative gene expression. Each of the four experimental groups contained three replicates: three different mice receiving the same treatment, arrayed on different slides. The Cy3 and Cy5 signal are detected, and a direct relationship is established between gene expression and the intensity of the fluorescence for each gene on the array.
|
|
Each of the four different experimental groups (Fig. 2) contained three replicates (three different mice receiving the same treatment and arrayed on three different slides). Based on our finding in Fig. 1, any spot that did not show a mean fluorescence intensity of 100 or more was excluded from analysis. GeneSpring software was used to perform a one-sample t-test to evaluate the reproducibility among replicates when comparing the expression levels of genes in the hyperoxia and hyperoxia/CsA groups with the baseline (Fig. 3). All statistical comparisons for the remaining genes were made by one-way analysis of variance (ANOVA). These analyses were followed up by Student-Newman-Keuls multiple range test to compare the differences between groups. P values below 0.05 were considered significant (Fig. 3).

View larger version (13K):
[in this window]
[in a new window]
|
Fig. 3. Selecting genes of interest. A series of technical and statistical criteria were used to eliminate all artifactual results from microarray analysis of global gene expression.
|
|
DNA sequencing.
Clones that passed all the data analysis criteria (Fig. 3) were purchased in glycerol stock from the University of Cincinnati Microarray Core, which had purchased the clones from Incyte as cDNAs inserted into either the pT3T7 or Bluescript SK- cloning vectors and transformed into Escherichia coli host plasmids. DNA from two colonies per clone were isolated (Plasmid Mini-kit; Qiagen, Valencia, CA) and sequenced (University of Cincinnati DNA Core). Our sequence results were compared with public database information on each clone through BLAST ("basic local alignment search tool"). Stromal cell-derived factor 1 was not sequence-confirmed, due to the lack of availability of viable cultures.
Northern blot analysis.
To confirm the gene expression changes observed through microarray analysis, Northern blot analysis was performed. A surfactant protein C (SP-C; AA222208) DNA probe was labeled with 32P using the Rediprime II random prime labeling system (Amersham Pharmacia Biotech UK, Buckinghamshire, UK). After prehybridization with 1x prehybridization/wash solution (Molecular Research Center) of the nylon membrane at room temperature, the labeled, denatured probe was added to the hybridization solution (Molecular Research Center). Hybridization was performed overnight at 65°C. The membrane was washed 23 times with 1x wash solution for 10 min, then quantified by phosphor imaging (ImageQuant, PhosphorImager; Molecular Dynamics, Sunnyvale, CA).
Western blot analysis.
Distal lung tissue from mice in each of the four experimental groups was frozen at -80°C, weighed, homogenized in sample buffer (1 ml/0.1 g tissue), sonicated, and equally loaded. After resolution on a 12% SDS-polyacrylamide gel and transfer to a nitrocellulose membrane (Schleicher & Schuell, Dassel, Germany), the blots were blocked with 5% nonfat milk and 0.1% Tween and incubated with cyclin B1 antibody (Neomarkers, Freemont, CA) diluted 1:1,000 in blocking solution, washed, and incubated with secondary antibody in blocking solution. After washing, the blots were developed with chemiluminescence detection system (Amersham Pharmacia Biotech, Piscataway, NJ). Protein bands were quantified using macros designed (Louis W. Kutcher, University of Cincinnati) for use with Image-Pro and Excel. Equal loading was confirmed through Ponceau staining, and cyclin B1 band densities were normalized to
-tubulin.
Immunohistochemistry.
To identify the cell types expressing cyclin B1, immunohistochemical analysis was performed using cyclin B1 rabbit polyclonal IgG (Santa Cruz, Santa Cruz, CA). The signal was detected through the Vectastain ABC kit and the peroxidase DAB substrate kit (Vector Laboratories, Burlingame, CA). Immunohistochemistry was performed on lung tissue sections from all four experimental groups, the negative control (normal goat serum; no primary antibody), and the positive control (human breast cancer tissue) simultaneously. All sections were exposed to DAB substrate for the same amount of time. Antigen unmasking was performed with citrate buffer and microwave heating (Dr. Susan Wert, Childrens Hospital Research Foundation, Cincinnati, OH). After low-power examination, randomly selected high-power fields of alveoli were selected to determine the number of cells morphologically consistent with alveolar type II cells that were distinctly cyclin B1 reactive over the total number of alveoli counted. One hundred to four hundred alveoli per experimental condition were counted. Cells that demonstrated distinct staining of greater than 25% of the nucleus were considered positive.
 |
RESULTS
|
---|
Hyperoxia-regulated genes.
