Advertisement
JBC

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Originally published In Press as doi:10.1074/jbc.M305788200 on June 24, 2003

J. Biol. Chem., Vol. 278, Issue 41, 40198-40212, October 10, 2003
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
278/41/40198    most recent
M305788200v1
Right arrow Submit a Letter to Editor
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
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 PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowRequest Permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Galindo, C. L.
Right arrow Articles by Chopra, A. K.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Galindo, C. L.
Right arrow Articles by Chopra, A. K.
Social Bookmarking
Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?

Identification of Aeromonas hydrophila Cytotoxic Enterotoxin-induced Genes in Macrophages Using Microarrays*

Cristi L. Galindo {ddagger} §, Jian Sha § , Deborah A. Ribardo, Amin A. Fadl, Lakshmi Pillai and Ashok K. Chopra ||

From the Department of Microbiology and Immunology, University of Texas Medical Branch, Galveston, Texas 77555-1070

Received for publication, June 3, 2003 , and in revised form, June 20, 2003.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 EXPERMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 
A cytotoxic enterotoxin (Act) of Aeromonas hydrophila possesses several biological activities, and it induces an inflammatory response in the host. In this study, we used microarrays to gain a global and molecular view of the cellular transcriptional responses to Act and to identify important genes up-regulated by this toxin. Total RNA was isolated at 0, 2, and 12 h from Act-treated macrophages and applied to Affymetrix MGU74 arrays, and the data were processed using a multi-analysis approach to identify genes that might be critical in the inflammatory process evoked by Act. Seventy-six genes were significantly and consistently up-regulated. Many of these genes were immune-related, and several were transcription factors, adhesion molecules, and cytokines. Additionally, we identified several apoptosis-associated genes that were significantly up-regulated in Act-treated macrophages. Act-induced apoptosis of macrophages was confirmed by annexin V staining and DNA laddering. Quantitative reverse transcriptase-polymerase chain reaction (RT-PCR) and enzyme-linked immunosorbent assay were used to verify increased expression of some inflammatory and apoptosis-associated genes identified by the microarray analysis. To further confirm Act-induced increases in gene expression, real-time RT-PCR was also used for selected genes. Taken together, the array data provided for the first time a global view of Act-mediated signal transduction and clearly demonstrated an inflammatory response and apoptosis mediated by this toxin in host cells at the molecular level.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
EXPERMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 
Aeromonas species (spp.)1 are significant human pathogens that cause both gastrointestinal and non-intestinal diseases in children and adults (1). These bacteria have been isolated from freshwater, salt water, and a variety of foods and are frequently isolated from patients with diarrhea and wound infections. Aeromonas spp. produce an impressive array of virulence factors, including hemolysins, cytotonic and cytotoxic enterotoxins, proteases, lipases, leucocidins, endotoxin, adhesions, and an S-layer (2-4). Our laboratory isolated and molecularly characterized an aerolysin-related cytotoxic enterotoxin (Act) from Aeromonas hydrophila that possesses several biological activities (5). Act causes lysis of red blood cells, is cytotoxic to intestinal and non-intestinal cells, evokes intestinal fluid secretion, and is lethal in nanogram quantities (27 ng) when injected intravenously into mice (2, 5, 6). In subsequent studies, we demonstrated for the first time that Act up-regulated the expression of genes encoding proinflammatory cytokines (tumor necrosis factor-{alpha} (TNF-{alpha}), interleukin-1{beta} (IL-1{beta}), and interleukin-6 (IL-6)) and the inducible nitric-oxide synthase gene in murine macrophages. Act activated arachidonic acid metabolism via the induction of group V phospholipase A2 (PLA2) and cyclooxygenase-2 (COX-2), resulting in the production of prostaglandin E2 (PGE2). Act also up-regulated cellular cAMP concentration, which, along with PGE2, contributed to the enterotoxic response associated with Act (7).

More recently, we demonstrated in macrophages that Act induced the rapid mobilization of calcium from intracellular stores, as well as an influx of calcium from the extracellular medium. The rise in intracellular calcium modulated the production of TNF-{alpha} and PGE2 via activation of NF-{kappa}B (7, 8). In addition to activation of NF-{kappa}B, Act treatment of macrophages resulted in activation of the transcription factor cyclic AMP-response element-binding protein (CREB). Taken together, the data demonstrated a direct role for Act in the induction of the host inflammatory response, which was confirmed in vivo (6).

The induction of key inflammatory mediators seen in Act-treated macrophages might result from involvement of Act in one or more signal transduction pathways (8). Act binds to cholesterol, which could target the toxin to lipid rafts and initiate signaling (5). Our investigation of macrophage stress response factors showed that Act caused activation of kinase cascades (tyrosine kinase and protein kinase A), induced redox stress factors (Ref. 1), and increased reactive oxygen species (ROS) production (8). Interestingly, Act also induced the anti-apoptotic protein Bcl-2 in murine macrophages (7). Act treatment of macrophages causes cell death, which is both dose- and time-dependent, and up-regulation of Bcl-2 suggests that apoptosis may be the mechanism of cell death or an attempt by the cell to avert apoptosis triggered by stress (9, 10). The present study demonstrated that Act indeed evoked apoptosis in macrophages, and we identified potential signaling molecules that could be involved in Act-induced apoptosis.

Although the impetus for signaling is not known, it seems probable that Act directly modulates one or more host cell signaling pathways. To investigate global cellular responses to Act and possibly identify new genes involved in Act-induced signaling, we applied DNA array technology. In an attempt to address the notorious variability commonly seen with data generated from microarrays, we performed the experiments in triplicate with three time points, generating a total of nine arrays. Additionally, we employed several different analysis techniques and expected a high level of consistency to allow a treatment-induced change in gene expression to be considered significant. Our data were quite consistent, with 76 genes identified as significant across all analyses. The microarray data clearly demonstrated a strong host cell inflammatory response to Act, up-regulation of several genes not previously known to be induced by Act, and the strong likelihood of apoptosis as the mechanism of Act-induced cell death.

We verified Act-induced changes in the expression of several genes, including macrophage inflammatory protein-2 (Mip-2), Mip-1{alpha}, RANTES (regulated on activation normal T cell expressed), JunB, TNF receptor-associated factor 1 (TRAF1), T-cell death-associated gene 51 (TDAG51), suppressor of cytokine signaling 3 (SOCS3), and growth arrest and DNA-damage-inducible protein GADD45{beta}, by enzyme-linked immunosorbent assay (ELISA) or real-time RT-PCR. Selection of genes to be verified was based on their known involvement in inflammatory, apoptotic, or stress responses. For example, Mip-2, Mip-1{alpha}, and RANTES are potent chemokines that contribute to the recruitment of circulating monocytes into inflamed tissues (11). SOCS3, on the other hand, is itself activated by cytokines and serves to alter subsequent cytokine signaling, possibly as a negative feedback mechanism to dampen excessive inflammation (12).


   EXPERMENTAL PROCEDURES
TOP
 ABSTRACT
 INTRODUCTION
 EXPERMENTAL PROCEDURES
RESULTS
 DISCUSSION
 REFERENCES
 
Cell Culture—The murine macrophage cell line RAW 264.7 was purchased from the American Type Culture Collection (Manassas, VA). The cells were cultured at 37 °C and 5% CO2 in Dulbecco's minimal essential medium (Invitrogen) containing 4.5 g/liter glucose, 10% fetal bovine serum, 2 mM L-glutamine, and antibiotics penicillin (100 units/ml), and streptomycin (0.1 mg/ml). For each experiment, 5 x 105 cells/ml were plated in 35-mm dishes and allowed to attach overnight. The medium was removed, and fresh medium containing the stimulant (6 ng/ml, unless otherwise stated, lipopolysaccharide (LPS)-free Act) was added (7). After each time point, the supernatant was collected and RNA extracted using the RNAqueous kit (Ambion, Austin, TX) according to the instructions from the manufacturer.

