Histopathology.  The method was established after conduction of the i.p. sensitization study, thus applied only in the i.n. sensitization study. Following bronchoalveolar lavage, lungs were inflated and immersion-fixed in neutral buffered formalin (10%), paraffin-sectioned at 5 μm thickness and stained with haematoxylin and eosin (H&E) or Periodic Acid Schiff (PAS). The inflammatory infiltrate and staining of goblet cells were evaluated by light microscopy of the H&E and PAS sections, respectively. All pathology scoring was performed by the same investigator (HR) that was aware of the animal grouping, but blinded to all other results. The intensity of

the perivascular and peribronchial inflammatory infiltration was scored according CHIR-99021 molecular weight to the following grading scheme. Lung sections graded as 0 showed no inflammatory Torin 1 infiltration. Sections graded as 1 demonstrated 1 or 2 minimal foci of perivascular and peribronchial infiltration, while grade 2 presented 3–6 foci of perivascular and peribronchial infiltration. Sections graded as 3 presented multiple foci of perivascular and peribronchial infiltration, many of which formed multilayered cuffs, while grade 4 presented multiple multilayered dense inflammatory infiltrates, primarily affecting the central parts of the lungs. Sections graded as 5 were as grade 4 but more extensive by affecting both central and peripheral parts

of the lungs. Staining of goblet cells in the bronchi was graded as 0, 1, 2, 3 and 4, corresponding to PAS-positive staining of 5% or less, 5–15%, 15–30%, 30–50% and more than 50% of bronchial epithelial cells. A Zeiss Axioplan 2 microscope (Carl Zeiss, Göttingen, Germany) with Plan-Neoflux 10 ×/0.30 lenses was used to magnify the histology slides. An RT Spot digital camera with the Spot RT slider v.4.6 software was used for image acquisition, addition and merger of electronic scale bar [using a Nikon MBM 11100 stage micrometre type A (Nikon, Tokyo, Japan) for objective calibration]. Adobe Photoshop CS4 v. 11.0

(Adobe Systems Inc., San Jose, CA, USA) was used for proportional else resizing of the images. Image resampling during resizing was performed as bicubic sharper. Pixel order was interleaved (RGBRGB), and no compression was applied upon saving. Auto colour and auto contrast correction was applied to the entire image. No other adjustment of the images was performed. Study design and statistical analysis.  A factorial design was used for both the i.p. and i.n. studies, which were analysed statistically by the General Linear Model procedure in Minitab v.15 (Minitab Inc., State College, PA, USA) with sex, age and allergen dose as fixed factors. When necessary, data were logarithmically or square root transformed to obtain equal variance and normal distribution of the residuals. Statistically significant main and interaction effect are reported.