An examination of global gene expression changes in distal lung cells in response to hyperoxic injury and CsA therapy and stringent filtering of data resulted in 12 hyperoxia-regulated clones. One of the 12 (AA125385; sequence confirmed to be cyclin B1) demonstrated a reversal of this change in expression with CsA during hyperoxia (Fig. 4). Fluorescence intensity values generated from hybridization to individual DNA spots are representative of the levels of gene expression, and the resultant intensity ratios are derived from the comparisons in gene expression (43) between an experimental sample and the untreated sample (Figs. 2 and 4).

View larger version (26K):
[in this window]
[in a new window]
|
Fig. 4. Hyperoxia-induced changes in distal lung gene expression. Hyperoxia-induced increases (top) and decreases (bottom) in gene expression. Twelve FVB/n mice were evenly divided into four groups: 1) control (C), 2) CsA, 3) hyperoxia (H), and 4) hyperoxia/CsA (H+CsA). Values are means ± SE. Intensity values generated from hybridization to individual DNA spots are representative of the levels of gene expression, and the resultant intensity ratios are derived from the comparisons in gene expression between two samples. RNA from the mice (n = 3) in each of the four experimental groups was labeled with Cy5-dUTP, whereas RNA from one or a pool of six untreated mice was labeled with Cy3-dUTP, for reverse transcription, competitive hybridization, and image acquisition. Changes in expression levels between control and hyperoxia were statistically significant (P 0.05) for the 12 genes of interest.
|
|
SP-C, uteroglobin (also called Clara cell secretory protein or CC10), cyclin B1 (EST AA125386 sequence results), and an EST (W87197) with high homology to glutathione S-transferase have been previously demonstrated to be regulated by hyperoxia or oxidative stress (10, 24, 25, 31, 46). A hyperoxia-stimulated increase in the glutathione S-transferase gene (P = 0.06 not P < 0.05) expression further validated the results obtained with the EST (W87197) with high homology to glutathione S-transferase (not shown). The RNAs, which have not been previously demonstrated to be regulated by hyperoxia included cyclin G, EST AA241295, EST AA036517, EST AA267567, AA217009, W82577, syntaxin (P = 0.05), and stromal cell-derived factor 1 (Fig. 4).
CC10 is an abundant 10-kDa polypeptide synthesized and secreted primarily by nonciliated Clara cells of the bronchiolar epithelium (39). The gene expression ratio of uteroglobin went down significantly with hyperoxic injury, as did the expression of SP-C (Fig. 4, bottom). SP-C is synthesized and secreted exclusively by the type II cells of the alveolar epithelium (14). Other genes that demonstrated significant hyperoxia-induced decreases in gene expression include stromal cell-derived factor 1, EST AA036517, EST AA267567, and cDNAs AA217009 and W82577 (Fig. 4, bottom).
An EST (W87197) with high homology to an enzyme important to the eukaryotic cells response to oxidative stress, glutathione S-transferase, demonstrated an increase in gene expression after hyperoxia exposure. Hyperoxia also stimulated a dramatic increase in mRNA expression of a recently identified cyclin protein, cyclin G (Fig. 4, top). An increase in gene expression of EST AA125385 with hyperoxia and was reversed with CsA therapy. Our sequence data for this EST cDNA showed high homology to murine cyclin B1. Other genes that demonstrated hyperoxia-induced increases in gene expression include syntaxin (P = 0.05) and EST AA241295. EST AA241295 demonstrates high homology to dimethylarginine dimethylaminohydrolase 1 (Ddah 1), which is involved in nitric oxide synthesis. Although Ddah 1 is not directly regulated by hyperoxia, the nitric oxide pathway is regulated by hyperoxia (8, 33). EST AA061216 (P = 0.06), EST AA476157 (P = 0.064), and protein kinase C (P = 0.077) also showed increases in gene expression in response to hyperoxia (not shown).
Northern blot analysis.
An SP-C (AA222208) DNA probe was hybridized to mouse distal lung RNA, revealing the presence of a single transcript of
0.78 kb (Fig. 5). SP-C hybridization band intensities were normalized to the intensities of both the 28S and 18S bands. The normalized intensities for each of the four experimental groups demonstrated a 26% decrease in intensity from 71.5 for the control to 53.7 after hyperoxia. A similar decrease of 21% was observed between the control and the hyperoxia/CsA groups. The CsA alone group was similar to the control group.

