| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Call For Papers: 2nd International Symposium on Animal Functional Genomics
1 Immunogenetics Laboratory
2 Center for Animal Functional Genomics, Department of Animal Science, Michigan State University, East Lansing, Michigan
ABSTRACT
The objective of this study was to characterize a large portion of the bovine neutrophil transcriptome following treatment with the anti-inflammatory glucocorticoid dexamethasone (Dex). Total RNA was isolated from blood neutrophils of healthy cattle (5 castrated male Holsteins) immediately following cell purification (0 h) or after ex vivo aging for 4 h with or without added Dex. Additional neutrophils were cotreated with a glucocorticoid receptor (GR) antagonist (RU486) and Dex for 4 h. RNA was amplified, dye labeled (Cy3 or Cy5), and hybridized to a series of National Bovine Functional Genomics Consortium (NBFGC) microarrays. LOWESS data normalization followed by mixture model analyses showed that 11.15% of the spotted NBFGC cDNAs (2,036/18,263) were expressed in 4-h (untreated) neutrophils. Subsequent two-step mixed-model analysis detected (P 
0.05) 1,109 differentially expressed genes, of which contrast analysis indicated those that were independently responsive to aging (1,064), Dex (502), RU486 + Dex (141), or RU486 (357). In silico analysis revealed that 416 of the differentially expressed genes are unknown, 59 did not cluster well based on known function, and 634 clustered into 20 ontological categories. Independent validation of differential expression was done for 14 of the putatively Dex-responsive genes across these categories. Results showed that Dex induced rapid translocation of GR into the neutrophil nucleus and signaled dramatic alterations in expression of genes that delay apoptosis, enhance bactericidal activity, and promote tissue remodeling without inflammation or fibrosis. Thus these findings revealed hitherto unappreciated plasticity of blood neutrophils and potentially novel anti-inflammatory/wound-healing actions of glucocorticoids.
cDNA microarray; National Bovine Functional Genomics Consortium microarray; polymorphonuclear neutrophil; dexamethasone; RU486
GLUCOCORTICOIDS ARE STEROID hormones well known for their potent therapeutic anti-inflammatory properties (25, 51, 76). Secretion of endogenous glucocorticoids from the anterior pituitary following cytokine activation of the hypothalamus-pituitary-adrenal axis also plays a role in dampening immune responses following infection (62, 74). As such, glucocorticoids traditionally have been considered as immunosuppressive hormones. Although a paucity of published mechanistic studies on glucocorticoid actions in neutrophils exist, some studies suggest that the steroid may work by downregulating expression of key genes involved in cell trafficking to inflamed tissues (9, 73), shutting down oxidative metabolism and corresponding production of tissue-damaging reactive oxygen species (ROS) (21) as well as inhibiting the cell's ability to degranulate and thus to spill ROS and potent extracellular matrix (ECM)-degrading enzymes onto tissues (37). Accumulating evidence indicates that the anti-inflammatory actions of glucocorticoids in neutrophils may be via inhibition of proinflammatory signaling pathways that are important in most cells, including NF-
B, AP-1, CAMP response element binding protein, signal transducer and activator of transcription (STAT)5, and others (13, 20).
The targets and effects of glucocorticoid action in most cells are twofold: nongenomic and genomic. Nongenomic effects of glucocorticoids occur very rapidly (minutes) through three main mechanisms: 1) nonspecific interactions with cellular membranes changing their physiochemical properties, 2) specific interaction with cytoplasmic receptors (glucocorticoid receptor; GR) and subsequent modification of cytosolic signaling pathways, and 3) specific interactions with membrane-bound GR (present in only a few cell types) resulting in rapid signaling events that alter cellular function (reviewed in Refs. 11, 25, 64). These nongenomic effects seem to require pharmacological doses of the steroid, such as those achieved with high-dose glucocorticoid therapy.1
In the present study, the classical genomic mechanism of glucocorticoid action was of primary interest. Initiation of this mechanism requires binding of the steroid to cytoplasmic GR and subsequent translocation of hormone-bound receptors into the nucleus where they modify transcription of glucocorticoid-responsive target genes (reviewed in Refs. 25, 47, 51). In the absence of agonistic ligands such as dexamethasone or cortisol, GR is sequestered in the cytoplasm by its association with chaperone proteins. On glucocorticoid binding, GR becomes phosphorylated and dissociates from its chaperones, exposing dimerization and nuclear translocation motifs that enable it to translocate through pores in the nuclear membrane. Once translocated to the nucleus, agonist-activated GR has the capacity to bind directly to specific DNA sequences known as glucocorticoid response elements (GRE) in regulatory regions of glucocorticoid-sensitive genes, with subsequent induction or inhibition of gene expression. In addition, agonist-activated GR can interact with multiple other transcriptional factors (e.g., NF-B), altering gene expression through protein-protein-DNA-binding events that either interfere with or enhance target gene transcription (reviewed in Ref. 20). These genomic effects of the steroid occur rapidly (1–4 h, depending on hormone concentration), profoundly affect cell behavior, and are inhibited by GR antagonists such as RU486 (5).
RU486 uses both passive and active antagonism to block glucocorticoid-induced GR transactivation/repression and subsequent effects on cell behavior. Passive antagonism is simply the blocking of agonist (e.g., dexamethasone or cortisol) binding to GR. However, the active antagonism of RU486 is more complex and involves its bulky side chain, which alters GR phosphorylation status and subcellular location of GR and inhibits coactivator binding to GR with simultaneous recruitment of corepressors (29, 31). Accordingly, RU486 is particularly effective in blocking GRE-dependent transactivation/repression of glucocorticoid-responsive target genes (29).
The hypothesis of the current study was that glucocorticoid treatment of isolated bovine neutrophils would induce an expression profile in acute responding target genes involved in key innate immune and inflammatory processes of the cells, and that RU486 would reverse these effects. Our main objectives were to 1) characterize the transcriptome of isolated neutrophils aged in the presence or absence glucocorticoid hormone for 4 h using cDNA microarray analysis and 2) substantiate glucocorticoid effects on key pathway genes and end point cell behaviors elucidated by the microarray results through independent mRNA abundance and cell phenotyping assays.
MATERIALS AND METHODS
Neutrophil Donors, Isolation, and Culture
Blood (100–200 ml) for neutrophil isolations was obtained as needed from two to six (depending on experiment; see below) donor animals (healthy male Holstein cattle that were castrated at 1 mo of age, referred to as steers in the remainder of the text). Steers were fed and housed according to standard operating procedures at the Michigan State University (MSU) Dairy Teaching and Research Facility, and they were 3–6 mo of age during blood collection periods. Steer use for the described experiments was approved by the MSU All University Committee on Animal Care and Use (approval no. 07/04-104-00).
