INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Early growth-response factor 1 (Egr-1), also known as nerve growth factor induced-A (NGFI-A), Krox-24, ZIF268, ETR103, and TIS8, is an 80- to 82-kD phosphorylated protein discovered independently by a number of laboratories searching for factors regulating cell growth and proliferation (1, 2). The gene for Egr-1, located on the q31.1 "cytokine cluster" region of chromosome 5 in humans, encodes a zinc-finger protein of 533 amino acids in length. Egr-1 is an "immediate-early response" protein because it is rapidly and transiently induced by a large number of growth factors, cytokines, and injurious stimuli (reviewed in [3]).

It has been hypothesized that Egr-1 functions in a regulatory manner by linking growth and injurious stimuli to the induction of genes directing the expression of effector molecules that ultimately results in pathology (4, 5). For example, a variety of proinflammatory events can induce the expression of Egr-1 (6); once induced, Egr-1 interacts with a G+C-rich consensus-binding site (GCG[T/G]GGGCG) on DNA and is known to alter the transcriptional regulation of several disease-associated effector genes in vitro, including tumor necrosis factor- (TNF-) (7), 5-lipoxygenase (Alox-5), platelet-derived growth factor-A (PDGF-A), PDGF-B, fibroblastic growth factor-2 (FGF-2), macrophage colony-stimulating factor (M-CSF), transforming growth factor-1 (TGF-1), copper-zinc superoxide dismutase, interleukin-2 (IL-2), p53, and intracellular adhesion molecule-1 (ICAM-1) (reviewed in [5]).

Given Egr-1's ability to be induced by many stimuli associated with airway disease, its ability to regulate genes involved in the pathogenesis of airway disease, and its hypothetical role in the pathogenesis of diverse inflammatory and disease processes, we speculated that Egr-1 may be an important transcription factor mediating allergen-induced airway inflammation and reactivity. To explore this hypothesis, we studied Egr-1 knock-out (KO) mice in a well-established ovalbumin (OVA)-induced model of airway inflammation and reactivity and compared them to wild-type (WT) littermate control mice.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animals: Immunization and Challenge Protocol

Egr-1 KO mice were generated and maintained on a 129 × C57BL/6 background as previously described (8). To minimize environmental effects mice were housed in isolation cages in a specific pathogen-free (SPF) environment, and specimens from sentinel animals were routinely screened to ensure SPF status; these findings confirmed the SPF status of all animal cohorts. Commercial sterilized pelleted mouse food and water were provided ad libitum. Because female KO mice are infertile, heterozygous mice were bred and offspring genotyped at the Egr-1 locus by a standard polymerase chain reaction (PCR) protocol (8). Male mice at approximately 4-5 wk of age were entered into the protocol and studied 3 wk later; littermates were used as control mice in all experiments. Mice were sensitized and challenged with OVA as detailed previously (9). In brief, on Day 0, all mice were immunized via intraperitoneal injection with 10 µg chicken OVA (Grade III; Sigma Chemical Co., St. Louis, MO) mixed with 1 µg Al(OH)₃ (J. T. Baker Chemical, Phillipsburg, NJ) in 0.2 ml phosphate-buffered saline (PBS). An identical booster injection was given on Day 7. On Days 14-20, mice were exposed to either aerosolized 6% OVA dissolved in PBS or PBS alone for 25 min/d. All mice were studied on Day 21 of the protocol, approximately 24 h after the last OVA exposure.