View larger version (73K):
[in this window]
[in a new window]
|
Fig. 5. Expression of surfactant protein C (SP-C). Each lane contains 20 µg of total RNA from the mouse distal lung, separated by electrophoresis in formaldehyde agarose gels, and transferred to Hybond N (Amersham). Lane 1, control; lane 2, CsA; lane 3, hyperoxia; and lane 4, hyperoxia/CsA. Northern hybridization was performed with 32P-labeled SP-C isolated from the an Incyte clone (accession no. AA222208).
|
|
Cyclin B1 protein expression.
The amount of active protein in a cell is often not reflected by the level of mRNA expression (3), and cyclin B1 has been demonstrated to be posttranscriptionally regulated (21, 34, 45). Western blot analysis of cyclin B1 demonstrated no change in protein expression after hyperoxia (Fig. 6). Immunohistochemical evaluation of cyclin B1 expression in the distal lung adds valuable cell-specific information to correlate to our microarray studies. The bronchiolar epithelial cell nucleus and cytoplasm demonstrated distinct positive staining. The endothelial cells, alveolar type I cells, and alveolar type II cells also demonstrated positive staining (Fig. 7). Quantification revealed a marked decrease in the number of positive staining type II cells per alveoli, 0.84 for control to 0.36 after hyperoxia and an increase to 0.89 with CsA therapy during hyperoxia. CsA treatment alone produced a modest decrease to 0.59.

View larger version (79K):
[in this window]
[in a new window]
|
Fig. 7. Distal lung immunohistochemical detection of Cyclin B1. Dose of cyclosporin A (CsA) was 10 mg·kg-1·day-1. Representative cyclin B1 immunoreactivity in endothelial cells (right inset), bronchiolar epithelial cells, alveolar type I cells, and alveolar type II cells. Arrows represent cyclin B1-positive cells. Left inset: lung sections without primary antibody. A, alveoli; B, bronchioles; V, vessels.
|
|
 |
DISCUSSION
|
---|
Hyperoxia induced changes in the expression of RNAs for genes encoding proteins (SP-C, glutathione S-transferase, cyclin B1, and CC10) that have been previously demonstrated to play a key role in the pathways involved in or associated with hyperoxic injury or oxidative stress (Fig. 8). Moreover, our microarray analyses allowed for the identification of hyperoxia-induced changes in the expression of RNAs for cyclin G, EST 241295, (high homology to mRNA for Ddah 1), syntaxin, EST AA036517, EST 267567, AA217009, W82577, and stromal cell derived factor 1, which have not previously been demonstated to be regulated by hyperoxia.

View larger version (31K):
[in this window]
[in a new window]
|
Fig. 8. Gene expression changes during hyperoxic lung injury. Hyperoxic lung injury is known to stimulate capillary congestion, increase neutrophil infiltration, edema, and hyaline membrane formation (22). Gene expression changes in major lung proteins (CC10 and SP-C), key enzymes in the eukaryotic cells response to oxidative stress (glutathione S-transferase) and cell cycle checkpoint proteins (cyclin G and cyclin B1) resulted from hyperoxic damage. CsA reversed the cyclin B1-mediated changes in gene expression.
|
|
Consistent with previous studies (25, 46) gene expression of two major lung structural proteins, CC10 and SP-C, was decreased. Previous studies have demonstrated an important role for CC10 in reduction of monocytic and neutrophilic infiltration (16). CC10 (-/-) mice exposed to hyperoxic injury demonstrated rapid lung edema and increased lung expression of pro-inflammatory cytokines IL-3, IL-6, and IL-1ß and in the case of IL-1ß, this was localized to the lung parenchyma. Furthermore, it has been previously demonstrated that the role of CC10 in reducing lung injury and inflammation is not limited to the bronchiolar regions but extends into the alveolar regions of the lung (18). Therefore, reductions in CC10 gene expression may explain the increased neutrophil influx and increased edema that was previously observed after hyperoxic injury (22).