Blood was collected into acid citrate dextrose (ACD) anticoagulant, and neutrophils were isolated using Percoll density gradient centrifugation, as described in Weber et al. (73). The isolated cells were enumerated (Z1 Coulter Particle Counter; Beckman Coulter, Miami, FL) and suspended to between 1 x 106 and 3 x 107 cells/ml (depending on the experiment; see below) in basic culture medium [RPMI 1640 (Invitrogen, Carlsbad, CA), 1.0% fetal bovine serum (low-endotoxin FBS; Hyclone, Logan, UT), and 25 U/ml penicillin plus 25 µg/ml streptomycin (Invitrogen)]. All neutrophil preparations were checked flow cytometrically (FACSCaliber flow cytometer and CellQuest software; Becton Dickinson, San Jose, CA) for purity (immunostaining done with the granulocyte-specific marker G1) and viability (propidium iodide uptake), as described previously (41, 72). Neutrophil purity ranged from 95 to 99%, with eosinophils being the main contaminating cell type. Viability of the freshly isolated neutrophils was always 
98%. Dexamethsone (Dex) (Azium; Schering Plough, Animal Health, Kenilworth, NJ) was used at 10–7 M for GR agonism and mifepristone (RU486) (Sigma Chemical, St. Louis, MO) at 10–6 M for GR antagonism, as in Madsen-Bouterse et al. (41). Cells were incubated (39°C in moist 5% CO2 air) for 30 min with no added chemicals or with RU486 and then for an additional 4–12 h with no added chemicals or with Dex (16, 41) before use in the various experiments of this study.
Protein Isolation and Western Blotting
Isolated neutrophils (2 x 107 cells) from a randomly selected steer were lysed in 1x Laemmli sample buffer and used to document expression of GR (
-isoform, which is the ligand-binding form of the receptor) and progesterone receptor (PR; forms A and B) by Western blot analysis (see below). Lysate from snap-frozen endometrial tissue of one Holstein cow (collected from a local slaughterhouse) was used as the positive control tissue for both receptors. This analysis was performed because RU486, selected as the GR antagonist for the gene expression and cell behavior experiments, also acts as a potent PR antagonist (31). Thus it was necessary to determine whether bovine neutrophils express PR.
Subsequently, Western blot analysis of GR subcellular location was assessed in neutrophils (3 x 107 cells) following aging for 4 h in basic culture medium containing the various treatments to be used in the gene expression and cell phenotype experiments (none, Dex, RU486 + Dex, or RU486 alone). The nuclear and cytosolic fractions were obtained as described by Kulyte et al. (35) from a second randomly selected steer. Briefly, neutrophils were washed twice with ice-cold PBS and suspended and swelled in solution A [10 mM NaCl, 10 mM Tris·HCl, pH 7.5, 3 mM MgCl2, 0.05% IGEPAL CA-630 (Sigma), 1 mM phenylmethylsulfonyl fluoride, 1 µg/ml leupeptin, 1 µg/ml pepstatin, 5 mM NaF, and 1 mM Na3VO4] at 4°C for 15 min. The cell suspensions were vortexed vigorously for 10 s, an equal volume of solution B was added (solution A containing 0.6 M sucrose), and the mixture was centrifuged at 1,500 g for 5 min at 4°C. After centrifugation, supernatants (cytosolic fractions) were transferred to new tubes and subjected to 10,000 g centrifugation for 15 min. Pellets (nuclear fractions) were washed with a mixture of solutions A + B (1:1, vol/vol) and centrifuged at 1,500 g for 5 min at 4°C. Nuclei pellets were then washed in solution C (solutions A + B without IGEPAL CA-630) and checked under a light microscope for purity and integrity before lysis in 1x Laemmli sample buffer.
For Western blotting, 40 µg of total protein from each sample were loaded onto 10% SDS-PAGE gels and separated electrophoretically. After transfer to nitrocellulose membranes (Bio-Rad Laboratories, Hercules, CA), separated proteins were probed with anti-PR (A and B forms) antibody (MA1-410, a mouse monoclonal IgG1, 0.5 µg/ml working concentration; Abcam, Cambridge, UK) and/or anti-GR antibody (PA1-511A, rabbit polyclonal IgG, 0.5 µg/ml working concentration; Affinity BioReagents, Golden, CO). Goat anti-rabbit horseradish peroxidase-conjugated antibody (1:10,000 working dilution; Pierce, Rockford, IL) was used as the secondary antibody. Western blots were developed using the SuperSignal West Pico Chemiluminescent Substrate system (Pierce), as described (73), photographed, stripped (16), and re-probed with ß-actin antibody (mouse monoclonal IgG1, catalog no. AC-15: ab6276; Abcam) as the lane-loading control (41, 73).
Assay of a Candidate Glucocorticoid-Responsive Gene
Quantitative real-time RT-PCR and nuclear run-on assays were performed to document transcriptional effects of Dex (and antagonistic effects of RU486) on expression of a candidate target gene, L-selectin (CD62L). The CD62L gene was selected for this experiment because the adhesion molecule it encodes is known to be profoundly downregulated on the surface of blood neutrophils of glucocorticoid-challenged cattle (9, 72, 73). Also, demonstration of transcriptional downregulation of CD62L in Dex-treated cells would help link Western blot observations of GR subcellular location with the receptor's putative genomic effects in bovine blood neutrophils. Finally, CD62L is not represented on the National Bovine Functional Genomics Consortium (NBFGC) array used in the microarray experiment (see below), so documentation of Dex and RU486 effects on CD62L gene expression through this preliminary experiment acted as an important validation that these treatments cause expected genomic effects in neutrophils.