Measurement of Airway Resistance and Responsiveness to Methacholine

Airway resistance and responsiveness to methacholine were measured in anesthetized mice with a sealed constant-mass plethysmograph as previously described in detail (10). In brief, under surgical anesthesia with pentobarbital sodium (100 mg/kg intraperitoneally), a 19-gauge tracheostomy cannula was surgically placed and secured. The right internal jugular vein was catheterized with a saline-filled Silastic catheter (0.06 cm o.d.) and attached to a 0.1-ml Hamilton microsyringe for methacholine (Sigma Chemical Co.) administration. A thoracotomy was performed so that pleural pressure was atmospheric pressure. The mice were placed in a sealed constant-mass/temperature plethysmograph chamber and attached to a rodent ventilator (Harvard Apparatus, Division of Ealing Scientific, Natick, MA). Ventilator settings were 150 breaths/min with a tidal volume of 5-6 µl/g and a positive end expiratory pressure of 3-4 cm H₂O; these settings have been shown to maintain normal PCO₂. Pulmonary resistance was determined from measured transpulmonary pressure and lung volume change. A dose-response curve to methacholine was obtained by administering sequential increasing doses from 33 to 3,300 µg/kg. Enough time was allowed between doses for pulmonary resistance to return to within 10% of baseline value. Each animal's dose-response curve was log transformed and subjected to regression analysis to calculate the dose required for a 1.5-fold increase in pulmonary resistance (ED₁₅₀ RL).

Bronchoalveolar Lavage Cell Counts and Differentials

Bronchoalveolar lavage (BAL) was performed after resistance measurements by instillation of 2 ml of PBS into the lung and retrieval of the fluid with gentle suction. The lavagate was centrifuged at 600 × g for 10 min, the supernatant separated from the cell pellet, and aliquots were frozen in liquid nitrogen and stored at 80° C. The cell pellets were resuspended in 1 ml of PBS at 4° C and counted in a hemocytometer by an investigator blinded as to genotype and treatment group. Differential counts were tabulated from cytocentrifuge prepared slides that were stained with Wright-Giemsa (HEME 3; Biochemical Sciences, Inc., Swedesboro, NJ). Protein extracts from BAL cells were prepared as previously detailed (11). Aliquots were immediately frozen on dry ice and stored at 80° C.

Determination of Blood Cell Counts, Differentials, and Serum Antibody Titers

After BAL mice were sacrificed by exsanguination by cardiac puncture; whole blood was collected for total leukocyte count. Blood smears for differential counts were preserved with CytoPrep (Fisher Scientific, Pittsburgh, PA) and stained with Wright-Giemsa (HEME 3). Serum was separated from whole blood after clotting by centrifugation, frozen in liquid nitrogen, and stored at 80° C. Total immunoglobulin E (IgE) and OVA-specific IgE were measured in serum by enzyme-linked immunosorbent assay (ELISA) as previously described in detail (12).

Lung Pathology

After exsanguination by cardiac puncture, the lungs were removed and separated by rapid dissection (< 3 min). The left lung was immediately frozen in liquid nitrogen, and the right lung was fixed for 4 h in 4% paraformaldehyde and stored in 70% ethanol at 4° C. The fixed lung was embedded in paraffin, cut in 5-µm sections, stained with hematoxylin and eosin (H&E) and periodic acid-Schiff (PAS)/alcian blue (pH 2.5), and examined by light microscopy in a masked fashion. A semiquantitative grading scale was used to score the degree of inflammation: 0 for absent, 1 for mild, 2 for moderate, and 3 for severe. Three lung compartments (perivenular, peribronchial, and intraalveolar) were each examined for two cellular components (granulocytic and mononuclear cells). In addition, the degree of mucous cell hyperplasia and number of intraalveolar giant cells were also graded.

Bone Marrow-derived Mast Cells

Mast cells were derived from in vitro culture of bone marrow progenitors from WT and KO mice as previously described (13). In brief, bone marrow cells were isolated from the femurs and tibias of mice following lung harvest. Unfractionated bone marrow cells were suspended in 25-ml flasks (Corning, Corning, NY) at a concentration of 0.5-1 × 10⁶ cells/ml in RPMI 164 medium containing 100 U/ml penicillin, 100 µg/ml streptomycin, 10 µg/ml gentamycin, 2 mM L-glutamine, 0.1 mM nonessential amino acids, 10% fetal calf serum, and 100 U/ml of IL-3 (R&D Systems, Minneapolis, MN). The purity of the cultured cells was assayed by staining cytocentrifugation slides with toluidine blue dye. Cell populations used for subsequent analysis were > 99% pure and analyzed at 3-4 wk of culture.