Microarray Analysis—RNA was isolated at 0, 2, and 12 h from Act-treated macrophages, and 20 µg of total RNA was processed for microarray analysis. Briefly, cDNA synthesis, in vitro transcription, and labeling and fragmentation to produce the oligonucleotide probes were performed as instructed by the GeneChip manufacturer (Affymetrix, Santa Clara, CA). The probes were first hybridized to a test array (Affymetrix) and then to the GeneChip murine genome MGU74Av2, both performed using the GeneChip Hybridization Oven 640. The chips were washed in a GeneChip Fluidics Station 400 (Affymetrix), and the results were visualized with a Gene Array scanner using the Affymetrix software. Data were analyzed using MAS 5.0 software (Affymetrix), Significance Analysis of Microarrays (SAM; Stanford University, Stanford, CA), GeneSpring 4.2 (Silicon Genetics, Redwood City, CA), and analysis of variance (ANOVA). Computational hierarchical cluster analysis was performed using Spotfire DecisionSite 7.1 (Spotfire Inc., Somerville, MA), Cluster/Treeview (Eisen Laboratory, University of California, Berkeley, CA),2 CLUSFAVOR 6.0 (Baylor College of Medicine, Houston, TX), and ArrayMiner2 (Optimal Design, Brussels, Belgium). Principal component analysis (PCA) was performed using Spotfire 7.1.

RT-PCR—The Advantage RT-PCR kit (Clontech, Palo Alto, CA) was used to synthesize cDNA from total RNA of Act-induced RAW cells as described by the manufacturer. Briefly, to 1 µg of total RNA was added oligo(dT)18 primer (20 µM) in a total volume of 12.5 µl made up with diethyl pyrocarbonate (DEPC)-treated water. The RNA and primer were heated at 70 °C for 2 min and then quenched on ice immediately. Subsequently, the following reagents were added: 4.0 µl of 5x reaction buffer (Clontech), 1.0 µl of deoxynucleoside triphosphate (dNTP) mixture (10 mM each), 0.5 µl of recombinant RNase inhibitor (40 units/µl), and 1.0 µl of recombinant Moloney murine leukemia virus RT (200 units/µl). The mixture was incubated at 42 °C for 1 h, and the cDNA synthesis was stopped by heating the reaction mixture at 94 °C for 5 min. The reaction mixture was then diluted to a final volume of 100 µl with DEPC-treated water. The Advantage 2 PCR kit (Clontech) was used to amplify the cDNA by PCR as described by the manufacturer. Briefly, the following reagents were added to 5 µl of cDNA sample: 5 µl of 10x PCR buffer, 36 µl of sterile water, 1 µl of dNTP mixture (10 mM each), 1 µl of Advantage Taq DNA polymerase (5 units/µl), 2 µl of premixed human GAPDH amplimer (10 µM (each) primer; for internal control), and 2 µl each of the tested primer pairs (20 µM) for mouse Mip-1{alpha} (fragment size, 154 bp) and mouse RANTES (fragment size, 177 bp) (obtained from BioSource International, Camarillo, CA) or 1 µl each of custom-designed 5' and 3' oligonucleotides for JunB (fragment size, 848 bp), TRAF1 (fragment size, 423 bp), and TDAG51 (fragment size, 349 bp). Custom-designed primers used were as follows: 5'-ACGGAGGGAGAGAAAAGCTC-3' (forward) and 5'-ATGTGGGAGGTAGCTGATGG-3' (reverse) (JunB), 5'-GCACTTGGTGAAGGAGAAGC-3' (forward) and 5'-GTCTTCTTGCCTGAGCCATC-3' (reverse) (TRAF1), and 5'-GAAGGATGCTGGAGAACAGC-3' (forward) and 5'-CGAAAGTCGATCTCTTTGCC-3' (reverse) (TDAG51). PCR was performed by temperature cycling (30 cycles of 94 °C for 45 s, 60 °C for 45 s, and 72 °C for 2 min and a final extension at 72 °C for 7 min). The PCR product was run on a 0.8% agarose gel (for JunB) or a 2% agarose gel and subjected to densitometric scanning, and the -fold increase in the level of a particular cDNA was normalized to the GAPDH product. Each experiment included a negative control in which RNA was omitted from the RT mixture and cDNA was omitted from the PCR.

ELISA—The purified anti-Mip-2 capture antibodies (Pharmingen, San Diego, CA) were diluted to 1-4 µg/ml in binding solution (0.1 M Na2CO3, pH 9.0) and then added to the wells of an enzyme immunoassay high-binding microtiter plate (Corning Costar, Corning, NY). After overnight incubation at 4 °C, the capture antibodies were removed, and any nonspecific binding was blocked by adding 200 µl of blocking buffer (1% bovine serum albumin in phosphate-buffered saline (PBS)) to each well. The plate was incubated at room temperature for 1-2 h, after which wells were washed three times with PBS-Tween 20 (0.05%) buffer. An aliquot (100 µl) of sample or standard was added, and the plate was incubated overnight at 4 °C. After incubation, the plate was washed four times with PBS-Tween, detection antibodies (biotinylated anticytokine; Pharmingen) were diluted to 0.5-2 µg/ml in blocking buffer-Tween and added to the wells, and the plate was incubated for 1 h at room temperature. After the plate was washed four times with PBS-Tween, an enzyme conjugate (streptavidin-conjugated horseradish peroxidase) was diluted to an optimal concentration in blocking buffer-Tween and added to the wells. The plate was incubated at room temperature for 30 min and then washed five times with PBS-Tween. Next, 2,2'-azinobis-(3-ethylbenzthiazoline-6-sulfonic acid) (ABTS) substrate solution (150 mg of ABTS in 0.1 M anhydrous citric acid, adjusted to pH 4.35 with sodium hydroxide) was vortexed, and 100 µl of 3% H2O2 was added to each 11 ml of substrate solution. A 100-µl aliquot was dispensed in each well, and the plate was incubated (5-80 min) for color development. The color reaction was stopped by adding 50 µl of stopping solution (20% SDS, 50% dimethyl formamide), and the optical density was read with a microtiter ELISA plate reader (Molecular Devices Corp., Sunnyvale, CA) at 405 nm.

Real-time RT-PCR—Real-time quantitative RT-PCR was performed on the LightCycler thermal cycler system (Roche Diagnostics, Indianapolis, IN) using SYBR Green I dye (Qiagen, Valencia, CA) as described by the manufacturer. Briefly, 200 ng of RNA was placed into a 20-µl reaction volume containing 1 µg of each primer, 10 µl of SYBR Green PCR master mix, and 0.2 µl of reverse transcriptase. A typical protocol included reverse transcription at 55 °C for 20 min and a denaturation step at 95 °C for 15 min followed by 40 cycles with 95 °C denaturation for 15 s, 55 °C annealing for 20 s, and 72 °C extension for 20 s. The temperature transition rate was set at 20 °C/s. Detection of the fluorescent product was performed at the end of the extension period at 80 °C for 10 s. To confirm amplification specificity, the PCR products were subjected to a melting curve analysis. Negative controls containing water instead of RNA were concomitantly run to confirm that the samples were not cross-contaminated. Targets were normalized to reactions performed using GAPDH amplimers (BIOSOURCE), and -fold change was determined with the LightCycler analysis software, as previously described (14). Custom primers were designed as follows: 5'-AAGCTGACAGGATATCGGCGA-3' (forward) and 5'-AAGATGACTCGGAGGATCTTCGGTGCC-3' (reverse) (JunB), 5'-AAGAGATCGACTTTCGGTGCC-3' (forward) and 5'-TGCTTCTGCCTGGTAGACTTGA-3' (reverse) (TDAG51), 5'-CGAGTTTCAATTTGGTTGCCC-3' (forward) and 5'-CCGATCCTCATCATCTCTCAGG-3' (reverse) (TRAF1), 5'-AAGATTCCGCTGGTACTGAGCC-3' (forward) and 5'-TTCTCATAGGAGTCCAGGTGGC-3' (reverse) (SOCS3), 5'-CCTGCGGAACAGTGAAATGTG-3' (forward) and 5'-AGCGATCTGTCTTGCTCAGCAC-3' (reverse) (GADD45{beta}). These primers would amplify DNA fragments in the size range of 100-250 bp that is optimal for real-time RT-PCR.

Detection of Macrophage Apoptosis by Flow Cytometry Analysis—Variously stimulated macrophages (1 x 106) were stained with annexin V conjugated to fluorescein isothiocyanate (FITC) and propidium iodide (PI) by using a TACS annexin V-FITC staining kit (R&D Systems, Minneapolis, MN) according to the instructions from the manufacturer. Briefly, cells were suspended in 100 µl of annexin V binding buffer and incubated with 1 µl of annexin V and 10 µl of PI for 15 min at room temperature in the dark. Four hundred µl of annexin V binding buffer was then added, and at least 20,000 cells were acquired in a FACSCalibur flow cytometer using the CellQuest 3.0.1 software (Becton Dickinson, Mountain View, CA). Percentages of cells undergoing apoptosis were determined by dual-color analysis. This staining allowed us to distinguish three subsets of cells: viable cells (annexin V-negative and PI-negative), early apoptotic cells (annexin V-positive and PI-negative), and late apoptotic or necrotic cells (annexin V-positive and PI-positive). Immediately after staining, the cells were analyzed on a flow cytometer using 488-nm excitation and a 525-nm bandpass filter for FITC and a 620-nm filter for PI detection.