A fully quantitative real-time RT-PCR assay was developed to profile CD62L mRNA abundance in 3 x 107 neutrophils/culture that were treated for 4 h with nothing, Dex, RU486 + Dex, or RU486 alone. A linear (R2 = 0.99, P = 0.0001) six-point standard curve was developed using a CD62L cDNA prepared by PCR. The PCR primers used to generate this CD62L cDNA were designed based on the mRNA sequence for Bos taurus selectin L (lymphocyte adhesion molecule-1) (SELL; GenBank accession no. NM_174182; GI:27901800) and were as follows: forward 5'-CCC AAC AAC AGG AAG AGT AAG-3' (base pairs 483–503) and reverse 5'-GCC TAT AGT TGC ATA TGT ATC AAA TTT TCA-3' (base pairs 1656–1627). The final PCR reaction mixture contained 1x PCR buffer, 1.5 mM MgCl2, 0.2 mM each dNTP, 0.5 µM forward primer, 0.5 µM reverse primer, 25 ng of cDNA template (from bovine total leukocytes), and 1 U/reaction Taq DNA polymerase (Invitrogen), all brought to a final volume of 25 µl with sterile Milli-Q water. PCR reactions were carried out in a RoboCycler Gradient 96 (Stratagene, La Jolla, CA) under the following conditions: denature at 95°C for 3 min followed by 40 cycles of 95°C for 30 s (denature), 50°C for 30 s (anneal), and 72°C for 30 s (extend), and a final extension at 72°C for 10 min. The PCR amplification product was then ligated into pGEM-T Easy vector (Promega, Madison, WI), and the recombinant plasmid was transformed into JM109 competent cells (Promega). Following DNA sequence verification (Agencourt Bioscience, Beverly, MA), large quantities of CD62L cDNA were generated by PCR using the conditions described above and the CD62L plasmid as template. This PCR product was gel purified using a 1% PCR Low Melt agarose gel (Sigma) and Wizard PCR Preps DNA Purification system (Promega) and serially diluted in sterile DNase-free water (10–12 to 10–18 g, standards) to construct the real-time RT-PCR standard curve. The cDNA standards or 2.5 ng of neutrophil test cDNA were added to reaction mixtures that contained CD62L real-time RT-PCR primers [forward 5'-ACG GGA AAA AAG GAT TAC TAT GGA-3' and reverse 5'-GCC TAT AGT TGC ATA TGT ATC AAA TTT TCA-3'; product length = 144 bp, melting temperature (Tm) = 74°C], and the SYBR Green PCR Master Mix system was used for real-time fluorescence detection in a PE7700 thermal cycler (Perkin Elmer Applied Biosystems), as described previously (40). CD62L mRNA abundance in the test samples was calculated using the equation for the linear standard curve, which converted the number of PCR cycles to threshold (Ct) for each sample into the concentration of CD62L mRNA (in fentograms; fg) present in the original starting sample. Each test sample, the standard curve, and a negative control (no cDNA template) were run in triplicate.
For nuclear run-on assay of CD62L gene transcription, nuclei in the variously treated neutrophils were isolated as described by Marucha et al. (44) with minor protocol modifications. Briefly, 2 x 107 neutrophils/sample were suspended in ice-cold homogenization buffer (0.3 M sucrose, 10 mM Tris·HCl, pH 7.4, 2 mM magnesium acetate, 3 mM CaCl2, 12 mM ß-mercaptoethanol, 0.25% NP-40) and incubated for 5 min at 4°C. The detergent-treated neutrophils were then under-layered with an equal volume of isolation buffer (0.6 M sucrose, 10 mM Tris·HCl, pH 8.0, 5 mM magnesium acetate, 12 mM ß-mercaptoethanol) and centrifuged at 500 g for 10 min at 4°C. Supernatants were removed, and pelleted nuclei were gently suspended in 110 µl of storage buffer (0.3 M sucrose, 10 mM Tris·HCl, pH 7.4, 2 mM magnesium acetate, 3 mM CaCl2, 12 mM ß-mercaptoethanol, and 50% glycerol), quick frozen in liquid nitrogen, and stored at –80°C until assay. At that time, the frozen nuclei were thawed on ice, pelleted by centrifuged (500 g for 10 min at 4°C), and suspended in 80 µl of 100 mM HEPES, pH 7.9, 180 mM NH4Cl, 10 mM MgCl2, 1 mM MnCl2, 0.2% EDTA, and 24% glycerol and subsequently incubated with 100 µCi of [-32P]UTP (6,000 Ci/mmol, Perkin Elmer) in nucleotide solution (0.4 mM ATP, 0.4 mM GTP, 0.4 mM CTP, 2 mM DTT, 10 µg/ml BSA) at 30°C for 120 min to label nascent RNA chains (68). Nuclei were then pelleted, suspended in 1 ml of 0.1 M Tris·HCl and 0.1 M EDTA containing 0.3% SDS, 100 µg/ml tRNA, 100 µg/ml proteinase K (Invitrogen), and 1,000 U of RNasin (Invitrogen) and incubated at 37°C for 30 min. Following incubation, the mixtures were transferred to 15-ml conical tubes, and the original tubes were rinsed with 1 ml of 100 mM sodium acetate, pH 5.0, and 20 mM EDTA and transferred to the same 15-ml tubes. Radiolabeled RNAs were extracted with equal volumes of phenol-chloroform-isoamyl alcohol (25:24:1) (Sigma), precipitated by adding 300 µl of 1 M Tris·HCl, pH 8.0, 200 µl of 3 M sodium acetate, pH 5.3, and 5.5 ml of 100% ethanol, and placed at –20°C overnight, and the precipitated RNA was recovered by centrifugation (3,000 g at 4°C for 30 min). The RNA (equal cpm per sample) was heat denatured at 65°C for 10 min and added to 4 ml of hybridization solution (50% formamide, 5x SSC, 50 mM sodium phosphate buffer, pH 6.5, 1x Denhardt's solution, 0.1% SDS) for hybridization to a series of nylon membrane strips (Roche Diagnostics, Indianapolis, IN), each spotted in excess with immobilized linearized/denatured plasmid DNA (
5 µg pGEM-T; Promega, San Luis Obispo, CA) containing the CD62L cDNA (test gene) or ß-actin cDNA [spot loading control described in Weber et al. (72)] or empty vector (as a nonspecific hybridization control). The membrane strips were prehybridized at 42°C in 50% formamide, 5x SSC, 50 mM sodium phosphate buffer, pH 6.5, 1x Denhardt's solution, 0.1% SDS, and 250 µg/ml tRNA (Roche) and hybridized for 60 h at 42°C. The blots were then washed at room temperature (2x SSC, 0.1% SDS) and at 50°C (1x SSC, 0.1% SDS) and exposed to BioMax MS film at –80°C for 120 h. Densities of resulting dots were quantified by use of scanning densitometry (GS-710 Calibrate Imaging Densitometer and Multi-Analyst software, Bio-Rad Laboratories), and the rate of CD62L gene transcription was expressed as a density ratio of CD62L to ß-actin (within blot). No probe hybridization to empty vector spots was ever detected.
