Adult female BALBc mice aged 6 to 8 weeks (Harlan Iberica, Barcelona, Spain) were used in the study. All animal procedures were approved by the Ethics Committee for Animal Research of the Universitat Autònoma de Barcelona.

Sensitization to HDM was induced following a procedure established by Cates et al. 29. Briefly, the mice were exposed to purified HDM extract (Alk-Abelló, Madrid, Spain) with a very low LPS content (<0.2 EU/dose, measured using the Charles River Endosafe Limulus Amebocyte Assay (Charles River Laboratories, Wilmington, Massachusetts, USA). The allergen was administered intranasally under light halothane anesthesia for 10 consecutive days at a dose of 25 μg/mouse in a 20-μl volume. Non-sensitized (control) animals received intranasal saline.

An antisense oligonucleotide strategy was used to selectively down-regulate the production of lung COX-2 mRNA but not COX-1 mRNA. One day before initiating exposure to HDM, and up to two days after withdrawing the allergen, the mice received intranasal saline (untreated), control mismatched antisense oligonucleotide, or selective COX-2 antisense oligonucleotide at 20 μg/mouse (Figure 1). On the days both products were administered, the treatment was always provided one hour before administering the HDM extract. The COX-2-selective oligonucleotide sequence (IK6 antisense oligonucleotide, 5'GGAGTGGGAGGCACTTGC3') was taken from Khan et al. 30. The control oligonucleotide contained a 7-base mismatched sequence (5'GGA

C

A

TTCA

G

COX-2 mRNA expression in the lung was assessed by real time PCR. Total RNA was extracted using Trireagent (Molecular Research Center Inc, Cincinnati, Ohio, USA), and traces of contaminating genomic DNA were removed with DNAfree (Ambion Inc, Austin, Texas, USA). COX-2 cDNA was generated using MMLV reverse transcriptase (Epicentre, Madison, Wisconsin, USA). For real-time PCR, 2 μg of total RNA from each animal was reverse transcribed and the resulting cDNA was placed in the glass capillaries together with 18 μl of a master mix. The COX-2 primers were designed with PrimerSelect software (DNASTAR Inc, Madison, Wisconsin, USA) and were as follows: forward primer, 5'AGCCAGCAAAGCCTAGAGCAACAA3'; and reverse primer, 5'TGACCACGAGAAACGGAACTAAGAGG3'. PCR was performed in a LightCycler instrument and the crossing point (defined as the point at which fluorescence increases appreciably above background fluorescence) was determined with LightCycler software (both from Roche Diagnostics, Mannheim, Germany) using the second derivative maximum method. Using the Relative Expression Software Tool (REST^©), the relative expression ratio was calculated on the basis of group means for COX-2 as the target gene versus the reference gene GAPDH, and the calculated group ratio was tested for significance using a statistical model known as the Pair Wise Fixed Reallocation Randomization Test^© 31. We took into account the PCR efficiency calculated for COX-2 and GAPDH, which was very similar for both. For purposes of graphical representation, the COX-2 mRNA expression ratio of the non-sensitized untreated mice was established as 1.0, and the average ratios of the other experimental groups were re-calculated on that basis.

To support lung COX-2 mRNA assessment, the enzyme's airway protein expression and activity were determined in some of the mice from each experimental group (from 3 to 6 mice see legend of Figure 2b and 2c). Briefly, proteins were extracted from the right lung lobe of each animal using a lysis buffer containing protease inhibitors (Mini complete tablet, Roche Diagnostics, Mannheim, Germany). The concentration of COX-2 protein was determined by ELISA (IBL, Hamburg, Germany).

Immunohistochemistry for COX-2 was performed in lung sections using a polyclonal antibody (sc-1746, Santa Cruz Biotechnology, Santa Cruz, California, USA) after boiling in 10 mM citrate buffer (pH 6) to retrieve the antigen. The sections were then incubated overnight at 4°C with or without the primary antibody. A rabbit anti-goat secondary antibody (Vector, Burlingame, California, USA) was used at room temperature for 1 hour, followed by incubation with horseradish peroxidase-conjugated avidin-biotin complex (Pierce, Rockford, Illinois, USA). Staining was then performed with diaminobenzidine to reveal immunolabeling. Additionally, prostanoids were extracted from BAL fluid and purified through Sep-Pak C₁₈ columns (Waters Corporation, Milford, Massachusetts, USA). After evaporation and resuspension in EIA buffer, the PGE₂ concentration was measured using a commercially available specific ELISA (Cayman Europe, Tallin, Estonia).

To assess the effect of impaired COX-2 production on airway function, we analyzed the in vivo airway reactivity to increasing doses of nebulized methacholine (6.25 to 100 mg/ml) 24 hours after the last exposure to HDM in either treated or untreated sensitized and non-sensitized mice (Figure 1), using a well established 323334 non-invasive whole-body plethysmography (WBP) technique (Buxco Europe Ltd, Winchester, UK), which is based on changes in the time elapsed between nasal and thoracic pressure fluctuations. The response to methacholine was averaged and expressed as Penh (Enhanced Pause) as previously reported 14. Results were compared by two-way ANOVA. In the BALB/c mice strain, Penh has been shown to correlate with lung mechanics (R_L) 35. This can also be inferred from recent studies, in which AHR was assessed using WBP in combination with invasive procedures 3637.