Analysis of DNA Fragmentation—DNA was extracted from stimulated macrophages using the Suicide-Track DNA Ladder Isolation Kit (Oncogene Research Products, Boston, MA). Monolayer cells and cells in the supernatants were lysed, centrifuged at 1000 x g and processed further as described by the manufacturer. The DNA was ethanol-precipitated, air-dried, and suspended in 50 µl of the kit resuspension buffer. Samples were run on a 1.5% agarose gel and visualized with ethidium bromide staining under UV light.

Gel Shift Analysis—Gel shift assays were performed as previously described (7) with only minor modifications. Briefly, oligonucleotides corresponding to the consensus AP-1 (5'-AGCTTGATGACTCAGCCGGAA-3') were annealed by the supplier (Bio-Synthesis, Inc., Lewisville, TX) and labeled using T4 polynucleotide kinase (Promega, Madison, WI) according to the manufacturer. Next DNA-binding reaction mixtures were assembled. We used unlabeled AP-1 consensus oligonucleotide as a specific competitor, unlabeled AP-2 or SP-1 consensus oligonucleotides as nonspecific competitors (Promega), and nuclear extracts from specific time points after Act treatment of RAW cells. Nuclear extracts were prepared using the NE-PER kit (Pierce) as described by the manufacturer. The reaction mixtures were incubated at room temperature for 10 min, followed by addition of 20,000 cpm of 32P-labeled transcription factor consensus oligonucleotide, and incubation for 20 min at room temperature. Subsequently, 1 µl of gel loading 10x buffer (Promega) was added to each reaction mixture and samples (5-20 µg) were loaded on a nondenaturing 4% polyacrylamide gel. The gel was pre-run in 0.5x Tris-borate-EDTA buffer for 30 min at 100 V before loading the samples. After completion of the run, the gel was transferred to Whatman 3MM paper, dried at 80 °C for 5 h, and exposed to x-ray film overnight to 48 h.

All of the experiments were performed at least in triplicate, and representative data are presented.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 EXPERMENTAL PROCEDURES
 RESULTS
DISCUSSION
 REFERENCES
 
Act Induces Up-regulation of Inflammatory Genes, Cytokines, Chemokines, Adhesion Molecules, and Transcription Factors in Murine Macrophages—RAW 264.7 cells were treated with a sublethal dose of Act (6 ng/ml) for 0, 2, and 12 h, and the RNA was isolated and applied to MGU74Av2 GeneChips. The experiment was performed in triplicate, and these independent experiments generated a total of nine arrays, each representing ~6,000 characterized murine genes and ~6,000 expressed sequence tag clusters. For each experiment (0-, 2-, and 12-h treatments), the genes represented on the three arrays were normalized using the Affymetrix Microarray Suite 5.0 software (MAS 5.0). Approximately 6,000 genes were disregarded in all analyses, except for clustering, as a result of their lack of expression on any of the nine arrays. The data was analyzed separately using four different techniques: MAS 5.0, SAM, GeneSpring 4.2, and ANOVA. A change in gene expression was considered significant if the p value was less than 0.05, the -fold change at least 2.0, and increased gene expression occurred in at least two out of three experiments. A total of six pairwise comparisons were performed (0 versus 2 and 0 versus 12 h for each of the three experiments) using the MAS 5.0 software. Based on these criteria, 157 genes were up-regulated and 35 down-regulated from 0 to 2 h after treatment of macrophages with Act. From 0 to 12 h, 220 genes were up-regulated and 167 were down-regulated (Table I). For comparison, the number of significant genes identified by SAM and GeneSpring 4.2 are also included in Table I.


View this table:
[in this window]
[in a new window]
  		TABLE I
Number of genes significantly up-regulated/down-regulated by Act in macrophages, as determined by using MAS 5.0, SAM, and Silicon Genetics GeneSpring 4.2 softwares

 

SAM software was used to perform two-class unpaired, blocked comparisons for 0 versus 2 and 0 versus 12 h of Act treatment in macrophages. An advantage of SAM is that it generates a table of {delta} values based on the input signal values that attaches a false detection ratio (FDR) to any given number of genes called significant (Table II). We chose {delta} values that captured the largest number of significant genes with the lowest FDRs. For 0- versus 2-h treatments, we chose a {delta} value of 0.38 and -fold change cutoff of 2.0, which yielded 151 significant genes with an FDR of 5.5% (Fig. 1, Table II). For the 0- versus 12-h treatment, SAM generated higher FDRs and yielded fewer genes. For example, a {delta} value of 1.09 yielded only 10 significant genes with an FDR of 19.9% (Table II). For every analysis method used, including clustering programs, gene expression profiles were much more consistent and reproducible for 0 versus 2 h treatments, and more genes were considered significantly changed by Act treatment of macrophages by 2 h than 12 h. The earlier time point (2 h) is most reliable for studying signaling events, which, given the fact that Act is cytotoxic, should prove useful for elucidating the mechanism of action of Act and its direct effect on host cells. We also performed a multi-class, blocked comparison of all 9 arrays using SAM, which generated 19 significant genes with a {delta} value of 1.58 and FDR of 4.7% (Table II). All 19 of these genes were also deemed significant by SAM for the two-class, unpaired blocked comparison of 0- and 2-h treatments, and 18 of them were considered significant by all four analysis methods. Given the fact that the expected number of falsely called genes was ~1 (FDR = 4.7% of 19) and that all but 1 of the 19 genes were also called significant using other analysis methods, we considered SAM to be a useful tool for data mining. The one gene detected by SAM using the multi-class option that was not later included in the most significant gene lists was nuclear autoantigen Sp-100, which was deemed non-significant based solely on the ANOVA value (0.83).


View this table:
[in this window]
[in a new window]
  		TABLE II
Significance table showing values of {delta} generated by significance analysis of microarrays software for Act-treated macrophages at 2 and 12 h

{delta} values calculated by SAM using Student's t test. FDR = (no. of genes falsely called)/(no. of genes called).

 


View larger version (18K):
[in this window]
[in a new window]
  		FIG. 1.
SAM plot for 0 versus 2 h of treatment of macrophages with Act. SAM was used to compare signal values for 0 and 2 h from each experiment (six arrays). Data were analyzed in Excel using the two-class, unpaired, and blocked options with the additional requirement of at least a 2-fold change in gene expression. The {delta} value (0.375) chosen yielded the greatest number of significant genes (151) with the lowest false detection ratio (8.3/151 = 5.5%). See Table II for example output {delta} values generated by SAM. Significantly up-regulated genes are shown in red, and significantly down-regulated genes are shown in green.

 

GeneSpring 4.2 software was used to further normalize the data, perform pairwise comparisons and hierarchical clustering, and to create Venn diagrams. Data were normalized across all genes and all arrays to allow replicate time points to be averaged and combined. In addition, the experiments were analyzed separately using GeneSpring. Pairwise comparisons were performed for each experiment (0- versus 2-h and 0- versus 12-h treatments), Venn diagrams created for each comparison, and genes included in the significant gene list only if they changed significantly for all three experiments. For 0- versus 2-h treatments, 88 genes were significantly up-regulated or down-regulated for all three experiments (Fig. 2A). Following the same trend seen with other analysis methods, only 45 genes were significantly changed for 0- versus 12-h treatments (Fig. 2B).



View larger version (35K):
[in this window]
[in a new window]
  		FIG. 2.
Venn diagrams generated by GeneSpring for 0 versus 2 h and 0 versus 12 h of treatment of macrophages with Act. Venn diagrams were created for 0 versus 2 h and 0 versus 12 h using GeneSpring 4.2. The numbers (added together) in each circle represent the number of genes considered significantly up- or down-regulated at least 2-fold using the treatment comparison function of GeneSpring. A, comparison of 0 versus 2 h of treatment from each experimental replicate. There were 88 genes that were considered significantly altered by Act treatment for all three experiments (center where all circles intersect). B, comparison of 0 versus 12 h of treatment from each experimental replicate. There were 45 genes that were considered significantly altered between 0 and 12 h for all three experiments. G98, G133, and G225 represent the three independent experimental array sets (0, 2, and 12 h).