Neutrophil RNA Used for the Microarray Experiments
TRIzol reagent (Invitrogen; 1.0 ml of TRIzol per 1 x 107 neutrophils) was used for the lysis and subsequent isolation of RNA (per manufacturer's instructions) from blood neutrophils (n = 5 steers) that were freshly isolated (0 h) or aged for 4 h plus or minus Dex or with RU486 + Dex. The RNA samples were treated with RQ1 RNase-free DNase (Promega), and concentration was determined (ND-1000 spectrophotometer; NanoDrop Technologies, Wilmington, DE). Because of the relatively low abundance of RNA in neutrophils, especially cells cultured for 4 h, a linear RNA amplification and dye coupling was performed using the SuperScript Indirect RNA Amplification system (Invitrogen) according to the manufacturer's instructions [which are based on the procedure described in Van Gelder et al. (70)]. Briefly, first-strand cDNA was synthesized using SuperScript III RT and an anchored oligo(dT) primer containing T7 promoter sequence, followed by second-strand synthesis and column purification of the double-stranded cDNA product. The cDNA template was then used in an in vitro transcription reaction (IVT) with T7 RNA polymerase and amino-allyl UTP to generate antisense RNA (aRNA). Following IVT and aRNA purification, the quality and quantity of the aRNA were evaluated using the Agilent Bioanlyzer RNA 6000 Nano LabChip (Agilent Technologies, Palo Alto, CA). The amino-allyl-modified aRNA was then labeled with Cy5 or Cy3 fluorescent dyes (Amersham Biosciences, Piscataway, NJ) and used to hybridize a series of 15 NBFGC microarrays [described in Suchyta et al. (66); clone information available at http://www.cafg.msu.edu] using the hybridization conditions described in Madsen et al. (40) and a GeneTAC Hybridization Station (Genomic Solutions, Ann Arbor, MI). Finally, arrays were removed, rinsed in 2x SSC, placed in individual open 50-ml conical tubes, dried by centrifugation (500 g for 3 min at room temperature), and scanned using a GeneTAC LS IV microarray scanner (Genomic Solutions) and accompanying software (version 3.01). Spot aligning was performed using MolecularWare DigitalGENOME Pro 2.7 (MolecularWare, Cambridge, MA), and total spot intensity values for each dye channel were stored as comma-separated value data files, exported into Excel, and loaded into SAS for data normalization and analyses (see below).
Statistical Analyses of Microarray Data
Data normalization.
Potential dye intensity biases in the microarray data sets were visualized using M-A scatter plots (79). Log2 intensity ratios, M = log2(Cy5/Cy3), were plotted against mean log2 intensities, A = (log2Cy5 + log2Cy3)/2, for each array. Array-specific data normalization was then performed using the locally weighted regression and smoothing scatter plot (LOWESS) procedure of the SAS software (55). The efficiency of LOWESS normalization was assessed by monitoring M-A plots for data from each array before and after LOWESS normalization. The normalized data were then back transformed before further statistical analyses using the following formulas: log2Cy3* = A – M*/2 and log2Cy5* = A + M*/2, where log2Cy3* and log2Cy5* are the normalized log intensities. Here, M* = M – 
represents each of the normalized M values, with being the LOWESS-predicted value for each spot.
Mixture model.
A mixture model technique was utilized to assess the proportion of nonexpressed genes among the probes represented on the NBFGC microarray. LOWESS-adjusted intensity values of the 4-h no-Dex neutrophils were used, and their empirical distributions were normalized across slides for location and scale adjustments. The average fluorescent intensity signals were then analyzed using a mixture model with two components, one relative to nonexpressed genes (background noise) and the other relative to expressed genes (background plus transcript signal). This analysis was performed using the software EMMIX (49).
Significance test.
LOWESS-adjusted log intensities were analyzed statistically for differential gene expression over aging time (0 vs. 4 h, no Dex) and steroid treatments (4 h with Dex, 4 h with RU486 + Dex) using a mixed-model approach consisting of two steps (77). The first step referred to across-slides normalization (global normalization), and the second step involved gene-specific analyses to compare the four experimental groups. The global normalization model included the effects of arrays and dyes and the interaction between arrays and dyes. The second step of the statistical analysis consisted of gene-specific models for the estimated residuals 
] obtained from the normalization approach described above. These models were as follows (52)
 |
represents the normalized fluorescent intensity signals for gene g; µ(g) is an overall mean value for gene g; D
is the gene-specific fixed effect of dye d; S
is the random effect of steer s; T
is the gene-specific fixed effect of treatment group t; ST
is the interaction between steer and treatment effects, assumed random; A
is the random effect of array a relative to steer s; and e is a stochastic error. The random terms of the model [S, T, ST, A, and e] were assumed normal with mean 0 and gene-specific variances. These analyses were computed using the MIXED procedure of SAS (55). To help distinguish between putative aging vs. GR agonist or antagonist treatment effects, specific mean contrasts were run for the 1,109 differentially expressed genes and included the following: 1) 0 vs. 4 h, no Dex; 2) Dex vs. no Dex, 4 h; 3) RU486 + Dex vs. Dex, 4 h; and 4) RU486 + Dex vs. no Dex, 4 h.
Ontological Clustering of Differentially Expressed Neutrophil Genes
The "Search Libraries" and "GeneLinks" functions of the MSU Center for Animal Functional Genomics (CAFG) interactive web site (http://www.cafg.msu.edu), which is annotated with expressed sequence tag (EST) sequence information, basic local alignment search tool (BLAST)n results, GenBank accession numbers, The Institute for Genomic Research (TIGR) cluster numbers, and putative gene names and functions [derived from the Gene Ontology (GO; http://www.godatabase.org), TIGR (http://www.tigr.org/) Bos Taurus Gene Index (BtGI), and GenBank and PubMed (http://www.ncbi.nlm.nih.gov) databases] for the NBFGC clones, were utilized to assign differentially expressed genes from the microarray experiment into putative ontological clusters (i.e., based on subcellular location, molecular pathway, and/or biological process). Fourteen genes were selected for further study based on 1) P values 0.05 for the main effect of treatment (refer to Supplemental Table 1; the online version of this article contains Supplemental Materials), 2) contrasts that indicated a significant (P 0.05) effect of Dex on their expression, and 3) our group's interest in the roles of neutrophil apoptosis and tissue remodeling in inflammation. Forty additional genes selected at random from across the ontology clusters were also tested for deeper validation of the microarray results. These 40 genes will not be discussed further, but results from the validation assays appear in Supplemental Table 2.
|
Available DNA sequences of the 14 test genes were obtained from TIGR (http://www.tigr.org/tigr-scripts/tgi/T_index.cgi?species=cattle), and primer sets were designed in-house (Primer Express software; Perkin Elmer Applied Biosystems, Foster City, CA) and synthesized at a commercial facility (Operon, Huntsville, AL). The NBFGC clone numbers, corresponding TIGR cluster numbers, putative gene names, and primer sets used for the real-time RT-PCR reactions of these genes are reported in Table 1. Gene expression changes were computed in duplicate reactions per sample using the 2–
Ct method of Livak and Schmittgen (39), for which ß-actin (Table 1) served as the normalizing (control) gene and the 4-h untreated neutrophils as the calibrator (i.e., these always had a ratio of 1.0). The ratio data were analyzed statistically, considering the nonparametric ANOVA approach of Friedman (80). Statistical significance was assigned if P 0.05. Because of large animal variation for some genes, trends were considered if 0.05 < P 0.07.