The observed decrease in SP-C mRNA expression after 72 h of >95% oxygen on both the microarray and Northern blot analysis is consistent with previous studies (25). In contrast to SP-B expression and function, SP-C may only play an indirect role in lung function during hyperoxic injury through an overall effect on surfactant protein dynamics. It may also play a role in oxygen radical scavenging. A decrease in surfactant C mRNA expression is therefore consistent with hyperoxia-induced damage to the alveolar-capillary barrier and changes in lung function (22), which leads to respiratory distress and death. The Incyte mouse GEM 1 array does not include cDNAs for surfactant proteins A, B, and D.
We have observed changes in the gene expression of glutathione S-transferase, an important enzyme in the eukaryotic cell response to oxidative stress. The glutathione system is important for reducing hydrogen peroxide (10). During hyperoxic damage, there are numerous sources of hydrogen peroxide. Activated neutrophils produce hydrogen peroxide and other reactive oxygen species during respiratory burst. An increase in glutathione S-transferase is likely an attempt by the body to protect against oxidant-induced damage to the alveolar-capillary barrier.
Any environmental stress has the potential to change the gene expression of key cell cycle checkpoint proteins. An environmental stress such as hyperoxia can cause DNA damage (28) and activate proteins which will prevent both the synthesis of DNA and division of cells with damaged DNA. Induction of apoptosis by DNA damage may also occur to eliminate cells in which damage is beyond repair (11). Cyclin G has been proposed to play a role in cell cycle checkpoint control (27). The expression of cyclin G is induced following DNA damage. The dramatic increase in cyclin G mRNA expression that we observed in our microarray (Fig. 4) may be the result of hyperoxia-induced DNA damage.
Another important cell cycle checkpoint protein is cyclin B1. Cyclin B1 is a known cell cycle marker of entry into M phase of the cell cycle. Although DNA damage activates G2 checkpoint through targeting cyclin B and cdc2 kinase activity, there is no direct evidence of G2 cell cycle checkpoint activation by hyperoxia (28). In G2 phase, cyclin B1 levels rise and cyclin B1 binds to cdc2, enters the nucleus, and promotes entry into M phase. In contrast to the public data bank information on clone AA125385 (suppressor of cytokine signaling 3 gene), our sequencing data from this clone demonstrated high homology to cyclin B1. With the high percentage of error (37.8% for a subset of IMAGE clones) on microarrays (15), it is possible that this gene of interest on our array is not AA125385 but rather cyclin B1. It is because of this high degree of error that we embarked on confirming our genes of interest with sequence data we obtained from the clone stocks.
mRNA and protein analyses must be employed to support microarray studies. The amount of active protein in a cell is often not reflected by the level of mRNA expression (3). The sequence of a gene does not describe posttranslational modifications, which may be essential to protein function and activity. Cyclin B1 has been shown to be posttranscriptionally regulated (21, 34, 45) in various tissues and species. Our microarray analysis of lung cyclin B1 mRNA expression indicates that cyclin B1 may be an important therapeutic target for treatment of hyperoxic lung damage. To test this hypothesis, we must first determine whether the observed changes in cyclin B1 mRNA expression are reflected at the protein expression level. Evaluation of cyclin B1 in the whole distal lung alone obscures the cell-type-specific expression of the protein. Distal lung cyclin B1 expression after hyperoxia may be the composite of increases and decreases in the various cells of the alveolar epithelium, endothelium, and bronchiolar epithelium, as well as the immune cells. All previous studies on the effects of hyperoxia on cell cycle progression were conducted in lung cell lines. After exposure to 95% oxygen, A549 cells, a distal lung epithelial cell line, demonstrated increased growth arrest associated with a decrease in cyclin B1 protein (24). The results from these studies must be confirmed through whole animal studies.