Reverse Transcription-PCR, Northern Blot, and Western Blot Analysis

Total RNA was extracted from frozen lung and mast cells with the Trizol reagent in accordance with the manufacturer's instructions (Life Technologies, Inc., Gaithersburg, MD). Samples were aliquoted and stored at 80° C.

Reverse transcription (RT)-PCR was performed with mouse TNF-, -actin, TGF-1, IL-2, p53, M-CSF, and c-myc Amplimer primer sets from Clontech Laboratories, Inc. (Palo Alto, CA). Alox-5 primers were (5'-CACGGGGACTACATCGAGTT) and (5'-AACCTCACATGGGCTACCAG). Egr-1 and Egr-Neo primers were as previously described (8). RT reactions were performed with 2 µg total RNA, 1 µg random hexamers (Life Technologies, Inc.), 50 nmol dNTP, 25 U rRNasin ribonuclease inhibitor (Promega), 200 U M-MLV RT (Promega, Madison, WI) in 25 µl total volume consisting of 50 mM Tris-HCl (pH 8.3), 75 mM KCl, 3 mM MgCl₂, and 10 mM DTT at 37° C for 60 min. PCR reactions were performed with 4 µl 1:5 diluted RT reactions, 40 nmol dNTP, 2 µl Amplimer primers, 2 U Taq polymerase (Promega) in 50 µl total volume consisting of 50 mM KCl, 10 mM Tris-HCl (pH 9.0), 0.1% Triton X-100 and 1.5 mM MgCl₂. The same RT reaction mixture was analyzed by each primer set for comparisons. Depending on the primer set, the annealing temperature and number of PCR cycles varied between 57-62° C and 27-30, respectively. PCR products were separated on a 1.5-2% agarose gel, photographed, and quantitated with Gel-Pro Analyzer densitometry software (Media Cybernetics, Silver Spring, MD).

RNA samples (10 µg) were separated by a 1% formaldehyde/agarose gel and transferred overnight to a Hybond nylon membrane (Amersham, Arlington Heights, IL). The membrane was UV cross-linked (Stratalinker UV Crosslinker, Stratagene, La Jolla, CA) and hybridized with cDNA probes, random-primer labeled (Megaprime DNA labeling Systems, Amersham) with [-³²P]dCTP (DuPont, NEN, Boston, MA), using ExpressHyb Hybridization solution (Clontech Laboratories) according to the manufacturer's instructions. The blots were washed with 0.1× SSC several times at 65° C and exposed to X-ray film at 80° C for 1-7 d. Autoradiographs were scanned and quantitated with Gel-Pro Analyzer densitometry software (Media Cybernetics).

Western blots were performed as previously described in detail (11). The membranes were probed sequentially with polyclonal antibodies to TNF-, Egr-1, and p65 (Santa Cruz Biotechnology, Santa Cruz, CA) at dilutions of 1:10,000, followed by enhanced chemiluminescent detection (Amersham) with 1:10,000 horseradish peroxidase-linked secondary antiserum. Blots were exposed to X-ray film for 1- 30 min, scanned, and quantitated with Gel-Pro Analyzer densitometry software (Media Cybernetics).

Statistical Analysis

Computations were performed with the JMP 3.15 statistical package or SAS version 6.12 (SAS Institute Inc., Cary, NC). For most analyses an ANOVA with Tukey-Kramer HSD multiple comparison test was used to assess differences among the genotypes and treatment groups. ANOVA and Wilcoxon tests were performed on the IgE data. Tests were conducted at the 5% significance level, and results are expressed as mean ± SD or ± SEM. To avoid the potential inflation of the type I error rate that can result from separate analyses of the eight inflammation scores, an omnibus test of group differences was performed. Given the relatively small sample size and apparent trend in lung inflammation, the null hypothesis of no group differences was tested against a constant shift alternative (i.e., a constant upward or downward difference across the eight inflammation scores). Because the inflammation scores are ordinal, this analysis was conducted with a proportional-odds model. The analysis also accounted for the correlation among the eight inflammation scores. A complementary analysis based on the Wilcoxon test of the averages of the eight inflammation scores was also performed and reported.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Egr-1 Modulates TNF- Expression