Neutrophil Phenotyping
Some striking themes of Dex-affected biological processes in neutrophils were implied by the results of the gene expression validation work. Accordingly, preliminary evidence of corresponding Dex-induced changes in neutrophil apoptosis, endocytosis, and tissue ECM-degrading activity was sought. Neutrophils from two to five donor steers (depending on the assay) were used in the phenotyping assays described below.
Apoptosis.
Apoptotic status of neutrophils (n = 5 steers; 1 x 106 per well of 96-well flat-bottom cell culture plates) aged in duplicate for 4 h in basic medium containing no added steroid, Dex, or RU486 + Dex and then for a subsequent 2 h with no additional treatment or with 100 ng/ml soluble FasL (recombinant human sFasL; Alexis, San Diego, CA) was assessed using two-color flow cytometry (FACSCalibur) and reagents and instructions supplied in a kit (Annexin V-FITC Apoptosis Detection kit; BD Biosciences Pharmingen, San Diego, CA). Data were acquired for 5,000 neutrophils per sample, and the percent apoptotic (annexin V-FITC+) cells was determined using the density plot analysis described previously (16, 40). Representative neutrophil cultures also were cytocentrifuged (Shandon Cytocentrifuge, Thermo Shandon Cytospin 4; Pittsburgh, PA) and stained with May-Grünwald-Giemsa stain (Sigma) for microscopic examination of apoptotic cells (Leica Microscope equipped with a Digital FireWire Color Camera system and IM50 software; Leica Microsystems, Bannockburn, IL). The flow cytometric data were analyzed statistically using the MIXED procedure of SAS (55) with a model that included treatments, incubation times, assay replications, and treatment x time interactions as fixed effects and animal and error terms as random effects. Multiple comparisons of means were performed using Tukey's procedure. Stated differences between observations were declared significant when P 0.05.
Endocytosis.
The ability of neutrophils (n = 2 steers) to take up nonspecific particles present in their environment was assessed using an assay in which green fluorescent latex beads (Fluoresbrite Carboxylate Microspheres YG, 1.75-µm size; Polyscience, Warrington, PA) served as targets for cells (1 x 106 neutrophils per well of 96-well flat-bottom cell culture plates) aged for 6 or 12 h in basic medium containing no added steroid, Dex, or RU486 + Dex. Following these treatments, the assay was performed in duplicate by combining the aged neutrophils with beads in a 1:10 ratio and incubating the mixtures for 30, 60, 90, or 120 min at 39°C with gentle shaking. Bead uptake by neutrophils was quantified flow cytometrically (%neutrophils that fluoresced green; FACSCalibur), and representative cells were captured photographically (Leica Microscope and Digital FireWire Color Camera system with IM50 software). The flow cytometric data were analyzed within assay time (30, 60, 90, and 120 min) using the MIXED procedure of SAS (55) with a model that included the fixed effect of experimental group (no steroid, Dex, and RU486 + Dex) and repetition (1, 2) and the random effects of steer and of steer x group interaction. Differences between experimental groups were declared significant when the group effect P 0.05. In such cases, the analyses were complemented with specific mean contrasts to distinguish between GR agonist and antagonist effects.
Tissue ECM-degrading activity.
The capacity for neutrophils to cause inflammatory tissue damage is mediated in part by the ability of the cell to release large amounts of ECM-degrading enzymes, including matrix metalloproteinase-9 (MMP-9). Gelatin zymography was used to measure MMP-9 activity in culture supernatants of neutrophils (n = 5 steers; 1.5 x 107 cells/well of 6-well culture plates) that were aged for 4 or 8 h in basic medium or in medium with added Dex, according to the method of Heussen and Dowdle (28). Briefly, 10 µl of culture supernatants per lane were electrophoresed in one-dimensional SDS-polyacrylamide gels impregnated with 0.25% (wt/vol) gelatin (Precast 10% Zymogram Ready Gel, Bio-Rad). All samples within animal were run on a single gel. Electrophoresis was conducted at 65 V for 30 min and then 120 V for an additional 60 min. Next, gels were washed twice for 30 min in 2.5% (vol/vol) Triton X-100 to remove SDS and then incubated in 50 mM Tris (pH 7.5), 5 mM CaCl2, and 0.02% NaN3 at 37°C for 18 h to enable MMP-9 activity (i.e., degradation of the gelatin). Gels were stained with Coomassie Brilliant Blue stain (25% methanol, 10% glacial acetic acid, 0.025% Coomassie Blue) for 90 min and destained (65% double distilled H2O, 25% methanol, 10% glacial acetic acid) for 60 min to visualize clear bands of digested gelatin. Band images were collected on the Fluor-S MultiImager (Bio-Rad), and the density of clearings at the 92-kDa marker was analyzed by scanning densitometry (GS-710 Calibrated Imaging Densitometer and Multi-Analyst software; Bio-Rad). MMP-9 activity was recorded as percent volume. The effect of Dex was determined by t-test (within 4- and 8-h incubation times) and declared significant when P 0.05.
RESULTS
Bovine Neutrophils Express Receptor for Glucocorticoid But Not for Progesterone
The main goal of the current study was to investigate direct effects of glucocorticoid on the transcriptome of isolated bovine blood neutrophils, using Dex as a GR agonist and RU486 as a GR antagonist. Western blot analysis was used to show that freshly isolated neutrophils expressed GR but not PR (Fig. 1A), indicating that GR is likely the main receptor through which Dex and RU486 carry out actions on expression of glucocorticoid-responsive genes in the cells. Further Western blot analyses of neutrophil cytosolic and nuclear fractions in cells aged ex vivo for 4 h showed that GR remained primarily cytosolic in untreated cells. In contrast, Dex caused GR translocation to the nuclei of the cells, a process that was clearly reduced by cotreatment with RU486 (Fig. 1B). These results were not artifacts of sample loading error, because the control gene (ß-actin) was expressed equally across lanes of these blots (Fig. 1, A and B).
|
Microarray Analysis of the Transcriptome in Dexamethasone-Treated Neutrophils
RNA amplification and quality.