 

Finally, ANOVA was performed on all genes from the nine arrays, and a change in gene expression was considered significant if the p value was less than 0.05. Genes that were deemed significant by all four analytical techniques (MAS 5.0, SAM, GeneSpring 4.2, and ANOVA) were compiled into lists that included 84 probe sets representing 76 genes for 0- versus 2-h treatments (Table III) and 10 genes for 0- versus 12-h treatments (data not shown). Examples of genes up-regulated from 0-12 h included LIM and SH3 protein 1 (Lasp1), interferon-induced 15-kDa protein, {beta} Fc receptor type II, and CD82 antigen. For all analysis methods, the 0- versus 2-h comparison yielded the greatest number of genes with the highest consistency, and we therefore focused on those 76 genes. The majority of the significant genes were associated with inflammation or stress response (e.g. TNF, granulocyte colony-stimulating factor, Mip, glutaredoxin 1, etc.), which was expected, given the inflammatory response induced by Act (7, 8). There were also several apoptotic genes (e.g. those for Bcl-10, BimEL, GADD45, TDAG51), which were up-regulated, suggesting that apoptosis could be the mechanism whereby Act kills host cells.


View this table:
[in this window]
[in a new window]
  		TABLE III
Most significant genes up-regulated by Act in macrophages from 0 to 2 h as determined by four separate analysis methods (grouped by function)

Genes listed twice were represented on the GeneChips by more than one probe set, and each was determined separately to be significantly up-regulated by Act. Minus sign (-) before the number indicated down-regulation of the gene.

 

Application of Principal Component Analysis to Assess Global Trends in the Data and to Functionally Group Act-induced Genes—To reduce the dimensions of the data and describe the general trend of gene expression changes induced in macrophages by Act treatment, we performed PCA using Spotfire DecisionSite 7.1 software. Three components were sufficient to describe 100% of the variability between treatments, with the three new axes (P1, P2, and P3) accounting for 97.8, 1.2, and 1.0% of the variability, respectively (Fig. 3A). We interpreted the first principal component (P1) as the average expression level of macrophage genes irrespective of treatment. We interpreted P2 to represent those genes, which were up-regulated between 2 and 12 h as a result of Act treatment. We considered the third principal component (P3) to primarily represent early induction of genes by Act treatment.



View larger version (29K):
[in this window]
[in a new window]
  		FIG. 3.
PCA of Act-treated macrophages at 0, 2, and 12 h. PCA was performed using Spotfire DecisionSite 7.1 software. Three components were sufficient to describe all the variation between time points. A, eigenvalue for each component with percentage of variance. Panel shows percentage of variation in data that can be attributed to each component. B, graphic representation of eigenvectors (time point versus component coefficient). Data reduction (from many dimensions down to two dimensions) results in a representation of the data (as shown graphically) that can be interpreted as a description of gene expression changes. The three components (PC1, PC2, and PC3) can be viewed as groups of macrophage genes that behave similarly upon treatment with Act: no change (PC1), late induction (PC2), and early induction (PC3).

 

PCA has previously been shown to be a powerful tool for grouping similarly expressed genes into functional sets (15). We interpreted the PCA results (based on component loadings) to attach biological meaning to the components and organized the genes that were most highly correlated with P2 and P3 into two separate lists of 100 genes each. In other words, we grouped together those genes for which expression patterns most resembled that of P2 and P3 as depicted in Fig. 3B. Most genes highly loaded on component 3 (genes up-regulated by 2 h) were related to inflammation as expected. Likewise, genes highly loaded on component 2 (genes up-regulated between 2 and 12 h) were mainly associated with an immune response, although with fewer cytokines and more genes related to metabolism and protein synthesis and turnover. The 15 most highly correlated genes for each component are listed in Table IV. The PCA results confirmed the data obtained from the other analysis methods used, and many of the genes that correlated with the third component were also included in the most significant gene list (Table III).


View this table:
[in this window]
[in a new window]
  		TABLE IV
Genes from Act-treated macrophages strongly correlated with PCA components 2 and 3

 

Inflammatory Genes Induced by Act Follow a Similar Expression Pattern—To identify groups of genes with similar expression patterns, we employed four separate software programs to perform hierarchical clustering: Cluster/Treeview, CLUSFAVOR 6.0, Spotfire DecisionSite 7.1, and ArrayMiner2. Average -fold change values (obtained using MAS 5.0 and GeneSpring 4.2) for 0 versus 2 h and 2 versus 12 h were clustered using the Cluster/Treeview software program. A cluster representing 53 probe sets (Fig. 4) demonstrated a set of genes that were up-regulated by 2 h (shown in red) and down-regulated between 2 and 12 h (shown in green). Genes for which expression did not change for the indicated time points are represented by black squares.



View larger version (65K):
[in this window]
[in a new window]
  		FIG. 4.
Hierarchical cluster of genes obtained using Cluster/Treeview. The Cluster/Treeview software programs were used to perform hierarchical clustering on -fold change values generated by Affymetrix MAS 5.0 software for 0 versus 2 h and 2 versus 12 h. The cluster shown represents those macrophage genes that were transiently up-regulated by Act at 2 h. Red squares represent genes that were up-regulated, green squares represent down-regulation, and black squares represent no change. Gene names are printed to the right of the cluster. G98, G133, and G225 represent the three independent experimental array sets (0, 2, and 12 h).

 

CLUSFAVOR 6.0, Spotfire DecisionSite 7.1, and ArrayMiner2 were used to cluster normalized intensity values from all nine arrays (Figs. 5, 6, 7). A cluster representing 49 probe sets was obtained using CLUSFAVOR 6.0 (Fig. 5) and exhibited the same pattern as that obtained using Cluster/Treeview (Fig. 4). Gene expression increased between 0 and 2 h (green to red) and fell back nearly to base line between 2 and 12 h (red to black/green). A linear representation of the gene expression pattern for the cluster is shown in Fig. 5B. Similar results were obtained with each of the clustering programs, and several genes consistently clustered together (Figs. 4, 5, 6, 7). Each of the time point replicates clustered together, with the 0- and 2-h time points clustering more closely together than the 12-h time points. There was one major cluster representing genes up-regulated by 2 h, the majority of which were related to immune response, stress, and apoptosis. There were 22 genes that clustered together for each of the four hierarchical clustering programs, all of which were also considered significant by the four analytical techniques: MAS 5.0, SAM, GeneSpring 4.2, and ANOVA (denoted by asterisk in Table III). The similar expression pattern of these genes, as determined by clustering, suggested their involvement in the same signaling pathway or in separate but co-regulated pathways.



View larger version (60K):
[in this window]
[in a new window]
  		FIG. 5.
Hierarchical cluster of genes obtained using CLUSFASVOR 6.0. Hierarchical cluster analysis was performed on signal values from 9 arrays (0-, 2-, and 12-h replicate experiments). A, cluster showing a set of genes that are transiently up-regulated by 2 h in Act-treated macrophages. Low signal values are bright green, high signal values are bright red, and black represents midrange values. B, graphic representation of the cluster shown in A. Time points are listed along the x axis, and relative gene expression is displayed on the y axis. G98, G133, and G225 represent the three independent experimental array sets (0, 2, and 12 h).

 


View larger version (57K):
[in this window]
[in a new window]
  		FIG. 6.
Hierarchical cluster of genes obtained using Spotfire DecisionSite 7.1. Normalization and treatment comparisons were made for all nine arrays using Spotfire software, and hierarchical clustering was performed on genes significantly altered by Act treatment (p value less than 0.05). Higher signal values are shown in red, and lower signal values are shown in green. The circled cluster represents genes that were transiently up-regulated at 2 h (genes listed to the right). G98, G133, and G225 represent the three independent experimental array sets (0, 2, and 12 h).

 


View larger version (101K):
[in this window]
[in a new window]
  		FIG. 7.
Hierarchical cluster of genes obtained using ArrayMiner2. A, hierarchical cluster generated using ArrayMiner2 shows early up-regulation of a subset of genes in Act-treated macrophages. Clustering was performed on signal values from all nine arrays, which are represented by nine columns (three 0-h time points, three 2-h time points, and three 12-h time points (from left to right)). Signal values range from low (bright green) to medium (black) to high (bright red) as shown in the color legend. As seen in the cluster represented, the 2-h time points (middle three columns) show a pattern of high gene expression compared with 0 h (left three columns) or 12 h (right three columns). B, graphic representation of the same cluster, with time points on the abscissa (three 0-h time points, three 2-h time points, and three 12-h time points (from left to right)) and relative gene expression on the coordinate. As shown by the graph, gene expression values increase from 0 to 2 h and fall back nearly to base line by 12 h. C, list of genes found in the cluster.