To identify genes regulated by Egr-1 we used RT-PCR to screen lung tissue from WT and KO mice after OVA/OVA protocol for differences in steady-state mRNA levels. We chose to screen seven genes based on in vitro data suggesting that Egr-1 may play a role in their regulation and the potential of these genes to alter airway inflammation and reactivity. The chosen genes included TNF-, ALOX-5, TGF-1, IL-2, M-CSF, p53, and c-myc. Egr-1, Egr-Neo, and -actin were used as controls. Of all these genes only TNF- showed a consistent and significant difference between WT and KO animals (Figure 1A), with densitometry analysis indicating a fall of approximately 38% in mRNA levels in the KO mice (relative densitometry 50.9 ± 10%) compared with WT controls (82.7 ± 22%, p < 0.01, n = 6 each). As expected, Egr-1 was detected only in WT mice and Egr-Neo only in KO mice. No differences in Egr-1 mRNA levels could be detected between OVA/PBS and OVA/OVA groups in whole lungs at the time of organ harvest (Figure 1B).

View larger version (77K): [in this window] [in a new window]   View larger version (27K): [in this window] [in a new window]   View larger version (48K): [in this window] [in a new window]   View larger version (94K): [in this window] [in a new window]   Figure 1.   Gene expression by RT-PCR, Western blot, and Northern blot analysis. (A) Steady-state expression of Egr-1-responsive genes was measured by RT-PCR using total RNA extracted from frozen whole lung. Representative examples are shown with base pair (bp) markers and positive control bands for each gene. Only primers for TNF- showed a consistent difference between WT and KO mice, with KO mice having approximately 38% less mRNA by densitometry (p < 0.01, n = 6). (B) Egr-1 mRNA was measured in whole lung by RT-PCR and OVA/ PBS and OVA/OVA protocols were compared. No difference in Egr-1 levels could be detected at Day 21. (C) Western blot analysis of BAL cell protein sequentially screened with antibodies to TNF-, p65, and Egr-1. Densitometry confirms approximately 48% less TNF- protein in KO mice than in WT control mice (p < 0.02). In contrast, levels of p65 protein are more uniform and not significantly different. Egr-1 control mice are shown. (D) Northern blot analysis of bone marrow-derived mast cells stimulated with IL-3. Steady-state levels of TNF- message are approximately 33% less in KO mice than in WT control mice (p < 0.02). In contrast, TGF-1 and -actin control mice show no differences.

Given the quantitative limitations of our screening RT-PCR technique, additional studies were performed to confirm differences in TNF- expression. First, Western blot analysis was performed on BAL cells to detect differences in TNF- protein. KO mice had a 48% decrease in protein levels (p < 0.01) (Figure 1C). In contrast, levels of the Rel A (p65) component of nuclear factor-B (NF-B) remained relatively constant with respect to genotype without a significant difference. As expected, Egr-1 expression was absent in KO mice.

To further demonstrate the contribution of Egr-1 to TNF- expression, bone marrow mast cells were harvested, grown in tissue culture with IL-3, and measured for TNF- mRNA by Northern blot analysis. Again, KO mice had a 33% decrease in steady-state TNF- mRNA levels compared with WT controls (p < 0.02) (Figure 1D). In contrast, TGF-1 and control gene -actin remained unchanged with respect to genotype. Not all conditions resulted in differentially expressed TNF- levels. For example, when mast cells were stimulated with IgE, no genotype-related differences in TNF- levels could be detected between KO and WT mice (data not shown). ELISA for TNF- on BAL supernatants showed very low levels of protein at Day 21 regardless of OVA or PBS treatment, as previously described (9). No obvious difference between OVA/ PBS and OVA/OVA groups in TNF- could be discerned (data not shown).

