(A–H) IT-Ad-IL-1β or Ad-control (Ad-C) was administered to adult mice to induce airway remodeling. (A–D) In Ad-IL-1β– (B) but not Ad-C–treated mice (A), 7 days after treatment, inflammation surrounds small airways (arrows, B), with increased collagen (arrowheads, D). Scale bar: 150 μm. (E) ELISA for IL-1β. (F) qPCR for Itgb8 (P = 0.01); (G) Serpine1 (P = 0.01); (H) Col1a2 (P = 0.04). (I) TGF-β bioassay of active and total TGF-β in BAL fluid from Itgb8^fl/– mice (n = 6) 14 days after IT-Ad-IL-1β without or with IT-Ad-Cre. (J) Mice treated with IT-Ad-IL-1β with or without systemic administration of control antibody, anti-β6, or anti–TGF-β were evaluated at 14 days. (K–N) IT-Ad-IL-1β or Ad-C was administered to adult Itgb8^fl/– mice expressing Cre-ER(T) under control of the fibroblast-specific Col1a2 promoter treated without or with tamoxifen (Tam). Tamoxifen causes specific deletion of Itgb8 in fibroblasts (Itgb8^fko). (K and L) Inflammation (K, left) or fibrosis (L, left) is diminished in Itgb8^fko mice (airway lumen, A). Arrows indicate inflammation in K; and arrowheads, collagen bundles, at low and high magnification (inset) in L in the presence of vehicle (K and L, left) or tamoxifen (K and L, right). Scale bar: 250 μm; 40 μm (inset). (M and N) Inflammation (M) or fibrosis (N) 7, 14, 21, or 42 days after IT-Ad-IL-1β without or with fibroblast-specific deletion of Itgb8. Shown are values minus the non–IT-Ad-IL-1β–treated controls. Raw data are shown in Supplemental Table 1. *P < 0.05, **P ≤ 0.01, ***P < 0.001.

(A and B) IL-β–dependent increases in lymphocyte and neutrophil numbers are inhibited in Itgb8^fko mice. BAL total cell counts (A) and differential (B) 7 days after IT-Ad-IL-1β in the presence (No Tam) or absence (Tam) of Itgb8 on fibroblasts. (C) Multicolor cell surface staining for lung CD4⁺ and CD8⁺ T cells, B cells, NK cells, PMNs, and macrophages (Macs) 14 days after IT-Ad-IL-1β treatment without or with fibroblast-specific deletion of Itgb8 (n = 3). Shown are gated cell numbers. (D) Multicolor intracellular staining of CD4⁺ cells for IL-17 and/or IFN-γ. Three populations are identified: Th-17 cells (IL-17⁺IFN-γ^–), IFN-γ⁺ Th-17 cells (IL-17⁺IFN-γ⁺), and Th1 (IL-17^–IFN-γ⁺). Shown are cell numbers. (E) ELISpot for IL-17. Shown are number of spots/10⁶ plated cells. (F and G) Multicolor cell surface staining for lung mDCs 14 days after IT-Ad-IL-1β treatment without or with fibroblast-specific deletion of Itgb8 (n = 6) (F) or with control IgG or systemic anti–TGF-β (n = 6) (G). Gating strategy for lung DCs is shown in Supplemental Figure 4, based on published methodology (54, 88). Shown is the increase in numbers of each subset relative to control Ad-LacZ plus corn oil or Ad-LacZ plus tamoxifen. *P < 0.05, **P < 0.01, ***P < 0.001.

(A) Time course of IT-Ad-IL-1β–induced increases in DCs in the MLN and dependence on Itgb8 (n = 6, each time point). Col-Cre-ER(T);Itgb8^fl/– mice were treated with IT-Ad-IL-1β with or without tamoxifen and the MLN harvested 7 or 14 days later. Gating strategy and representation of CD8α⁺, CD4⁺, 103⁺, 103^– and inflammatory DCs are identical to those in Ballesteros-Tato et al., with the exception that CD4^–CD8α^– DCs are not shown (54). Shown is the increase in cell number relative to the controls. (B) MLN DC subpopulations of IT-Ad-IL-1β–treated mice (14 days) without and with systemic anti–TGF-β (n = 6). Shown is the increase in cell number minus the controls. (C) MLN gating strategy (MHC-II [Ia] versus CD11c), with percentage of cells shown within each gate. (D) The percentages of gated MHC-II^hiCD11c^hi DCs positive or negative for 103 are shown. Indicated in C are gates used for identification of CD4⁺ and CD8α⁺ conventional DCs and inflammatory DCs (iDC). (E and F) CFSE-labeled BMDCs (2 × 10⁶/ mouse) were injected into the lateral tail vein 12 days after IT-Ad-IL-1β or Ad-C treatment and 36 hours prior to lung (E) and MLN (F) harvest of mice without or with fibroblast deletion of Itgb8 (n = 9). Numbers of MHC-II^hiCD11c⁺ CFSE-labeled cells are shown. *P < 0.05, **P < 0.01.