Inflammation was evaluated 48 hours after the last exposure to HDM (Figure 1) using two approaches. Total cellularity in BAL samples was counted immediately after fluid collection. BAL was performed by slowly infusing 0.3 ml of phosphate-buffered saline (PBS) (2% fetal bovine serum) twice and recovering it by gentle aspiration 30 seconds after delivery. A 20-μl aliquot of BAL fluid was stained with Turk solution (0.01% crystal violet in 3% acetic acid) and analyzed in a Neubauer chamber. In addition, an experienced blinded observer counted the number of Congo red-stained eosinophils twice in histological lung sections of the same animals. The tissue area was also measured using MIP 45 Advanced System image analysis software (Microm España, Barcelona, Spain) to express the number of eosinophils per mm². Group means (n = 12) were compared using an unpaired t test.

The expression of mPGE synthase and hPGD synthase was determined by conventional reverse transcriptase-PCR using previously described primer pairs 38 in mRNA extracted from the lungs. GAPDH was assessed as the reference gene. Samples were denatured for 5 min at 95°C, and the cycling parameters (35 cycles) for mPGES, hPGDS, and GAPDH were 95°C for 30 sec, 55°C for 30 sec, and 72°C for 30 sec. A final extension step of 8 min at 72°C was applied. Amplification products were separated by agarose gel electrophoresis. After staining with ethidium bromide, the optical density of the bands was analyzed using Quantity One software (Bio-Rad Laboratories, Hercules, California, USA). The ratios of the band densities were calculated for each enzyme versus GAPDH and the means (n = 12) were compared using an unpaired t test.

Data in the text, table, and figures are expressed as the mean ± standard error of the mean (SEM) unless otherwise stated. Differences in airway response between different groups were tested for statistical significance using ANOVA followed by a post hoc Bonferroni test. The unpaired t test was used for all other analyses. Differences were considered statistically significant when the probability value was less than 0.05.

Figure 2a shows the relative expression ratio of COX-2 mRNA in lung tissue, where 1.0 was established as the baseline level in the untreated non-sensitized mice. The level in the lung tissue of HDM-sensitized mice treated with the control oligonucleotide was the same as in the lung tissue of the untreated sensitized mice. In contrast, COX-2 mRNA expression was significantly diminished in sensitized mice treated with the COX-2-specific antisense oligonucleotide compared with untreated (55% decrease) or mismatched control oligonucleotide-treated HDM-sensitized mice. In the non-sensitized mice's airway, the antisense oligonucleotide did not exert any effect on COX-2 mRNA expression. Our data also revealed that there were no statistically significant differences in the expression of lung COX-2 between non-sensitized and HDM-sensitized mice.

The concentration of COX-2 in lung protein extracts from some of the mice was measured by ELISA. A significant 45% reduction in COX-2 protein concentrations was observed in the lungs of COX-2 antisense-treated sensitized mice compared with non-treated sensitized mice (Figure 2b). Analysis of COX-2 expression in the lung by immunohistochemistry showed the same results for untreated HDM-sensitized mice and mismatched control oligonucleotide-treated sensitized mice (Figure 3e), that is, they consistently had a visibly increased number of positive cells and a stronger staining intensity than selective COX-2 oligonucleotide-treated mice (Figure 3f).

The photomicrographs in Figure 3 also show that the pattern of COX-2 expression in the airways of the mice was restricted to secondary and tertiary bronchi and bronchiolar epithelial cells (Figure 3a and 3b), as well as alveolar macrophages (Figure 3c). No expression whatsoever was observed in the epithelium of the main bronchi in any of the animals studied. This distribution pattern of COX-2 protein was found to be identical in all the experimental groups, regardless of whether they were sensitized or not, or treated or not. Only representative pictures of the pattern are included.

Finally, the PGE₂ concentration in BAL fluid was significantly (66%) lower in the lungs of COX-2 antisense oligonucleotide-treated sensitized mice than in the untreated sensitized ones (Figure 2c). Thus, impairment of COX-2 expression was accompanied by a similar level of impaired activity.

Increasing doses of methacholine induced a dose-dependent rise in Penh in all experimental groups (Figure 4). The Penh increase was higher in the HDM-sensitized mice than in the non-sensitized mice, revealing the induction of AHR in the HDM-sensitized animals. As shown in Figure 4, untreated and control oligonucleotide-treated sensitized mice had almost identical responses to the bronchoconstrictor. However, AHR was significantly higher in the selective COX-2 oligonucleotide-treated mice, where the Penh value at the maximum methacholine concentration (100 mg/ml) was almost twice that found in the untreated or mismatched control-treated sensitized animals (7.40 ± 1.62 vs 14.12 ± 2.75).

HDM-sensitized mice developed a clear peribronchial and perivascular eosinophilic inflammation (Figure 3e) with goblet cell hyperplasia, compared with the non-sensitized animals. No differences were found in the total number of inflammatory cells or in eosinophil accumulation between untreated and mismatched control oligonucleotide-treated sensitized animals (Figure 5a and 5b). In contrast, when COX-2 was selectively inhibited in sensitized mice, the level of inflammation was clearly and significantly reduced to 50%–55% of its value in the untreated HDM-sensitized mice, depending on whether total cells or eosinophils were considered (Figure 5a and 5b, respectively). This reduced inflammation was also visible in the lung sections (Figure 3f).

The mRNA expression of mPGE₂ and hPGD₂ was normalized to the level of constitutive GAPDH (Table 1). No differences were observed between the non-sensitized and the untreated sensitized mice. The expression ratio in the untreated and the mismatched control oligonucleotide-treated mice was also the same for both enzymes. Likewise, no differences were found in the mRNA expression of hPGD₂ synthase in the lungs of mice from any of the experimental groups, whether or not they were treated with COX-2 antisense oligonucleotide. In contrast, HDM-sensitized mice treated with the selective COX-2 oligonucleotide displayed significantly reduced expression of mPGE synthase (40% reduction) compared with mismatched control oligonucleotide-treated or untreated sensitized animals.