Egr-1 KO Mice Have Elevated Total Serum IgE

Both total serum IgE and OVA-specific IgE were increased in response to aerosolized OVA treatment, consistent with previous reports (p < 0.01) (Figure 2) (14). It is interesting that WT mice had lower levels of total IgE than KO mice, p < 0.05 (Figure 2A). This trend was true for OVA-sensitized but PBS-challenged mice, 196 ± 42 versus 643 ± 131 ng/ml, respectively, as well as for OVA-sensitized and OVA-challenged mice, 912 ± 221 versus 1599 ± 219 ng/ml, respectively (n = 6 to 13). The trend was similar for OVA-specific IgE, with increased levels in the KO groups; however, the differences did not reach significance.

View larger version (23K): [in this window] [in a new window]   Figure 2.   Total and OVA-specific IgE measurement by ELISA. (A) Total IgE levels are expressed in nanograms per milliliter with mean ± SEM recorded (n = 6-13). Total IgE increases in both groups of mice following OVA airway challenge, p < 0.01. Moreover, KO mice had higher levels at baseline and following OVA challenge, p < 0.05. (B) OVA-specific IgE levels expressed as relative OD units with mean ± SEM recorded (n = 6- 13). As with the total IgE study, OVA-specific IgE increases in both groups following OVA challenge, p < 0.01. KO mice had higher levels of OVA specific IgE at baseline and following OVA challenge, but these values were not significant.

Egr-1 KO Mice Have Decreased Airway Responsiveness to Methacholine

A significant difference in baseline (OVA/PBS) airway responsiveness between WT and KO mice was detected (Figure 3). The Egr-1 KO mice were significantly hyporesponsive compared with WT controls, with a log ED₁₅₀ of 2.54 ± 0.25 and 2.13 ± 0.10, respectively (n = 7 and 9, p < 0.005). A similar difference was also noted in the OVA/OVA-treated group. Egr-1 KO mice were significantly hyporesponsive compared with WT controls, with a log ED₁₅₀ of 2.44 ± 0.22 and 2.07 ± 0.18, respectively (n = 10 and 9, p < 0.005). The 129 × C57BL/6 background of the Egr-1 KO mice was generally hyporesponsive to methacholine in our model and although OVA/OVA group had a lower log ED₁₅₀ compared with the OVA/PBS group, the differences did not reach significance.

View larger version (11K): [in this window] [in a new window]   Figure 3.   Airway responsiveness to methacholine. The log ED150 is plotted with mean ± SD for the OVA/PBS and OVA/OVA groups comparing WT and KO mice. The KO mice are significantly hyporesponsive compared with WT control mice in both OVA/PBS and OVA/OVA groups (p < 0.005). The 129 × C57BL/6 background of the Egr-1 KO mice is hyporesponsive to methacholine; thus, no difference in log ED150 could be detected between OVA/PBS and OVA/OVA groups.

Egr-1 Genotype Does Not Affect BAL Eosinophilia or Blood Leukocyte Levels following OVA Challenge

We detected no abnormalities in the cellular composition of BAL fluid from OVA/PBS-challenged KO mice compared with WT controls. BAL fluid from both WT and KO mice contained approximately 92% alveolar macrophages, 8% lymphocytes, and < 1% eosinophils or neutrophils, without significant differences between genotypes (n = 10 each group) (Figure 4A). After OVA/OVA challenge, both groups of mice showed a dramatic increase in eosinophil number and percentage, but there were no genotype-related differences between mice (Figure 4B). Total BAL cell number also remained similar between WT and KO groups (data not shown) without any significant difference. Peripheral blood cell counts (data not shown) and differentials (Figure 5) were also within normal limits in both WT and KO mice and did not change significantly after OVA/OVA treatment.