We therefore examined cyclin B1 in the distal lungs of mice exposed to hyperoxia and treated with CsA. In fact, quantification of our immunohistochemistry demonstrates a decrease in type II cell cyclin B1 expression after hyperoxia and a return to normal with CsA treatment during hyperoxia. mRNA and protein levels do not always correlate. Posttranscriptional modifications and differences among distal lung cell types may account for these differences. Modifications that lead to decreased protein turnover despite increased mRNA expression or a lag time before protein is actually expressed may explain these results. The absence of a change in the overall distal lung cyclin B1 expression (Fig. 6) may also be the result of cell-type-specific posttranslational and posttranscriptional regulation of cyclin B1. Type II cells are not only the progenitor cells of the distal lung, but they produce pulmonary surfactant, which is critical for lung function and survival. Cell-specific posttranscriptional and posttranslational modifications of cyclin B1 may be a mechanism by which the lung prevents these cells from entering mitosis after hyperoxia-induced DNA damage. CsA may be valuable in reversing this effect during injury and thereby making more pulmonary secreting type II progenitor cells immediately available for surfactant secretion and replacement of injured type I cells. An increase in cyclin B1 expression in one or more of the other distal lung cell types such as the type I cells, endothelial cells, or inflammatory cells may be obscured by decreases in cyclin B1 expression in type II cell, thereby producing no change in the overall distal lung protein expression of cyclin B1 (Fig. 6) . These correlations may be evaluated through antibody colabeling of distal lung cell types with both cyclin B1 and their cell-specific marker proteins. These studies may be followed by an examination of the changes in the functional activity of cyclin B1 with hyperoxia and CsA treatment during hyperoxia.
Syntaxin demonstrated an increase in gene expression in both hyperoxia and CsA/hyperoxia groups. Syntaxin plays a role in calcium regulated secretion (4). This is consistent with our previous characterization of hyperoxia and CsA-stimulated changes in secretion from lung cells (22). Moreover, the Munc-18-syntaxin complex is thought to be regulated by one of the members of the cdc2 family of cell division kinases (38).
Because of low fluorescence intensity values, lack of reproducibility among replicates, or an insignificant change in gene expression, many genes did not satisfy our selection criteria. Other genes encoding proteins such as heme oxygenase (9, 12), glutathione (36, 41), glutathione peroxidase,
-glutamyltransferase (19), ENaC (35), and NF-
B, as well as numerous apoptotic regulatory proteins (5, 29, 30), have been shown to be regulated by hyperoxia and are not present on the array. Mitochondria play a central role in apoptotic cell death during oxidative stress (2) such as short-term exposure to hyperoxia (5). The mitochondrial permeability pore is a major nontranscriptional target of CsA. The deleterious effects of oxidative damage on mitochondria have been demonstrated to be mediated through the opening of the CsA-sensitive and Ca2+-operated nonspecific pore (42). Our microarray analysis did not reveal changes in expression profiles of mitochondrial genes.
Microarrays, however, have the potential to be powerful hypothesis-generating tools for the evaluation of global gene expression. Specifically, we have discovered hyperoxic regulation of several ESTs and characterized genes. The reversal of the hyperoxia-induced increase of one gene with CsA treatment during hyperoxia enables the future elucidation of the role of this gene in the protection of lungs from hyperoxic injury by CsA. Among the potential 30,000 genes found in the mouse genome, often only sequence-confirmed ESTs are spotted on microarray slides. Potentially critical "unknown" genes not spotted on arrays are excluded. As we have found, gene arrays favor the identity of potentially abundant mRNAs such as SP-C and CC10. The majority of false positives are included in the low-fluorescence-intensity spots. These low-abundance mRNAs may represent channels, inflammatory cytokines, and transcription factors. Isolation of the respective cDNA clones and sequencing, or probing the array with known oligonucleotides, is necessary to confirm identity of individual spotted DNAs.
Because of the myriad of posttranscriptional and posttranslational modifications, global gene expression studies must be followed by an evaluation of protein expression at the cellular and subcellular level. Moreover, global gene expression studies do not provide information regarding the nontranscriptional effects of the drug or treatment. To summarize, our system for eliminating false positives from our cDNA microarray revealed hyperoxia induces changes in gene expression of major lung proteins, a protein involved in homeostasis during oxidative stress, and key cell cycle checkpoint proteins known to be involved in pathways associated with oxidative stress and DNA damage (Fig. 8).
Our future studies will include an examination of functional changes in cyclin B1 activity to elucidate a possible role for cyclin B1 in the therapeutic benefits of CsA therapy during hyperoxic lung injury. Furthermore, the death of cyclin B1 null mice in utero demonstrates that this gene is essential for development (6). Mice with cyclin B1 mutations targeted to various cells will be valuable in answering some of the questions raised by this study. As arrays with the complete murine genes become available, our global gene expression studies can expand. Pathway-specific arrays may also prove valuable both in examining the expression of transcription factors (NFAT, NF-
B, MEF1, and Elk-1) known to be associated with the pathways involved in this model of lung injury and therapy and in developing CsA-based treatment strategies. An examination of posttranslational modification of cyclin B1 may also prove valuable in developing CsA-based therapies for lung injury.