 

Confirmation of Up-regulated Genes by ELISA and Real-time RT-PCR—To confirm the microarray data, quantitative RT-PCR or ELISA was performed on selected genes or their products. Act-induced up-regulation of genes verified by standard RT-PCR was also confirmed and quantitated by real-time RT-PCR (Table V). Three macrophage inflammatory proteins (Mip-1{alpha}, Mip-1{beta}, and Mip-2) were significantly up-regulated by Act, based on all microarray analyses, and they clustered together for all four clustering programs used (Table III). We therefore verified up-regulation of Mip-1{alpha} by RT-PCR and of Mip-2 by ELISA. As shown in Fig. 8 and Table V, Mip-1{alpha} was up-regulated >500-fold by 2 h but less than 3-fold at 12 h. These results were consistent with the microarray analyses data and clustering results (early, transient up-regulation in response to Act). Likewise, Mip-2 was induced by Act at 8 h (earliest time point tested) with an additional increase in protein levels at 12 h (Fig. 9).


View this table:
[in this window]
[in a new window]
  		TABLE V
Confirmation by real-time RT-PCR of selected genes determined to be up-regulation by Act in macrophages using microarrays

 


View larger version (67K):
[in this window]
[in a new window]
  		FIG. 8.
Confirmation of gene expression changes by quantitative RT-PCR. RT-PCR was conducted using specific primers for glyceraldehyde-3-phosphate dehydrogenase (G3PDH) and Mip-1{alpha} (A), RANTES (B), JunB (C), TRAF1 (D), and TDAG51 (E) before examining the cDNA fragments on 2% agarose or 0.8% agarose (JunB) gels. Lane 1 represents untreated macrophages. Lane 2 represents 2-h Act treatment, and lane 3 represents 12-h Act treatment. Lane 4 is a positive control (glyceraldehyde-3-phosphate dehydrogenase) for RT-PCR.

 


View larger version (19K):
[in this window]
[in a new window]
  		FIG. 9.
Verification of Act-induced up-regulation of Mip-2 by ELISA. A representative ELISA of the supernatant from Act-treated macrophages for the presence of the Mip-2 antigen is shown. As shown in the graph, Mip-2 protein levels were up-regulated by Act at 8 h (earliest time point tested) and were further increased by 12 h.

 

Because most of the Act-induced genes were inflammatory in nature, we also examined up-regulation of the chemokine RANTES by RT-PCR. RANTES was not included in the significant list of genes up-regulated by Act between 0 and 2 h, but this was not surprising because RANTES is known to be transcribed later during the inflammatory process than other related chemokines, such as Mip-1{alpha} and Mip-2 (16, 17). As shown in Fig. 8 and Table V, RANTES was up-regulated 1.8-fold by 2 h and 7.5-fold by 12 h, which is consistent with later gene induction.

Because we were interested in the cell-signaling events induced by Act and toxin-induced cell death, we chose five other genes for verification: those for the AP-1 transcription factor JunB, the anti-apoptotic protein TRAF1, the stress and apoptosis-related TDAG51, SOCS3, and a cDNA with 94% homology to growth arrest and DNA-damage-inducible protein GADD45{beta}. According to the microarray analysis results, JunB was up-regulated by Act 3-5-fold by 2 h. As shown in Fig. 8 and Table V, we verified JunB up-regulation by RT-PCR (9.4-fold induction within 2 h). TRAF1 was up-regulated 25-32-fold by Act, and TDAG51 was up-regulated 31-42-fold, according to the microarray analysis. Up-regulation of both gene products was verified by RT-PCR (11-and 28.4-fold, respectively), which was similar to -fold inductions determined by microarrays (Fig. 8 and Table V). Likewise, Act-induced up-regulation of SOCS3 and the cDNA representing GADD45{beta} was confirmed by real-time RT-PCR, with -fold inductions of 6 and 54.8, respectively (Table V).

Act Induces Apoptosis in Murine Macrophages—Microarray analysis uncovered several genes involved in apoptosis that were induced upon Act treatment. It is known that Act is cytotoxic, and the array findings suggested that the mechanism of cell death could be apoptosis. To test this hypothesis, RAW 264.7 cells were cultured with various doses of Act for 0, 2, 4, 6, 8, and 12 h and assayed for annexin V and propidium iodide staining by flow cytometry. To minimize artifacts that result with adherent cells, we used a non-enzymatic solution to remove the cells from the tissue culture dishes for the assay. As shown in Fig. 10, Act caused apoptosis of macrophages, which was dose- and time-dependent (dose curve data not shown). Annexin staining of Act-treated cells at 2 h (2.9%) was comparable with that of untreated cells (4.3%). Cells treated with Act for 4 h exhibited apoptosis (16.9%), and by 6 h the percentage of apoptosis (28.4%) was nearly that of the cycloheximide-treated positive control (36.3%). Late apoptotic or necrotic cells increased with time, as demonstrated by increases in annexin and propidium iodine staining (Fig. 10, upper right quadrants). We also confirmed the apoptosis data by examining DNA laddering. Cells were treated at 0, 4, 6, 8, 10, and 12 h with 20 ng/ml Act, the DNA was isolated, and ladders were visualized by gel electrophoresis. As seen in Fig. 11, ladders were apparent beginning at 6 h, which confirmed that apoptosis was the method of cell death induced by Act. Ladders were completely absent in untreated cells and declined after 8 h, which was consistent with further DNA degradation that occurred later in apoptosis. Taken together, the results demonstrated apoptosis as the mechanism of cell death induced by Act.



View larger version (63K):
[in this window]
[in a new window]
  		FIG. 10.
Demonstration of Act-induced apoptosis of murine macrophages by annexin V staining and flow cytometry. Act-induced apoptosis was assessed by incubation of Act-treated murine macrophages with annexin V-FITC and propidium iodide and then examination by flow cytometry. A, untreated macrophages were used as a negative control for apoptosis; B, cycloheximide (100 µg/ml)-treated macrophages were used as a positive control for apoptosis. Macrophages were treated with 20 ng/ml Act for 2 h (C), 4 h (D), and 6 h (E). Quadrant 2 represents late apoptotic or necrotic cells, quadrant 3 represents viable cells, and quadrant 4 represents early apoptotic cells. As shown in C, only 2.9% of the cells treated with Act at 2 h were apoptotic (comparable to the negative control). By 4 h, Act induced 16.9% apoptosis (D), and by 6 h, 28.4% (E), of the cells were apoptotic (comparable to the positive control).

 


View larger version (111K):
[in this window]
[in a new window]
  		FIG. 11.
Demonstration of Act-induced apoptosis of murine macrophages by DNA laddering. DNA in lane 1 was collected from HL60 cells treated with 0.5 µg/ml actinomycin D as a positive control (PC). DNA was isolated from macrophages treated with 20 ng/ml Act at 0, 4, 6, 8, 10, and 12 h and loaded into lanes 2-7. The samples were subjected to electrophoresis using 1.5% agarose. No DNA laddering was observed for the negative control (lane 2) as expected, and DNA laddering of Act-treated macrophages was first apparent at 6 h (lane 4), which was comparable to the positive control (lane 1).

 

Act Causes Activation of the AP-1 Transcription Factor in Murine Macrophages—According to microarray analyses, AP-1 transcription factor subunits JunB and Fos-like antigen 1 (Fra-1) were up-regulated in Act-treated macrophages. Furthermore, Act has been shown to induce the production of TNF-{alpha}, which is known to activate the transcription factor AP-1 in macrophages (18). We therefore performed gel shift assays to detect AP-1 activation in these cells. Fig. 12 clearly demonstrates that Act caused nuclear translocation of a protein capable of binding a radiolabeled AP-1 binding sequence. The binding was strongly evident at 4 h but only weakly detectable by 8 h. We performed competition assays using unlabeled AP-1 (specific) and SP-1 or AP-2 (nonspecific) consensus oligonucleotides. As shown in Fig. 12, addition of the specific competitor completely abrogated binding, whereas the nonspecific competitor did not (data not shown), indicating that the band observed was indeed specific for AP-1.