View larger version (21K): [in this window] [in a new window]   Figure 4.   BAL lavage cell differentials. Cytocentrifuge-prepared slides were stained with Wright-Giemsa, counted, and tabulated by cell type. (A) OVA/PBS, WT versus KO mice. (B) OVA/OVA, WT versus KO mice. Mean values ± SEM are recorded, n = 10 each. Percentage of eosinophils increased dramatically with OVA nebulization; however, no differences in BAL differentials or total cell count (data not shown) could be detected between WT and KO mice.

View larger version (24K): [in this window] [in a new window]   Figure 5.   Peripheral leukocyte differential. Blood smears were stained with Wright-Giemas, counted, and tabulated by cell type. Unidentifiable cells were classified as degenerate cells. (A) OVA/PBS, WT versus KO mice. (B) OVA/OVA, WT versus KO mice. Mean values ± SEM are recorded, n = 10 each. No differences in peripheral leukocyte differentials or total cell count (data not shown) could be detected.

Egr-1 Genotype Does Not Affect Airway Inflammation following OVA Challenge

Histological examination of WT and KO lungs challenged with the OVA/PBS protocol revealed normal findings without any noted differences between genotypes (Figures 6A and 6B). Particular attention was given to alveolar architecture, peribronchial smooth muscle, interstitial connective tissue, and presence or absence of inflammatory cells. After OVA/OVA challenge, mice developed a characteristic inflammatory response throughout the lung consisting of focal cellular infiltrates centered on terminal bronchioles, alveolar ducts, and small blood vessels (Figures 6C and 6D). Cellular infiltrates also occurred in the perivascular and periairway interstitium, and consisted largely of lymphocytes and eosinophils. Mucous cell hyperplasia and hypertrophy were noted in bronchi, and in some airspaces giant cells of foreign body type were seen. Semiquantitative inflammation scoring of eight histological parameters are plotted as mean ± SEM (Figure 7). Although each of the eight inflammation scores in the KO mice is slightly less, statistical analysis of the data yielded a z-statistic of 0.50 (p > 0.60), indicating that the WT and KO groups are not significantly different across the eight inflammation scores (n = 7 for each group).

View larger version (125K): [in this window] [in a new window]   Figure 6.   H&E-stained lung sections. (A) WT lung after OVA/PBS protocol showing normal lung histology (×100). (B) KO lung after OVA/PBS protocol showing normal lung histology (×100). (C) WT lung after OVA/OVA protocol showing characteristic eosinophilic inflammatory process (×200). (D) KO lung after OVA/OVA protocol showing inflammatory process indistinguishable from that of WT lung (×200). (E ) WT lung after OVA/OVA protocol (×400). (F ) KO lung after OVA/ OVA protocol (×400).

View larger version (30K): [in this window] [in a new window]   Figure 7.   Inflammation score. H&E- and PAS/alcian blue-stained lung specimens from WT and KO mice were compared by a semiquantitative scoring system. Eight parameters based on cellular type and location were examined. A score of 0 indicates the absence of inflammation; a score of 3 indicates severe inflammation. Mean values ± SEM are plotted, n = 7 each. No significant difference in any particular parameter or overall inflammation score was detected between WT and KO mice.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Data from in vito studies indicate that a number of genes thought to be part of the inflammatory process may be regulated by Egr-1 (15). However, to our knowledge only three genesLH-, tissue factor, and apoprotein A1have been definitely linked to Egr-1's trans-activation potential in intact animal models (5, 8). For example, the mouse and human LH- gene promoter contains one Egr-1 binding site between 50 and 42 bp relative to the transcription initiation site, and female Egr-1 KO mice are infertile due to luteinizing hormone- (LH-) deficiency (16). The tissue factor gene promoter contains three Egr-1 binding sites between 109 and 55 bp, and its expression is diminished in Egr-1 KO mice exposed to hypoxemia (17). As a result, the KO mice deposit less fibrin in the pulmonary vasculature than WT mice and are more resistant to hypoxemia-associated thrombotic events. In a mouse model of nephrotic syndrome, levels in apoprotein A1, whose gene contains two Egr-1 binding sites between 221 and 180 bp, in Egr-1 KO mice were half those in the WT mice (18). On the basis of these data, we hypothesized that Egr-1 plays a role in the regulation of TNF- expression and hence (19) could lead to differences in airway inflammation and reactivity. To test our hypothesis we compared Egr-1 KO mice with WT control mice in a well-established model of OVA-induced airway inflammation and reactivity.