The quantity of isolated neutrophil total RNA across treatment groups ranged from 3 to 10 µg/culture, and it was considered relatively high in quality based on Agilent BioAnalyzer results (Fig. 2A). Unlike other cells and tissues, neutrophil total RNA always has a slightly lower abundance of 28S rRNA than 18S rRNA, which is most likely due to endogenous RNA degradation that is either normally ongoing in these terminally differentiated and short-lived leukocytes and/or occurs during the 2.5-h cell isolation procedure. Indeed, an RNA profile identical to that shown in Fig. 2A has been reported for isolated human neutrophils (81). Nonetheless, use of samples with moderate RNA degradation has been shown to yield meaningful microarray results (59), so the neutrophil RNA samples of this study were processed further. To ensure that enough neutrophil RNA would be available for the microarray and subsequent gene expression validation assays, RNA (1 µg) from cells of each culture scenario was amplified for use in the microarray experiment. As expected, amplification resulted in the removal of 28S and 18S rRNA and produced an 1,000-fold increase in mRNA that ranged in size from 200 to 1,500 bp (representative profile shown in Fig. 2B).
|
|
40% decrease) on expression of proapoptotic caspase 8 compared with that observed in untreated cells, and RU486 completely removed this Dex effect (Fig. 4A). In the case of anti-apoptotic translocase of the outer mitochondrial membrane 70A (TOMM70A), both Dex and RU486 + Dex increased gene expression approximately threefold relative to untreated cells (Fig. 4B). All three steroid treatment scenarios abolished expression of proapoptotic TF-1 cell apoptosis-related gene-19 (TFAR-19) relative to that observed in untreated cells (Fig. 4C), while expression of anti-apoptotic B lymphocyte-induced maturation protein-1 (Blimp-1) showed pronounced upregulation (6.5-fold) in response to Dex, which was dampened, but not removed, by RU486 (Fig. 4D). These results suggested that ligand-activated GR had gene-specific genomic (Dex) and nongenomic (RU486 + Dex, and RU486 alone) reprogramming effects in neutrophils that would result in delayed apoptosis. To test this possibility, neutrophils were aged for 4 h in the presence and absence of Dex ± RU486 followed by an additional 2 h in the presence or absence of a known apoptosis inducer, sFasL. The flow cytometric data summarized in Fig. 5 show that Dex delayed annexin V-FITC staining of neutrophils during spontaneous apoptosis (bars at left) and sFasL-induced apoptosis of the cells (bars at right), both of which were completely reversed by the addition of RU486. In substantiation of these results, microscopic examination of representative May-Grünwald-Giemsa-stained neutrophils from the variously treated cultures showed that more untreated cells (see Fig. 6B) and cells subjected to sFasL stimulation without or with RU486 treatment (see Fig. 6, C and E, respectively) were smaller and had more rounded and condensed nuclei than freshly isolated neutrophils (Fig. 6A) \. or cells treated with Dex but not RU486 (Fig. 6D). Taken together, the gene expression profiles in Fig. 4 and phenotyping results in Figs. 5 and 6 indicate that glucocorticoid delays neutrophil cell death through GR-mediated reprogramming of multiple apoptosis regulatory pathways.
|
|
|
(PPAR-); Fig. 7C], the antimicrobial arsenal of the cells [bactericidal/permeability-increasing protein (BPI) and light chain 3 (LC3); Fig. 7, D and E], and homophilic cell-cell adhesion (calsyntenin; Fig. 7F). Particularly striking were the 2,900-fold upregulation of BPI [a gram-negative bacteria and LPS-binding protein with potent bacteriacidal and opsonizing activities (1)], the 13-fold induction of PPAR- [a transcription factor well known for its wound-sensing capacity and corresponding anti-inflammatory, anti-apoptotic, and anti-fibrotic gene-regulating activities (8, 65)], and the 4.8-fold upregulation of LC3 [involved in vesicle trafficking (33), including the endocytic uptake and killing of intracellular pathogens in autophagic vesicles (7, 26, 34)]. Data in Fig. 7 also showed that treatment with RU486 reversed most of the Dex-mediated inductions of these genes. Together, these data suggested that some of the genomic effects of glucocorticoids might be to upregulate processes involved in neutrophil sampling of and response to its extracellular and intracellular environment. To begin to test this possibility, neutrophils were aged ex vivo for 6 or 12 h in the presence or absence of Dex ± RU486 before monitored uptake of green fluorescence particles (latex beads) during a 30-, 60-, 90-, or 120-min assay period. Data from the 12-h cells are summarized in Fig. 8A, \. and representative fluorescence histogram plots and microscopic photographs are shown in Fig. 8, B and C. These data demonstrate that neutrophils aged in Dex-containing medium had a significantly enhanced rate and magnitude of bead uptake than cells aged in medium without added Dex, and that RU486 reversed this Dex effect on endocytosis. The same numerical trends were observed for neutrophils cultured for 6 h in the various steroid treatments (not shown), but P values ranged from 0.08 to 0.70 for the Dex effect because of high variation within assay times tested. Interestingly, compared with the clean surroundings of untreated neutrophils (Fig. 8C, top), it was also noted that the environment of neutrophils in the Dex-treated cultures contained substantial debris (Fig. 8C, bottom). While further experimentation is required to identify the nature of this debris, it is possible that the Dex-induced increase in particle uptake came along with increased release of autophagic vesicles in light of the corresponding induction in LC3 gene expression (Fig. 7E). These collective observations point to a rather unexpected possibility, that glucocorticoids may enhance neutrophil antimicrobial capacity while downregulating the cell's proinflammatory potential through GR-mediated genomic reprogramming of multiple genes that regulate these processes.
|
|
|
|
A predominant theme of glucocorticoid-induced delay in neutrophil apoptosis has been revealed through the gene expression profiles and cell phenotyping of this study and is supported by our previous studies showing pronounced effects of the steroid on other apoptosis genes not present on the NBFGC array (16, 40, 41). Apoptosis is characterized morphologically by loss of mitochondrial membrane potential, cell blebbing, redistribution of outer plasma membrane lipids, cytoplasmic condensation with cell shrinkage, compaction of nuclear chromatin, chromosomal DNA fragmentation, and cell dismantling into small apoptotic bodies that are readily cleared in vivo by neighboring macrophages, fibroblasts, and epithelial cells to prevent inflammation (6, 19). Terminally differentiated, mature neutrophils normally apoptose within 12–24 h of release from bone marrow into blood (56) using at least two signaling pathways (41, 42, 50, 57) that converge directly or indirectly on caspase 8 activation (56), leading to cell death (6). Thus, caspase 8 figures prominently in neutrophil apoptosis. In light of this, it was notable that our Dex-treated neutrophils had delayed spontaneous and sFasL-induced apoptosis, maintained normal cell size and nuclear morphology during sFasL stimulation, and had significantly reduced caspase 8 gene expression compared with untreated cells. In contrast, the RU486-treated neutrophils expressed similar amounts of caspase 8 mRNA as untreated cells and showed the same aging-related onset of spontaneous apoptosis and sFasL-induced apoptosis. In previous studies from our group, RU486 was also shown to reverse glucocorticoid-mediated changes in multiple apoptosis regulatory genes and hormone-induced delay in spontaneous apoptosis, including the activation of caspases 8 and 9 (16, 41). Combined with the fact that GR translocated rapidly to nuclei of the Dex-treated neutrophils and had clear transcriptional regulatory roles while there, the results described above hint strongly at a genomic influence of GR on pivotal molecules that regulate extrinsic and intrinsic pathways of apoptosis, although this theory must be tested experimentally using nuclear run-on and GR DNA-binding assays.