View larger version (82K):
[in this window]
[in a new window]
  		FIG. 12.
Demonstration of AP-1 activation by gel shift analysis. Act activated the AP-1 transcription factor in murine macrophages, as determined by gel shift assay. Nuclear extracts from Act-treated RAW cells (0, 10, and 30 min and 2, 4, 8, and 24 h) were mixed with consensus oligonucleotides for AP-1 and subjected to nondenaturing 4% polyacrylamide gel electrophoresis (lanes 1-7). Nuclear extract from RAW cells treated for 4 h with Act was mixed with unlabeled AP-1 consensus oligonucleotide (~50-fold excess) before adding the labeled oligonucleotide (lane 8). The gel was dried and subjected to autoradiography.

 


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 EXPERMENTAL PROCEDURES
 RESULTS
 DISCUSSION
REFERENCES
 
The data obtained from the microarray analysis were consistent and reproducible. Several of the genes determined to be significantly up-regulated according to the microarray analyses were previously determined by traditional laboratory techniques to be induced by Act. For instance, up-regulation of TNF-{alpha} by Act in macrophages was demonstrated by Northern blot analysis and ELISA (7). Additionally, induction of IL-1{beta} by Act, which was deemed significant by three out of four analysis methods, was previously reported by Northern blot analysis (7). Activation of NF-{kappa}B was previously demonstrated by gel shift assay, and both the p105 and p49/p100 subunits were significantly up-regulated according to microarray data analysis (Table III). Inhibitor of NF-{kappa}B, I-{kappa}B{alpha}, was also up-regulated by Act according to all analyses and clustering techniques employed. This was expected, because NF-{kappa}B negatively regulates itself by transcriptional activation of I-{kappa}B{alpha} (19). Up-regulation of adhesion factors, growth factors, and immune response molecules is consistent with the idea that Act induces activation of macrophages leading to an acute inflammatory response.

Several apoptosis-related genes were up-regulated in murine macrophages by Act treatment (Table III), which prompted us to confirm apoptosis as the mechanism of cell death. Whether Act directly induces apoptosis or causes it secondary to cell stress awaits further study, but the array results strongly suggested the involvement of TNF-{alpha}. In macrophages, binding of TNF-{alpha} to its receptor activates the transcription factors NF-{kappa}B and AP-1, leading to induction of proinflammatory and immunomodulatory genes (18). TNF-{alpha} rarely triggers apoptosis unless protein synthesis is blocked, most likely because of the presence of survival factors regulated by NF-{kappa}B and c-Jun NH2-terminal kinase (JNK)/AP-1 that can suppress the apoptotic stimulus generated by TNF-{alpha} (20). For example, NF-{kappa}B induces the expression of genes that promote cell survival such as those coding for TNF receptor-associated factors 1 and 2 (TRAF1, TRAF2) and the cellular inhibitors of apoptosis 1 and 2 (c-IAP1, c-IAP2) (21). However, inhibition of either the NF-{kappa}B or JNK/AP-1 pathways sensitizes cells to apoptosis induction by TNF-{alpha} (22). The microarray results indicated that Act induced transcription of NF-{kappa}B (Table III), and previous studies by our laboratory demonstrated activation of NF-{kappa}B by gel shift assay (7, 8). Conversely, our laboratory previously demonstrated an absence of AP-1 (c-Jun/c-Fos) activation in Act-treated macrophages (7), and the c-Jun transcript was absent according to the microarray analyses. Additionally, c-Fos gene expression was possibly down-regulated by Act treatment (significant for only two out of four analytical methods and, therefore, not included in Table III). In this study, we were able to demonstrate activation of AP-1, but only at 4 h (Fig. 12), which preceded Act-induced apoptosis (Figs. 10 and 11). One explanation is that Act induces inhibition of TNF-mediated AP-1 (c-Jun/c-Fos) activity by action of the pro-apoptotic protein Bcl-10, the expression of which was up-regulated by Act treatment (Table III). Bcl-10 has been shown to bind TRAF2, competitively inhibit TRAF2 interaction with c-IAP, and attenuate the JNK signaling pathway activated by TNF stimulation (23). Alternatively, Act-induced apoptosis of macrophages may result from a change in the AP-1 composition from c-Jun- to JunB-containing complexes. Microarray analysis revealed that Act up-regulated expression of Fra-1 and JunB, the latter of which was confirmed by RT-PCR. Further studies are required to elucidate the significance of JunB up-regulation by Act and subsequent effects on transcription regulation.

Expressions of the anti-apoptotic protein TRAF1 and stress and apoptosis-related TDAG51 were up-regulated in Act-treated macrophages, both of which were subsequently verified by RT-PCR. TRAF1 works in conjunction with TRAF2 and c-IAPs to fully suppress TNF-induced apoptosis, which may be achieved via recruitment of c-IAPs that suppress caspase activation (21, 24). It is unclear why TRAF1 was up-regulated in Act-treated macrophages, given the fact that toxin exposure leads to cell death. The most likely explanation is that TRAF1 inhibition of apoptosis is perturbed by some other cellular factor. Alternatively, Act-induced apoptosis may occur via a pathway that does not involve TRAF1. An argument for the latter hypothesis is the up-regulation of TDAG51 by Act. Originally, TDAG51 was suggested to be required for T cell receptor-dependent induction of Fas/Apo1/CD95 expression in murine T cell hybridomas (26), but more recently TDAG51 has been shown to cause apoptosis via inhibition of protein synthesis (27). It is therefore possible that TDAG51 contributes to Act-induced apoptosis of macrophages.

Three myeloid differentiation (MyD) primary response and growth arrest DNA-damage (GADD) genes were up-regulated by Act: MyD42 (JunB), GADD45 protein (homologous to GADD45{alpha}), MyD118/GADD45{beta}, and also a cDNA representing GADD45{beta} (Table III). As mentioned earlier, JunB is a component of the transcription factor AP-1, which functions in terminal differentiation of hematopoietic cells, as well as in modulating stress-associated programmed cell death (28). GADD45{alpha} and GADD45{beta} are growth-suppressive and apoptotic proteins that interact with cell cycle proteins (28, 29). We confirmed Act-induced up-regulation of the cDNA representing GADD45{beta} by real-time RT-PCR (Table V), but the significance of GADD45 proteins in Act-treated macrophages awaits further study.

Several cytokines were up-regulated by Act, including Mip-1{alpha}, Mip-2, and RANTES, which were verified by RT-PCR or ELISA. Also up-regulated by Act was SOCS3, which is known to be induced by various cytokines and subsequently acts to alter cytokine signaling via feedback inhibition of the Janus kinase and signal transducer and activator of transcription pathway (12, 30) (Tables III and V). Given the up-regulation of SOCS3 by Act and the transient nature of cytokine production by Act-treated macrophages, it is possible that Act-induced expression of SOCS3 is responsible in part for the eventual decrease in cytokine gene expression. Surprisingly, up-regulation of IL-6 by Act, which was previously demonstrated by RT-PCR and ELISA (7), was not detected by microarray analysis. This underscores the need to verify genes discovered using microarray techniques and to expect some false negatives. Our laboratory previously demonstrated Act-induced activation of the transcription factor CREB by gel shift assay, but transcriptional up-regulation of CREB was not detected by microarray analysis. One explanation for this discrepancy may lie in an inherent flaw of DNA array techniques, which is their detection of transcript induction but not protein activity, such as regulation by phosphorylation or inhibitor degradation. Other examples of non-transcriptional cellular responses to Act that were not detected using microarrays included calcium mobilization, ROS production, and activation of kinase cascades.

Regardless of the limitations of microarray techniques, however, there were changes in gene expression detected by microarray analysis that have been implied by previous studies. For example, we previously demonstrated Act-induced production of PGE2 by ELISA. We also demonstrated, by Northern blot analysis, Act-induced up-regulation of a key prostaglandin synthesis enzyme, COX-2, and production of the necessary substrate, arachidonic acid, via up-regulation of PLA2 (8). Although it is unclear why the array data did not include up-regulation of either COX-2 or PLA2, we did not expect direct detection of PGE2 production using microarrays, because prostaglandins are not proteins. However, the enzyme responsible for PGE2 production, glucocorticoid-regulated inflammatory prostaglandin G/H synthase, was deemed to be significantly up-regulated by Act by all analyses and clustering methods (Table III). Our laboratory also demonstrated Act-induced cAMP production, which is in accordance with the up-regulation of adenosine A2b receptor as determined by microarray analysis (Table III). Overall, the microarray data was confirmed a priori by traditional laboratory techniques, by real-time RT-PCR, and by demonstration of Act-induced apoptosis.