(A) qPCR of mouse lung homogenates 14 days after IT-Ad-IL-1β treatment without or with fibroblast deletion of Itgb8 (n = 6). Shown are copy number increases relative to control mice. (B) IL-1β treatment of mouse lung fibroblasts induces a sustained induction of collagen expression that is dependent on Itgb8 and TGF-β. Mouse lung fibroblasts were cultured from lungs of Itgb8^fl/fl mice and treated with Ad-Cre or Ad-LacZ (Ad-C) (2.5 × 10⁶ PFU) overnight prior to treatment with recombinant IL-1β (1 ng/ml) with isotype control mAb or anti–TGF-β (40 μg/ml). Cells were lysed 72 hours or 1 week later, and Sircol assay was performed to determine collagen concentration. CCL2 (C and E) and CCL20 (D and F) as measured by ELISA from Itgb8^fl/fl mouse lung fibroblasts treated with Ad-Cre or anti–TGF-β (C and D) or from whole lung homogenates or BAL from Col-CreER(T);Itgb8^fl/– mice 14 days after IT-Ad-IL-1β without or with tamoxifen (E and F). *P < 0.05, **P < 0.01, ***P < 0.001.

Chronic ovalbumin sensitization causes airway inflammation, wall fibrosis, and increased mucin production (A–C), which is reduced by the absence of Itgb8 on fibroblasts (D–F). (A–I) Histologic (A–F) and morphometric (G–I) analysis of airway inflammation (A, D, and G); fibrosis (B, E, and H); and mucin (C, F, and I). Arrows in A indicate airway inflammation, and arrowheads in B, fibrosis. Scale bars: 150 μm; 40 μm (inset). ***P < 0.001.

(A) BAL cell counts and (B) cell differential of lymphocytes (Lymphs), PMNs, macrophages (Macs), and eosinophils (Eos) from BAL of mice after chronic ovalbumin challenge (n = 6). (C and D) Multicolor cell surface staining for lung DCs (C) or DCs from MLN (D) from mice after ovalbumin sensitization and challenge without or with fibroblast-specific deletion of Itgb8 (n = 6). Gating strategies were identical to those in Figure 2, F and G. (E and F) Ovalbumin conjugated to FITC was introduced intranasally 36 hours prior to lung (E) and MLN (F) harvest, and OVA-FITC–labeled cells were identified using gating strategies identical to those shown in Figure 3, C and D (n = 8). (G) qPCR analysis of cytokines present in lung after chronic ovalbumin challenge without or with fibroblast-specific deletion of Itgb8 (n = 6). Shown are copy number increases relative to control mice. (H and I) Ovalbumin induction of CCL2 (H) and CCL20 (I) as measured by ELISA on whole lung homogenates or BAL from Col-CreER(T);Itgb8^fl/– mice after chronic ovalbumin challenge without or with tamoxifen (n = 6). *P < 0.05, **P < 0.01, ***P < 0.001.

(A) ITGB8 determined by qPCR from normal and COPD fibroblasts treated with IL-1β (n = 5, P = 0.03). (B) Collagen in cell lysates from normal and COPD fibroblasts (n = 5) treated with or without IL-1β with isotype control, anti-β8, or anti–TGF-β. (C) TGF-β activation was measured by bioassay (arbitrary light units). (D) Selected human cytokine antibody array data (relative spot intensity). See Supplemental Figure 9 for complete array data. ELISA for CCL2 (E) and CCL20 (F). Results are expressed as fold increase compared with non–IL-1β–treated cells (n = 5). (G) Murine BMDC migration to fibroblast conditioned media (CM) or serum-free DMEM (Media). Mouse Itgb8^fl/fl fibroblasts were treated with or without IL-1β in the presence of Ad-Cre, anti–TGF-β, or isotype control for 48 hours (n = 6). (H) Migration of wild-type, Ccr2^–/–, or Ccr6^–/– BMDCs (n = 3). (I) Migration of wild-type BMDCs to untreated or IL-1β–treated mouse lung fibroblast CM in the presence of anti-CCL2, anti-CCL20, or both antibodies combined (n = 3). (J) Migration of human cord blood–derived DCs to unstimulated (Control) or IL-1β–stimulated isotype control–, anti-β8–, or anti–TGF-β–treated pooled normal and COPD fibroblast CM (n = 4). *P < 0.05, **P < 0.01, ***P < 0.001.

Stained sections were from a COPD cohort classified according to Global Initiative on Obstructive Lung Disease (GOLD) criteria (89), with GOLD stage 1–2 corresponding to mild to moderate and 3–4 to severe obstructive disease. Shown are sections of normal small airway wall (A) or COPD small airway wall (B) stained with ant-CD11c. DCs were identified by positive CD11c staining and the presence of long extended cellular processes (arrows in B). Airway epithelial cells (AE) and lamina propria (LP) are indicated. Scale bar: 20 μm. (C) CD11c immunostaining was graded on a 1–3 scale and compared according to disease status by GOLD stage. Normal (NL), n = 8; COPD, n = 13. ANOVA revealed a statistically significant difference (P = 0.02), and a P value for trend of 0.005. (D) CD11c immunostaining scores significantly correlated with β8 immunostaining scores (r = 0.6188, P = 0.0016).

(i) Injury (i.e., tobacco smoke) or antigen (Ag) leads to inflammasome activation and caspase-1–mediated cleavage and release of active IL-1β. (ii) IL-1β is released by epithelial cells or macrophages. (iii) IL-1β increases αvβ8 expression by fibroblasts, which leads to increased conversion of latent to active TGF-β. (iv) Active TGF-β binds to fibroblast TGF-β receptors, leading to increased TGF-β–dependent chemokine (CCL2, CCL20) secretion and increased recruitment of DCs from the peripheral circulation. (v) Immature DCs encounter Ag within the epithelium or lamina propria and migrate toward fibroblasts via αvβ8-dependent chemotactic gradients. (vi) Mature DCs gain access to draining lymphatics. (vii) Mature DCs enter the draining MLN, where they prime adaptive immune responses. (viii) Differentiated memory/effector lymphocytes migrate to the airway to establish pathologic inflammation.