Results of the current study also revealed additional (but only recently appreciated) apoptosis-regulating genes that were strongly induced by Dex, including TOMM70A, symplekin, Blimp-1, MT-II, and PPAR-. In addition to energy production, mitochondria have critical functions in apoptosis (50). Mitochondria require hundreds of nuclear-encoded proteins to function properly, and the large majority of these must be transported from the cytosol into the organelle by a complex of membrane-bound proteins called translocases of the outer mitochondrial membrane, one of which is TOMM70A (14, 22, 23). As such, TOMM70A has likely roles in the regulation of mitochondrial membrane potential, morphology, and physiology (2, 22). Its upregulation in response to Dex may be connected to the steroid's apoptosis-delaying activities vis-à-vis maintenance of mitochondrial membrane integrity (41). Interestingly, our microarray results suggested that Dex also induced expression changes in dozens of mitochondrial structural genes, electron transport chain genes, and other redox regulatory genes that might be connected to apoptosis regulation. Although we have yet to validate these expression changes by real-time RT-PCR and their impacts on the cells using phenotypic assays, genes in these ontological categories reasonably may be coregulated with TOMM70A by GR to depress mitochondrial ROS production (10, 21, 54), preserve mitochondrial membrane integrity (41), and deactivate any released ROS, thus facilitating the glucocorticoid-induced delay in neutrophil apoptosis.
The cytoprotective properties of heat shock proteins (Hsp) include anti-apoptotic effects linked to their ability to prevent activation of caspases 8 and 9 and to regulate the activity of several prosurvival signaling cascades, including NF-B (reviewed in Ref. 6). Thus it is critical that the Hsp response be rapidly initiated for cell survival during stress. One important nuclear protein that induces Hsp responses is symplekin (78), the gene expression of which was observed to be increased approximately fourfold in our Dex-treated neutrophils relative to untreated and RU486-treated cells. In addition, the microarray results revealed 14 other cellular defense response genes that were putatively differentially expressed in the variously treated neutrophils, including Hsp60, Hsp70-1, Hsp70-IP, and members of the Hsp90 family. Thus ligand-activated GR appears to have genomic-level effects on the Hsp system in neutrophils, which may help influence the apoptotic status of the cells. In light of the extraordinary number of other apoptosis regulatory genes that have been uncovered in this and our previous studies (10, 16, 40, 41), the putative Hsp response to Dex warrants further validation in real-time RT-PCR and cell phenotyping assays.
The more than sixfold induction of Blimp-1 gene expression observed in the Dex-treated neutrophils is somewhat more tricky to connect directly with neutrophil apoptosis, although this potent transcriptional repressor is well known as a master regulator of B cell differentiation into long-lived antibody-secreting plasma cells in bone marrow (30, 61). During development of these plasma cells, increasing Blimp-1 expression induced by the bone marrow environment represses key genes involved in cell cycle progression and cell division, relieves repression from other transcription factors to enable continuous immunoglobulin secretion without the need for antigen stimulation, and, through its five Krüppel-type zinc fingers, interacts with DNA, coactivators, and histone deacetylases to provide stable chromatin modifications that maintain differentiation, maturation, and plasma cell longevity (15, 60, 61). Blimp-1 is likely involved in differentiation of monocytes and neutrophils from progenitor myeloid lineage cells in bone marrow, is known to be expressed in long-lived T cells (43) and circulating monocytes and neutrophils (15), and appears to be required for proper adherence and phagocytic functions of myeloid lineage cells (15). In light of the pronounced upregulation of Blimp-1 gene expression in Dex-treated neutrophils and the interactions known to occur between Blimp-1 and chromatin, it may be relevant that 64 additional transcription factor genes and 29 genes encoding chromatin-remodeling proteins were identified in the microarray experiment as putatively differentially expressed in the various steroid-treated neutrophils. Any of these gene expression changes could logically support genomic reprogramming for extended survival of these otherwise short-lived leukocytes.
All of the data discussed thus far point to key genomic effects of Dex-activated GR in the regulation of neutrophil apoptosis. However, GR also appeared to have nongenomic effects with respect to neutrophil apoptosis, as was clearly demonstrated by patterns of TFAR-19 gene expression in Dex- and RU486-treated cells. The TFAR-19 gene (also coded PDCD5 for programmed cell death 5) encodes a cytosolic protein that rapidly translocates to and accumulates in the nucleus of cells just before apoptosis-related phosphatidylserine flipping in the plasma membrane and chromosomal DNA fragmentation in the nucleus (18). Although its precise role in programmed cell death is not understood, transfer of anti-TFAR-19 antibodies (53) and introduction of siRNA against TFAR-19 into cultured mammalian cell lines (17) significantly delay apoptosis induced by chemotherapy or by proapoptotic Bax overexpression. In the current study, both Dex and RU486 totally inhibited TFAR-19 mRNA expression in bovine neutrophils, suggesting that GR actions on abundance of this transcript might be predominantly nongenomic and occur rapidly in the cytosol.
Another theme revealed through the microarray experiment and validated by real-time RT-PCR in this study was one of Dex-induced inhibition of the neutrophil tissue-remodeling system, comprising profibrotic TGF-ß-associated proteins and the tissue ECM-degrading gelatinase MMP-9. An additional 20 genes that encode ECM proteins and various other proteins that regulate their deposition and structure were also detected as putatively differentially expressed in the microarray experiment. However, the number of affected TGF-ß-related genes was striking and thus pursued for initial validation. TGF-ß comprises a family of pleiotrophic cytokines that are immunosuppressive and proapoptotic and promote fibrosis and vascularization by inhibiting expression of MMPs and enhancing the production, deposition, and spatial conformation of tissue ECM proteins (38, 71, 75). While an intact TGF-ß system is necessary for wound healing of otherwise normal tissue, it becomes pathogenic if its regulation is aberrant. The various isoforms of TGF-ß mediate their effects in cells by interacting with several binding proteins, an important one being TGF-ßRIII (also known as betaglycan). Betaglycan/TGF-ßRIII is expressed predominantly on endothelial cells but also on several other cell types (including leukocytes). It is a membrane-anchored TGF-ß-binding protein that presents TGF-ß directly to a type II signaling receptor, which by itself has limited ability to bind the cytokine (reviewed in Ref. 38). Thus betaglycan/TGF-ßRIII expression is critical for the synthesis of ECM components (collagens, fibronectin, elastin, and TFGßIP) and for angiogenesis in response to TGF-ß during normal wound healing (75). However, betaglycan/TGF-ßRIII also can be shed by cells to act as a potent blocker of TGF-ß overstimulation and inhibitor of pathological fibrotic processes (38, 63). While requiring proof through further experimentation, it is possible that the downregulation of betaglycan/TGF-ßRIII in response to Dex effects the same result as receptor shedding by inhibiting excessive deposition of tissue ECM proteins in the presence of TGF-ß. One such TGF-ß-inducible ECM protein is TGFßIP, which also acts as an important anchoring protein for integrins expressed on the plasma membranes of adherent cells (3). Uncontrolled upregulation of TGFßIP leads to diseases of abnormal ECM protein deposition and cell adhesion, including ovarian endometriosis (4), several corneal dystrophies that cause visual impairment (3, 27), cancer (58), and fibrous airway disease (38). That Dex and RU486 both inhibited betaglycan/TGF-ßRIII gene expression and Dex downregulated TGFßIP gene expression in neutrophils of the current study suggest that ligand-bound GR has rapid nongenomic as well as more gradual genomic regulatory effects on the TGF-ß system, which might act to prevent tissue fibrosis by neutrophils recruited to participate in wound healing (46, 67). Furthermore, GR-induced downregulation of the TGF-ß system in neutrophils is expected to be anti-apoptotic, which is congruent with the theme of extended cell survival revealed through the expression profiles of numerous other apoptosis regulatory genes discussed above.