The data obtained using microarrays were consistent with previous in vivo experiments, which demonstrated that Act caused extensive inflammation and intestinal damage (2, 6, 31). The transcripts determined to be up-regulated by Act, using microarrays, were also consistent with previous findings using RAW 264.7 cells (7, 8), and many were inflammatory in nature. However, it is possible that signaling pathways induced by Act in RAW 264.7 cells may differ from those that occur in primary macrophages recruited during infection. Recent literature suggests that caution should be taken when extrapolating data from RAW 264.7 cells to primary cells (32-35). For instance, LPS has been shown to activate all three families of mitogen-activated protein kinases (MAPKs) in a rat alveolar macrophage cell line (NR8383) as well as in two murine macrophage cell lines (RAW 264.7 and J774A.1); however, LPS activates only the extracellular signal-regulated kinase 1/2 MAPK pathway in rat primary alveolar macrophages (32-34). Additionally, rapamycin-induced inhibition of DNA synthesis was much less pronounced in bone marrow-derived macrophages than in the murine macrophage cell line BAC1.2F5 (35).

Although it is generally accepted that cultured mammalian cells may differ to some extent in gene expression profiles compared with cells in vivo, RAW 264.7 cells behave similarly to primary macrophages in many respects (25, 36-39). For example, RAW 264.7 cells have been shown to secrete the potent vasorelaxant adrenomedullin in response to phorbol ester, retinoic acid, LPS, and interferon-{gamma}. Mouse peritoneal macrophages were also shown to produce and secrete adrenomedullin at a rate similar to that of RAW 264.7 cells, and secretion was enhanced by LPS treatment (36). RAW 264.7 cells behave similarly to primary human alveolar macrophages in response to Mycobacterium tuberculosis infection, specifically that M. tuberculosis-induced production of TNF-{alpha}, but not of nitric oxide, is toll receptor-dependent (37). Additionally, phagocytosis of diesel exhaust particles by RAW 264.7 cells or primary rat alveolar macrophages has been shown to lead to induction of apoptosis via production of ROS (38). Anthrax lethal toxin kills both RAW 264.7 cells and primary mouse macrophages via a process that requires proteasome activity (39). Double-stranded RNA elicits similar signaling responses from RAW 264.7 cells and primary mouse macrophages: inducible nitric-oxide synthase expression and IL-1 release, which depend upon NF-{kappa}B activation (25).

Although RAW 264.7 cells have been utilized widely and successfully as a model for the study of macrophage responses to a variety of stimuli, confirmation of data in primary macrophages is essential. Our future studies will therefore be aimed at confirming microarray results using murine peritoneal macrophages and a human monocyte/macrophage cell line (e.g. U937 cells). Act-induced apoptosis will also be examined in primary murine cells and U937 cells and the pathway and intermediates involved delineated. Such microarray studies using primary cells will be crucial in dissecting the definitive role of signaling molecules that ultimately leads to the disease state in the host during Aeromonas infection.

Based on several analysis methods, the microarray results revealed a set of consistent and significant genes that were induced in murine macrophages by Act. Our laboratory previously confirmed the up-regulation of many of these genes at the transcript and protein levels, lending credence to the data obtained using microarrays. In addition, we confirmed several Act-induced gene expression changes by ELISA and real-time RT-PCR. Twenty-two of the significant genes discovered using microarrays were also found to cluster together, suggesting their coordinated expression in one or more related pathways. Microarray analysis also revealed that Act induced up-regulation of several genes that were related to apoptosis, which we subsequently confirmed by annexin V staining and DNA laddering. Although our laboratory previously demonstrated that Act induces an inflammatory response in vitro and in vivo, the use of microarrays provided a global picture of early transcriptional responses to this toxin. Additionally, the microarray analysis uncovered genes induced by Act that were previously not suspected to play a role in the host response to A. hydrophila. Act-induced genes that were discovered using microarrays, such as Jun-B, TDAG51, and GADD45, are prospects for future research and may aid in the further elucidation of the mechanism of action of Act.


   FOOTNOTES
 
* This work was supported by Grant AI41611 from NIAID, National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Back

{ddagger} Predoctoral fellow and recipient of a grant from the National Science Foundation. Back

§ These authors contributed equally to the paper. Back

Supported in part by a McLaughlin postdoctoral fellowship. Back

|| To whom correspondence should be addressed: Dept. of Microbiology and Immunology, Medical Research Bldg., 301 University Blvd., Galveston, TX 77555-1070. Tel.: 409-747-0578; Fax: 409-747-6869; E-mail: achopra{at}utmb.edu.

1 The abbreviations used are: spp., species; Act, cytotoxic enterotoxin; RT, reverse transcriptase; TNF, tumor necrosis factor; IL, interleukin; PLA2, phospholipase A2; COX-2, cyclooxygenase-2; PGE2, prostaglandin E2; CREB, cyclic AMP-response element binding protein; ROS, reactive oxygen species; Mip, macrophage inflammatory protein; RANTES, regulated on activation normal T cell expressed; TRAF, tumor necrosis factor receptor-associated factor; TDAG51, T-cell death-associated gene 51; SOCS3, suppressor of cytokine signaling 3; ELISA, enzyme-linked immunosorbent assay; GADD, growth arrest and DNA-damage-inducible protein; LPS, lipopolysaccharide; SAM, Significance Analysis of Microarrays; PCA, principal component analysis; DEPC, diethyl pyrocarbonate; PBS, phosphate-buffered saline; ABTS, 2,2'-azinobis-(3-ethylbenzthiazoline-6-sulfonic acid); FITC, fluorescein isothiocyanate; PI, propidium iodide; MAS, Microarray Suite; FDR, false detection ratio; Lasp1, LIM and SH3 protein 1; JNK, c-Jun NH2-terminal kinase; c-IAP, cellular inhibitor of apoptosis; MyD, myeloid differentiation; CIS3, cytokine-inducible SH2-containing protein 3; MAPK, mitogen-activated protein kinase; ANOVA, analysis of variance. Back

2 Eisen laboratory (rana.lbl.gov/EisenSoftware.htm). Back


   ACKNOWLEDGMENTS
 
Drs. T. Wood and B. Luxon from the Department of Human Biological Chemistry and Genetics, University of Texas Medical Branch, Galveston, TX, provided facilities of their core laboratories for these studies. Celso Gutierrez, Jr. provided computer software and graphics expertise. Mardelle Susman provided editing services. Rebecca Alyea aided in proofreading the paper.