 |
DISCLOSURES
|
---|
This work was supported by Shriners Hospital for Children Grant 01-HDQ-005.
 |
ACKNOWLEDGMENTS
|
---|
We gratefully acknowledge members of the Dedman laboratory (Dr. Marcia Kaetzel, Dr. Bailing Li, Dr. Yong Ji, and Shirelyn Beauman) for valuable technical advice. Directed by Dr. Craig Tomlinson, Saykumar Karyala and Paolo Polintan at the University of Cincinnati Microarray Core performed labeling, hybridization, and imaging of the slides. Dr. Bruce Aronow, Sara Williams, and colleagues in Pediatrics and Bioinformatics (Childrens Hospital Research Foundation, Cincinnati, OH) provided valuable assistance in the use of GeneSpring software. We thank Glenn Doerman for excellence and skill in arranging many of the figures, and we thank Dr. Paul Succop for providing assistance with ANOVA and Student-Newman-Keuls statistical analysis in SAS.
 |
FOOTNOTES
|
---|
Article published online before print. See web site for date of publication (http://physiolgenomics.physiology.org).
Address for reprint requests and other correspondence: E. Matthew, 231 Albert Sabin Way, Univ. of Cincinnati Medical Center, PO Box 670576, Cincinnati, OH 45267-0576 (E-mail: mattheeh{at}email.uc.edu).
10.1152/physiolgenomics.00130.2002.
 |
References
|
---|
- Adamson IYR and Bowden DH. Type 2 cells as progenitor of alveolar epithelium regeneration: cytodynamic study of mice after exposure to oxygen. Lab Invest 30: 1523, 1974.
- Adrain C, Creagh EM, and Martin SJ. Apoptosis-associated release of Smac/DIABLO from mitochondria requires active caspases and is blocked by Bcl-2. EMBO J 20: 66276636, 2001.[Abstract/Free Full Text]
- Anderson L and Seilhammer J. A comparison of selected mRNA and protein abundances in human liver. Electrophoresis 18: 533537, 1997.[ISI][Medline]
- Atlas D. Functional and physical coupling of voltage-sensitive calcium channels with exocytotic proteins: ramifications for the secretion mechanisms. J Neurochem 77: 972985, 2001.[ISI][Medline]
- Budinger GRS, Tso M, McClintock DS, Dean DA, Sznajder JI, and Chandel NS. Hyperoxia-induced apoptosis does not require mitochondrial reactive oxygen species and is regulated by Bcl-2 proteins. J Biol Chem 277: 1565415660, 2002.[Abstract/Free Full Text]
- Brandeis M, Rosewell I, Carrington M, Crompton T, Jacobs MA, Kirk J, Gannon J, and Hunt T. Cyclin B2-null mice develop normally and are fertile whereas cyclin B1-null mice die in utero. Proc Natl Acad Sci USA 95: 43444349, 1998.[Abstract/Free Full Text]
- Brody AR, Liu JY, Brass D, and Corti M. Analyzing the genes and peptide growth factors expressed in lung cells in vivo consequent to asbestos exposure an in vitro. Environ Health Perspect 105: 11651171, 1997.[ISI][Medline]
- Chang L, Ma L, Zhang X, and Chen Y. The role of nitric oxide in hyperoxic lung injury in premature rats. J Tongji Med Univ 21: 7881, 2001.[Medline]
- Choi AM, Sylvester S, Otterbein L, and Holbrook NJ. Molecular responses to hyperoxia in vivo: relationship to increased tolerance in aged rats. Am J Respir Cell Mol Biol 13: 7482, 1995.[Abstract]
- Comhair SA and Erzurum SC. Antioxidant responses to oxidant-mediated lung diseases. Am J Physiol Lung Cell Mol Physiol 283: L246L255, 2002.[Abstract/Free Full Text]
- Clement A, Henrion-Caude A, Besnard V, and Corroyer S. The role of cyclins in epithelial response to oxidants. Am J Respir Crit Care Med 164: S81S84, 2001.[Abstract/Free Full Text]
- Dennery PA, Lee CS, Ford BS, Weng YH, Yang G, and Rodgers PA. Developmental expression of heme oxygenase in the rat lung. Pediatr Res 53: 4247, 2003.[Abstract/Free Full Text]