In this model, systemic sensitization to OVA, by intraperitoneal injection, followed by repeated OVA nebulization to the airways generally results in pulmonary eosinophilia, increased IgE and leukotriene production, and airway hyperresponsiveness (20, 21). We now show that when Egr-1 KO mice are compared with WT control mice in this model they are significantly different in three respects: (1) KO mice expressed lower levels of TNF- mRNA and protein; (2) KO mice had elevated levels of IgE; and (3) the KO mice had airways that were hyporesponsive to methacholine challenge. Differences in IgE level and airway responsiveness occurred at baseline (OVA/PBS) and after OVA/OVA treatment, suggesting that airway inflammation was not essential to these observations. Although we had speculated that Egr-1 KO mice would have diminished airway inflammation compared with WT mice when challenged with our OVA/OVA protocol, this was not the case. Despite Egr-1's suspected role in the regulation of genes involved in cell differentiation, growth, apoptosis, and inflammation and our finding of diminished TNF- expression, no differences in airway inflammation or BAL cell composition could be detected between WT and KO mice. These data demonstrate that the Egr-1 transcriptional pathway plays a functional role in the regulation of TNF- expression, IgE production, and airway reactivity in mice. However, even in the absence of this transcription factor, compensatory mechanisms allow for the full appreciation of airway inflammation that occurs after allergen stimulation.

Suppressed levels of TNF- mRNA and protein were not unexpected. The effect of Egr-1 genotype on the expression of TNF- is probably related to Egr-1's ability to bind and activate the TNF- promoter. Kramer and coworkers have identified a G+C-rich Egr-1-responsive element on the TNF- promoter between 170 to 160 bp relative to the transcriptional start site (7). Deletion of this site resulted in a reduction in promoter activity, and cotransfection studies overexpressing Egr-1 demonstrated a 2- to 4-fold augmentation in TNF- promoter activity. These data are consistent with the approximate 40% reduction in TNF- mRNA and protein we observed in the KO mice. In Egr-1 KO mice these Egr-1 binding sites in the promoter would be unoccupied, or perhaps occupied by a less potent transcription factor, such as Sp1, resulting in decreased TNF- transcription and expressed protein (22). Although no differences in TNF- or Egr-1 could be detected between OVA/PBS and OVA/OVA groups on Day 21 of our protocol, we speculate that levels of TNF- and Egr-1 protein change over time in certain tissue compartments as airway inflammation evolves. Furthermore, since TNF- itself may serve as an inducer of Egr-1 expression (6), it is possible that these two proteins could augment the production of each other in a positive-feedback fashion that amplifies the effects of this inducible pathway in vivo. However, even with this amplification, the maximal affect on TNF- expression was only about 40% in vivo or in bone marrow-derived mast cells stimulated by IL-3. The latter experiments are particularly informative, as they show that not all stimuli resulted in differences in TNF- expression. For example, when bone marrow-derived mast cells were stimulated by IgE, TNF- mRNA levels increased dramatically, yet no differences in TNF- mRNA could be detected between WT and KO mice. This finding suggests that only certain stimuli (e.g., IL-3) lead to TNF- transcription activation via the Egr-1 pathway, with other stimuli (e.g., IgE cross-linking) augmenting TNF- transcription through an Egr-1-independent pathway.