   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 EXPERMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Janda, J. M., and Abbott, S. L. (1998) Clin. Infect Dis. 27, 332-344[Medline] [Order article via Infotrieve]
  2. Chopra, A. K., and Houston, C. W. (1999) Microbes Infect. 1, 1129-1137[CrossRef][Medline] [Order article via Infotrieve]
  3. Merino, S., Aguilar, A., Nogueras, M. M., Regue, M., Swift, S., and Tomas, J. M. (1999) Infect. Immun. 67, 4008-4013[Abstract/Free Full Text]
  4. Merino, S., Rubires, X., Knochel, S., and Tomas, J. M. (1995) Int. J. Food Microbiol. 28, 157-168[CrossRef][Medline] [Order article via Infotrieve]
  5. Ferguson, M. R., Xu, X. J., Houston, C. W., Peterson, J. W., Coppenhaver, D. H., Popov, V. L., and Chopra, A. K. (1997) Infect. Immun. 65, 4299-4308[Abstract]
  6. Sha, J., Kozlova, E. V., and Chopra, A. K. (2002) Infect. Immun. 70, 1924-1935[Abstract/Free Full Text]
  7. Chopra, A. K., Xu, X., Ribardo, D., Gonzalez, M., Kuhl, K., Peterson, J. W., and Houston, C. W. (2000) Infect. Immun. 68, 2808-2818[Abstract/Free Full Text]
  8. Ribardo, D. A., Kuhl, K. R., Boldogh, I., Peterson, J. W., Houston, C. W., and Chopra, A. K. (2002) Microb. Pathog. 32, 149-163[CrossRef][Medline] [Order article via Infotrieve]
  9. Adams, J. M., and Cory, S. (2001) Trends Biochem. Sci. 26, 61-66[CrossRef][Medline] [Order article via Infotrieve]
  10. Adams, J. M., and Cory, S. (1998) Science 281, 1322-1326[Abstract/Free Full Text]
  11. Cavaillon, J. M. (1994) Biomed. Pharmacother. 48, 445-453[CrossRef][Medline] [Order article via Infotrieve]
  12. Kovanen, P. E., and Leonard, W. J. (1999) Curr. Biol. 9, R899-R902[CrossRef][Medline] [Order article via Infotrieve]
  13. Deleted in proof
  14. Wittwer, C. T., Herrmann, M. G., Moss, A. A., and Rasmussen, R. P. (1997) BioTechniques 22, 130-138[Medline] [Order article via Infotrieve]
  15. Crescenzi, M., and Giuliani, A. (2001) FEBS Lett. 507, 114-118[CrossRef][Medline] [Order article via Infotrieve]
  16. Ortiz, B. D., Nelson, P. J., and Krensky, A. M. (1997) Immunol. Today 18, 468-471[CrossRef][Medline] [Order article via Infotrieve]
  17. Schall, T. J., Jongstra, J., Dyer, B. J., Jorgensen, J., Clayberger, C., Davis, M. M., and Krensky, A. M. (1988) J. Immunol. 141, 1018-1025[Abstract]
  18. Tartaglia, L. A., and Goeddel, D. V. (1992) Immunol. Today 13, 151-153[CrossRef][Medline] [Order article via Infotrieve]
  19. Scott, M., Fujita, T., Liou, H., Nolan, G., and Baltimore, D. (1993) Genes Dev. 7, 1266-1276[Abstract/Free Full Text]
  20. Ashkenazi, A., and Dixit, V. M. (1998) Science 281, 1305-1308[Abstract/Free Full Text]
  21. Wang, C.-Y., Mayo, M. W., Korneluk, R. G., Goeddel, D. V., and Baldwin, A. S., Jr. (1998) Science 281, 1680-1683[Abstract/Free Full Text]
  22. Beg, A. A., and Baltimore, D. (1996) Science 274, 782-784[Abstract/Free Full Text]
  23. Yoneda, T., Imaizumi, K., Maeda, M., Yui, D., Manabe, T., Katayama, T., Sato, N., Gomi, F., Morihara, T., Mori, Y., Miyoshi, K., Hitomi, J., Ugawa, S., Yamada, S., Okabe, M., and Tohyama, M. (2000) J. Biol. Chem. 275, 11114-11120[Abstract/Free Full Text]
  24. Park, Y. C., Ye, H., Hsia, C., Segal, D., Rich, R. L., Liou, H. C., Myszka, D. G., and Wu, H. (2000) Cell 101, 777-787[CrossRef][Medline] [Order article via Infotrieve]
  25. Heitmeier, M. R., Scarim, A. L., and Corbett, J. A. (1998) J. Biol. Chem. 273, 15301-15307[Abstract/Free Full Text]
  26. Park, C. G., Lee, S. Y., Kandala, G., and Choi, Y. (1996) Immunity 4, 583-591[CrossRef][Medline] [Order article via Infotrieve]
  27. Hinz, T., Flindt, S., Marx, A., Janssen, O., and Kabelitz, D. (2001) Cell. Signal. 13, 345-352[CrossRef][Medline] [Order article via Infotrieve]
  28. Liebermann, D. A., and Hoffman, B. (2002) Oncogene 21, 3391-3402[CrossRef][Medline] [Order article via Infotrieve]
  29. Sheikh, M. S., Hollander, M. C., and Fornance, A. J., Jr. (2000) Biochem. Pharmacol. 59, 43-45[CrossRef][Medline] [Order article via Infotrieve]
  30. Alexander, W. S., Starr, R., Metcalf, D., Nicholson, S. E., Farley, A., Elefanty, A. G., Brysha, M., Kile, B. T., Richardson, R., Baca, M., Zhang, J. G., Willson, T. A., Viney, E. M., Sprigg, N. S., Rakar, S., Corbin, J., Mifsud, S., DiRago, L., Cary, D., Nicola, N. A., and Hilton, D. J. (1999) J. Leukocyte Biol. 66, 588-592[Abstract]
  31. Xu, X.-J., Ferguson, M. R., Popov, V. L., Houston, C. W., Peterson, J. W., and Chopra, A. K. (1998) Infect. Immun. 66, 3501-3509[Abstract/Free Full Text]
  32. Rao, K. M., Meighan, T., and Bowman, L. (2002) J. Toxicol. Environ. Health A 65, 757-768[CrossRef][Medline] [Order article via Infotrieve]
  33. Barbour, S. E., Wong, C., Rabah, D., Kapur, A., and Carter, A. D. (1998) Mol. Immunol. 35, 977-987[CrossRef][Medline] [Order article via Infotrieve]
  34. Rao, K. M. (2001) J. Leukocyte Biol. 69, 3-10[Abstract/Free Full Text]
  35. Hamilton, J. A., Byrne, R., Whitty, G., Vadiveloo, P. K., Marmy, N., Pearson, R. B., Christy, E., and Jaworowski, A. (1998) Int. J. Biochem. Cell Biol. 30, 271-283[CrossRef][Medline] [Order article via Infotrieve]
  36. Kubo, A., Minamino, N., Isumi, Y., Katafuchi, T., Kangawa, K., Dohi, K., and Matsuo, H. (1998) J. Biol. Chem. 273, 16730-16738[Abstract/Free Full Text]
  37. Means, T. K., Jones, B. W., Schromm, A. B., Shurtleff, B. A., Smith, J. A., Keane, J., Golenbock, D. T., Vogel, S. N., and Fenton, M. J. (2001) J. Immunol. 166, 4074-4082[Abstract/Free Full Text]
  38. Hiura, T. S., Kaszubowski, M. P., Li, N., and Nel, A. E. (1999) J. Immunol. 163, 5582-5591[Abstract/Free Full Text]
  39. Tang, G., and Leppla, S. H. (1999) Infect. Immun. 67, 3055-3060[Abstract/Free Full Text]

Add to CiteULike CiteULike   Add to Complore Complore   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us   Add to Digg Digg   Add to Reddit Reddit   Add to Technorati Technorati    What's this?


This article has been cited by other articles:


Home page
 CVIHome page
B. Desnues, M. Ihrig, D. Raoult, and J.-L. Mege
Whipple's Disease: a Macrophage Disease
Clin. Vaccine Immunol., February 1, 2006; 13(2): 170 - 178.
[Full Text] [PDF]

Home page
 Infect. Immun.Home page
C. L. Galindo, A. A. Fadl, J. Sha, L. Pillai, C. Gutierrez Jr., and A. K. Chopra
Microarray and Proteomics Analyses of Human Intestinal Epithelial Cells Treated with the Aeromonas hydrophila Cytotoxic Enterotoxin
Infect. Immun., May 1, 2005; 73(5): 2628 - 2643.
[Abstract] [Full Text] [PDF]

Home page
 Infect. Immun.Home page
J. E. Comer, C. L. Galindo, A. K. Chopra, and J. W. Peterson
GeneChip Analyses of Global Transcriptional Responses of Murine Macrophages to the Lethal Toxin of Bacillus anthracis
Infect. Immun., March 1, 2005; 73(3): 1879 - 1885.
[Abstract] [Full Text] [PDF]

Home page
 Infect. Immun.Home page
S. Jeyaseelan, H. W. Chu, S. K. Young, and G. S. Worthen
Transcriptional Profiling of Lipopolysaccharide-Induced Acute Lung Injury
Infect. Immun., December 1, 2004; 72(12): 7247 - 7256.
[Abstract] [Full Text] [PDF]

Home page
 J. Biol. Chem.Home page
C. L. Galindo, A. A. Fadl, J. Sha, C. Gutierrez Jr., V. L. Popov, I. Boldogh, B. B. Aggarwal, and A. K. Chopra
Aeromonas hydrophila Cytotoxic Enterotoxin Activates Mitogen-activated Protein Kinases and Induces Apoptosis in Murine Macrophages and Human Intestinal Epithelial Cells
J. Biol. Chem., September 3, 2004; 279(36): 37597 - 37612.
[Abstract] [Full Text] [PDF]

Home page
 Infect. Immun.Home page
C. L. Galindo, A. A. Fadl, J. Sha, and A. K. Chopra
Microarray Analysis of Aeromonas hydrophila Cytotoxic Enterotoxin-Treated Murine Primary Macrophages
Infect. Immun., September 1, 2004; 72(9): 5439 - 5445.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
278/41/40198    most recent
M305788200v1
Right arrow Submit a Letter to Editor
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
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 PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowRequest Permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Galindo, C. L.
Right arrow Articles by Chopra, A. K.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Galindo, C. L.
Right arrow Articles by Chopra, A. K.
Social Bookmarking
Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
 All ASBMB Journals   Molecular and Cellular Proteomics 
 Journal of Lipid Research   ASBMB Today 
Copyright © 2003 by the American Society for Biochemistry and Molecular Biology.
Advertisement
spacer
Advertisement
Advertisement