- Forkert PG. Conjugation of glutathione with the reactive metabolites of 1,1-dichloroethylene in murine lung and liver. Microsc Res Tech 36: 234242, 1997.[ISI][Medline]
- Glasser SW, Korfhagen TR, Wert SE, Bruno MD, McWilliams KM, Vorbroker DK, and Whitsett JA. Genetic element from human surfactant protein SP-C gene confers bronchiolar-alveolar cell specificity in transgenic mice. Am J Physiol Lung Cell Mol Physiol 261: L349L356, 1991.[Abstract/Free Full Text]
- Halgren RG, Fielden MR, Fong CJ, and Zachareweski TR. Assessment of clone identity and sequence fidelity for 1189 IMAGE cDNA clones. Nucleic Acids Res 29: 582588, 2001.[Abstract/Free Full Text]
- Harrod KS, Mounday AD, and Stripp BR. Clara cell secretory protein decreases lung inflammation after acute virus infection. Am J Physiol Lung Cell Mol Physiol 275: L924L930, 1998.[Abstract/Free Full Text]
- Ho S, Clipstone N, Timmerman L, Northrop J, Graff I, Fiorentino D, Nourse J, and Crabtree G. The mechanisms of action of cyclosporin A and FK506. Clin Immunol Immunopathol 80: S40S45, 1996.[ISI][Medline]
- Johnston CJ, Mango GW, Finkelstein JN, and Stripp BR. Altered pulmonary response to hyperoxia in Clara cell secretory protein deficient mice. Am J Respir Cell Mol Biol 17: 147155, 1997.[Abstract/Free Full Text]
- Kim KK, Whitin JC, Sukhova NH, and Cohen HJ. Increase in the extracellular glutathione peroxidase in plasma and lungs of mice exposed to hyperoxia. Pediatr Res 46: 715721, 1999.[Abstract]
- Liu S, Liu P, Borras A, Chatila T, and Speck S. Cyclosporin A-sensitive induction of the Epstein-Barr virus lytic switch is mediated via a novel pathway involving a MEF2 family member. EMBO J 16: 143153, 1997.[Abstract/Free Full Text]
- Maity A, McKenna WG, and Muschel RJ. Evidence for posttranscriptional regulation of cyclin B1 mRNA in cell cycle following irradiation in HeLa cells. EMBO J 14: 603609, 1995.[Abstract]
- Matthew E, Pun R, Simonich M, Iwamoto H, and Dedman J. Cyclosporin A protects lung function from hyperoxic damage. Am J Physiol Lung Cell Mol Physiol 276: L786L795, 1999.[Abstract/Free Full Text]
- Matthew E, Warden G, and Dedman J. A murine model of smoke inhalation. Am J Physiol Lung Cell Mol Physiol 280: L716L723, 2001.[Abstract/Free Full Text]
- McGrath-Morrow SA, and Stahl J. Growth arrest of A549 cells during hyperoxic stress is associated with decreased cyclin B1 and increased p21 (Waf1/Cip1/Sdi1) levels. Biochim Biophys Acta 1538: 9097, 2001.[ISI][Medline]
- Minoo P, King J, and Coalson JJ. Surfactant proteins and lipids are regulated independently during hyperoxia. Am J Physiol Lung Cell Mol Physiol 263: L291L298, 1992.[Abstract/Free Full Text]
- Nabeyrat E, Corroyer S, Besnard V, Cazals-Laville V, Bourbon J, and Clement A. Retinoic acid protects against hyperoxia-mediated cell-cycle arrest of lung alveolar epithelial cells by preserving late G1 cyclin activities. Am J Physiol Lung Cell Mol Physiol 263: L291L298, 1992.[Abstract/Free Full Text]
- Okamoto K and Beach D. A transcriptional target of the p53 tumor suppressor. EMBO J 13: 48164822, 1994.[Abstract]
- OReilly MA. DNA damage and cell cycle checkpoints in hyperoxic lung injury: breaking to facilitate repair. Am J Physiol Lung Cell Mol Physiol 281: L291L305, 2001.[Abstract/Free Full Text]
- OReilly MA, Staversky RJ, Watkins RH, Reed CK, de Mesy Jensen KL, Finkelstein JN, and Keng PC. The cyclin-dependent kinase inhibitor p21 protects lung from oxidative stress. Am J Respir Cell Mol Biol 24: 703710, 2001.[Abstract/Free Full Text]