The mechanism responsible for the airway hyporesponsiveness observed in Egr-1 KO mice at baseline may be related to decreased TNF- expression. TNF- is a potential mediator of airway inflammation and reactivity in asthma and other forms of airway disease (19). TNF- is increased in the airways of mice with asthma (23), stimulates the release of inflammatory mediators (24), alters the late asthmatic inflammation reaction through its effects on cell adhesion molecules (25), and may induce bronchial hyperresponsivenss when exogenously administered (26). TNF- may also have indirect effects on airway smooth muscle function through its ability to release inflammatory mediators or its effects on the autonomic nervous system. For example, TNF- has been reported to increase airway responsiveness through the release of leukotrienes and other bronchoconstrictors (27). Prejunctional effects of TNF- on the parasympathetic nerve pathway may also increase responsiveness of bronchial tissue (28), and TNF- can alter -adrenergic signal transduction in airway smooth muscle cells (29). Furthermore, several studies have suggested that TNF- modulates calcium transients through effects on G proteins and other signal-transduction components, leading to increased airway smooth muscle tone and hyperresponsiveness to agonists such as methacholine (30). Since a decrease in the level of activation of any or all of these pathways would decrease airway responsiveness, these data are consistent with our observations.

Not all studies have demonstrated altered airway responsiveness mediated by TNF-. For example, systemic administration of endotoxin leads to airway hyperresponsiveness (AHR) and increased TNF- expression in mice; however, inactivation of TNF- with antibodies to TNF- had no effect on AHR (31). In a model more closely related to our own, treatment with antibodies to TNF- partially decreased AHR to OVA, but the decrease was not significant (32). Consistent with these findings, we have been unable to detect differences in AHR between TNF KO mice and WT controls in our model (Silverman and coworkers, unpublished data). We believe that the differences in TNF- detected between Egr-1 KO and WT mice are unlikely to completely explain the observed differences in airway reactivity and that other mechanisms are likely to be involved. The fact that differences were detected in both OVA/OVA and OVA/PBS groups suggests that inflammation is not essential to our findings and the mechanism responsible for the differences in airway reactivity is more likely to be related to an intrinsic property of airway smooth muscle function. Airway response to methacholine can be modified by a number of pathways that control the composition or cycling rate of actin-myosin cross-bridging, including various kinases, phosphatases, ion channels, and receptors, as well as genes regulating myosin isoform and the autonomic nervous system. The Egr-1 transcriptional pathway may alter the expression of any one or more of these diverse genes, thereby altering airway smooth muscle tone and response to methacholine.

The role of Egr-1 in the regulation of serum IgE is difficult to explain. Egr-1 KO mice had significantly higher levels of total IgE than WT mice at baseline and after OVA nebulization. A similar trend was observed for OVA-specific IgE. The OVA/ OVA protocol increased total and OVA-specific IgE levels compared with OVA/PBS, but did not seem to alter differences between WT and KO mice that were present following the OVA/PBS protocol. The effect of Egr-1 on the expression of antibodies was unexpected, but its mechanism, although unknown, is potentially important with respect to human disease. Although it is possible that Egr-1 directly alters immunoglobulin gene transcription or isotype switching, we believe it is more likely that Egr-1 regulates these events indirectly through the regulation of other genes. In other words, Egr-1 may alter the expression of one or more growth factors or cytokines affecting leukocyte maturation, differentiation, and antibody production. In support of this hypothesis, Egr-1 is present in lymphocytes and B cells and regulates the expression of activation-associated leukocyte genes such as the adhesion protein CD44 (33). It is interesting that the integral impact of Egr-1 deficiency is to enhance IgE levels, thus suggesting that there are mechanisms resulting in chronic suppression of IgE expression.

In this study we have for the first time established a role for the Egr-1 transcriptional pathway in vivo in the regulation of TNF-, IgE, and baseline airway responsiveness to methacholine in a mouse model. Because linkage analysis has identified susceptibility loci regulating serum IgE levels, AHR, and clinical asthma on chromosome 5q, and Egr-1 is located in this region of the human genome (34), we speculate that Egr-1 may be involved in similar regulatory events in humans and that there exist mutations in the Egr-1 gene that may alter its expression. Extrapolation from our murine data supports the hypothesis that the Egr-1 transcriptional pathway, if activated in asthma, subtly but significantly modifies the regulation of atopy and airway hyperresponsiveness characteristic of this disease.