It would make biological sense that a GR-induced downregulation of the TGF-ß system would be balanced by simultaneous inhibition of the MMP system, as we observed for MMP-9, so that risk of excessive tissue ECM degradation by neutrophil influx to wounded tissues is reduced (67). In addition, MMP-9 is known to process the chemokine IL-8 into a more potent form that amplifies its neutrophil-attracting and -degranulating properties (69). That ligand-activated GR inhibited expression of MMP-9 in our treated bovine neutrophils and depressed its activity in 4- and 8-h cultures of the cells at the same time that it inhibited betaglycan/TGF-ßRIII and TGFßIP gene expression suggests that acute downregulation of neutrophil recruitment, ECM-degrading capacity, and ECM protein deposition may be key anti-inflammatory properties of Dex that promote wound healing without causing damage from excessive fibrosis. If true, this would support a wound-healing role for glucocorticoids through their effects on neutrophils, which is congruent with six classical wound-healing genes that were identified as putatively differentially expressed in the cells through the microarray experiment, as well as the validated Dex-induced upregulation of PPAR-, LC3, and BPI (described above), which could promote wound healing by controlling inflammation, preventing fibrosis, and promoting bacterial uptake and killing (45, 46).
Finally, the real-time RT-PCR expression patterns for a diverse group of genes shown to be potently Dex responsive and inhibitable by RU486 indicated that glucocorticoids induce a complex genomic reprogramming of neutrophils that, overall, is cyroprotective, potently anti-inflammatory, and bactericidal. For example, the Dex-induced increases in cytoprotective MT-II (50-fold induction) (24, 32), anti-inflammatory PPAR- (13-fold induction) (8, 65), antimicrobial BPI (2,900-fold induction) (1, 12, 36), and endosomal/autophagosomal vesicle trafficking LC3 (4.8-fold induction) (33, 34) in Dex-treated vs. untreated neutrophils hinted of massive coordination by nuclear GR of blood neutrophil development into longer-lived cells with profoundly augmented antimicrobial capacity and reduced tissue-damaging potential. The significant increase in uptake of latex beads by the Dex-treated compared with untreated and RU486-treated neutrophils substantiates the gene expression profiles we observed. It thus appears that these terminally differentiated leukocytes have been misunderstood in the past, because they appear to be capable of further maturation under the influence of glucocorticoids. This may be needed for balanced innate immune activity during episodes of physiological stress and is likely a significant and hitherto unexpected benefit of anti-inflammatory steroid therapy.
In the past, immunologists have propagated a misconception that, because neutrophils are short-lived, terminally differentiated, professional phagocytes possessing highly condensed nuclei and relatively few mitochondria, and participate nonspecifically in innate immune defense, it is neither necessary nor possible for these cells to induce gene expression (46, 48). On the contrary, the current study as well as others (10, 40, 48, 67) have established that neutrophils express abundant mRNA species and are capable of extensive, rapid, and complex changes in gene expression involving significant proportions of all mRNAs present in the cells. We have demonstrated here that neutrophils show remarkable plasticity in response to a single anti-inflammatory stimulant, glucocorticoid, possibly reflecting the cells' increasingly recognized and diverse roles as deciders of their own fate, immune response decision shapers, tissue remodelers and wound healers, and, of course, pathogen destroyers (46). Given these observations and possibilities, development of new GR ligands with specific gene or gene cluster regulatory effects may prove beneficial in the treatment of various forms of inflammation while reducing or eliminating many of the unwanted side effects that occur with available glucocorticoid therapies.
GRANTS
Portions of this study were supported by funds from the Michigan Agricultural Experiment Station Project No. MICL02035 [for J. L. Burton's participation in United States Department of Agriculture (USDA) Multistate Research Project NC-1010], USDA-IFAFS Grant No. 2001-52100-11211 (for NBGFC microarray development, microarray analysis, and the real-time RT-PCR work), and by USDA National Research Initiative Grant No. 2001-35204-10798 (for Western blot analyses of GR and the CD62L gene expression work) and Grant No. 2006-35205-16706 (for support of the May 2006 2nd International Symposium on Animal Functional Genomics and presentation of our data there), both from the USDA Cooperative State Research, Education, and Extension Service.
ACKNOWLEDGMENTS
We are grateful to Ling-Chu Chang and Kelly Buckham for assistance with animal handling, blood collections, and neutrophil preparations for all assays and Bob Kreft and staff at the MSU Dairy Teaching and Research Facility for the excellent care of the animals used in this study.
Present addresses: S. A. Madsen-Bouterse, Perinatology Research Branch, NICHD/NIH/DHHS, Detroit, MI 48201; and G. J. M. Rosa, 456 Animal Sciences Building, University of Wisconsin-Madison, Madison, WI 53706.
FOOTNOTES
Address for reprint requests and other correspondence: J. L. Burton, Immunogenetics Laboratory, 1205E Anthony Hall, Dept. of Animal Science, Michigan State Univ., East Lansing, MI 48824 (e-mail: burtonj{at}msu.edu).
Article published online before print. See web site for date of publication (http://physiolgenomics.physiology.org).
1 The 2nd International Symposium on Animal Functional Genomics was held May 16–19, 2006, at Michigan State University in East Lansing, MI, and was organized by Jeanne Burton of Michigan State University and Guilherme J. M. Rosa of University of Wisconsin, Madison, WI (see meeting report by Drs. Burton and Rosa, Physiol Genomics 28: 1–4, 2006). 
REFERENCES