- OReilly MA, Staversky RJ, Huyck HL, Watkins RH, LoMonaco MB, DAngio CT, Baggs RB, Maniscalco WM, and Pryhuber GS. Bcl-2 family gene expression during hyperoxia induced lung injury. Lab Invest 80: 18451854, 2000.[ISI][Medline]
- OReilly MA, Staversky RJ, Stripp BR, and Finkelstein JN. Exposure to hyperoxia induces p53 expression in mouse lung epithelium. Am J Respir Cell Mol Biol 18: 4350, 1998.[Abstract/Free Full Text]
- Otto WR. Lung epithelial stem cells. J Pathol 197: 527535, 2002.[ISI][Medline]
- Pepperl S, Dorger M, Ringel F, Kupatt C, and Krombach F. Hyperoxia upregulates the NO pathway in alveolar macrophages in vitro: role of AP-1 and NF-
B. Am J Physiol Lung Cell Mol Physiol 280: L905L913, 2001.[Abstract/Free Full Text]
- Pines J and Hunter T. Isolation of a human cyclin cDNA: evidence for cyclin mRNA and protein regulation in the cell cycle and for interaction with p34 cdc2. Cell 58: 833846, 1989.[ISI][Medline]
- Rafii B, Tanswell AK, Otulakowski G, Pitkanen D, Belcastro-Taylor R, and OBrodovich H. O2-induced ENaC expression is associated with NF-
B activation and is blocked by superoxide scavenger. Am J Physiol Lung Cell Mol Physiol 275: L764L770, 1998.[Abstract/Free Full Text]
- Rahman I, Bel A, Mulier B, Donaldson K, and MacNee W. Differential regulation of glutathione oxidants and dexamethasone in alveolar epithelial cells. Am J Physiol Lung Cell Mol Physiol 275: L80L86, 1998.[Abstract/Free Full Text]
- Shaw KT, Ho AM, Raghavan A, Kim J, Jain J, Park J, Sharma S, Rao A, and Hogan PG. Immunosuppressive drugs prevent a rapid dephosphorylation of transcription factor NFAT1 in stimulated immune cells. Proc Natl Acad Sci USA 92: 1120511209, 1995.[Abstract]
- Shuang R, Zhang L, Fletcher A, Groblewski G, Pevsner J, and Stuenkel E. Regulation of Munc-18/Syntaxin 1A interaction by cyclin-dependent kinase 5 in nerve endings. J Biol Chem 273: 49574966, 1998.[Abstract/Free Full Text]
- Singh G and Katyal SL. Clara cells and Clara cell 10kD protein (CC10). Am J Respir Cell Mol Biol 17: 141143, 1997.[Free Full Text]
- Sugimoto TS, Stewart S, and Guan K. The calcium/calmodulin-dependent protein phosphatase calcineurin is the major Elk-1 phosphatase. J Biol Chem 272: 2941529418, 1997.[Abstract/Free Full Text]
- Sullivan SJ, Oberley TD, Roberts RJ, and Spitz DR. A stable O2-resistant cell line: role of lipid peroxidation by products in O2-mediated injury. Am J Physiol Lung Cell Mol Physiol 262: L748L756, 1992.[Abstract/Free Full Text]
- Takeyama N, Matsuo N, and Tanaka T. Oxidative damage to mitochondria is mediated by the Ca2+-dependent inner membrane permeability transition. Biochem J 294: 719725, 1993.[ISI][Medline]
- Taniguchi M, Miura K, Iwao H, and Yamanaka S. Quantitative assessment of DNA microarrays: comparison with northern blot analyses. Genomics 71: 3439, 2001.[ISI][Medline]
- Tian J and Karin M. Stimulation of Elk1 transcriptional activity by mitogen- activated protein kinase is negatively regulated by protein phosphatase 2B (calcineurin). J Biol Chem 274: 1517315180, 1999.[Abstract/Free Full Text]
- Trembley JH, Kren BT, and Steer CJ. Post-transcriptional regulation of cyclin B mRNA expression in regenerating rat liver. Cell Growth Differ 5: 99108, 1994.[Abstract]
- Wikenheiser KA, Wert SE, Wispe JR, Stahlman M, DAmore-Bruno M, Singh G, Katyal SL, and Whitsett JA. Distinct effects of oxygen on surfactant protein B expression in bronchiolar and alveolar epithelium. Am J Physiol Lung Cell Mol Physiol 262: L32L39, 1992.[Abstract/Free Full Text]