Deficient Follistatin-like 3 Secretion by Asthmatic Airway Epithelium Impairs Fibroblast Regulation and Fibroblast-to-Myofibroblast Transition

Abstract

Bronchial epithelial cells (BECs) from healthy children inhibit human lung fibroblast (HLF) expression of collagen and fibroblast-to-myofibroblast transition (FMT), whereas asthmatic BECs do so less effectively, suggesting that diminished epithelial-derived regulatory factors contribute to airway remodeling. Preliminary data demonstrated that secretion of the activin A inhibitor follistatin-like 3 (FSTL3) by healthy BECs was greater than that by asthmatic BECs. We sought to determine the relative secretion of FSTL3 and activin A by asthmatic and healthy BECs, and whether FSTL3 inhibits FMT. To quantify the abundance of the total proteome FSTL3 and activin A in supernatants of differentiated BEC cultures from healthy children and children with asthma, we performed mass spectrometry and ELISA. HLFs were cocultured with primary BECs and then HLF expression of collagen I and α-smooth muscle actin (α-SMA) was quantified by qPCR, and FMT was quantified by flow cytometry. Loss-of-function studies were conducted using lentivirus-delivered shRNA. Using mass spectrometry and ELISA results from larger cohorts, we found that FSTL3 concentrations were greater in media conditioned by healthy BECs compared with asthmatic BECs (4,012 vs. 2,553 pg/ml; P = 0.002), and in media conditioned by asthmatic BECs from children with normal lung function relative to those with airflow obstruction (FEV₁/FVC ratio < 0.8; n = 9; 3,026 vs. 1,922 pg/ml; P = 0.04). shRNA depletion of FSTL3 in BECs (n = 8) increased HLF collagen I expression by 92% (P = 0.001) and α-SMA expression by 88% (P = 0.02), and increased FMT by flow cytometry in cocultured HLFs, whereas shRNA depletion of activin A (n = 6) resulted in decreased α-SMA (22%; P = 0.01) expression and decreased FMT. Together, these results indicate that deficient FSTL3 expression by asthmatic BECs impairs epithelial regulation of HLFs and FMT.

Keywords: FSTL3; activin A; airway remodeling; asthma; epithelial cells.

Figure 1.

Quantification of follistatin-like 3 (FSTL3) via targeted proteomics in primary bronchial epithelial cells (BECs) derived from healthy donors and donors with asthma. (A) The peptide primary sequence, locations and numbers of cleavage ions, and representative images showing histograms of the indicated product ions quantified from two independent FSTL3 peptides in isolates from media samples obtained from healthy BECs compared with asthmatic BECs. (B) FSTL3 abundance was assessed by summing the peak area quantification of the product ions from two technical replicates from two independent experiments using healthy donors (n = 4) and donors with asthma (n = 4). FSTL3 abundance was decreased in media samples obtained from asthmatic BECs compared with healthy BECs. Significance was assessed using an unpaired Student’s t test (*P < 0.05).

Figure 2.

Secreted levels of activin A and FSTL3 in media obtained from primary BECs derived from healthy donors and donors with asthma. (A) No significant difference was observed in activin A secretion between the healthy (n = 21) and asthmatic (n = 21) BECs (323 vs. 549 pg/ml; P = 0.3). (B) FSTL3 secretion from asthmatic BECs was significantly lower than that from healthy BECs (4,012 vs. 2,553 pg/ml; P = 0.002). When the asthmatic BECs were dichotomized into BECs from subjects with normal airflow (NA, n = 12) and BECs from subjects with airflow obstruction (AO, n = 9), asthmatic BECs from donors with AO (FEV₁/FVC < 0.8) demonstrated significantly less secretion of FSTL3 compared with asthmatic BECs from donors with normal airflow (3,026 vs. 1,922 pg/ml; P = 0.04).

Figure 3.

Knockdown studies using a red fluorescent protein (RFP)-tagged, doxycycline (Dox)-inducible lentiviral shRNA targeted against activin A and FSTL3 in primary BECs. (A) Secretion of activin A by primary BECs transduced with Dox-inducible shRNA targeting activin A and nontargeting (NT) control shRNA (n = 10). Addition of Dox to the media led to a 68% reduction of activin A secretion (n = 11 asthmatic BEC lines, n = 5 healthy BEC lines; 62.6 pg/ml vs. 159.8 pg/ml; P < 0.0001). (B) Secretion of FSTL3 by primary BECs transduced with Dox-inducible shRNA targeting FSTL3 and NT control shRNA. Addition of Dox to the media led to a 46% reduction of FSTL3 secretion (n = 6 asthmatic BEC lines, n = 11 healthy BEC lines; 3,048 pg/ml vs. 5,750 pg/ml; P < 0.0001). (C and D) RFP expression in FSTL3 shRNA lentiviral-transfected BECs without and with the addition of Dox, respectively.

Figure 4.

Relative expression of collagen I and α-smooth muscle actin (α-SMA) by human lung fibroblasts (HLFs) conditioned with primary BECs derived from healthy donors and donors with asthma. Collagen I expression was fourfold greater in HLFs cocultured with asthmatic BECs (n = 6) compared with HLFs cocultured with healthy BECs (n = 8; P = 0.02), and α-SMA expression was twofold greater in HLFs cocultured with asthmatic BECs compared with HLFs cocultured with healthy BECs (P = 0.01).

Figure 5.

(A) Relative expression of collagen I and α-SMA by HLFs conditioned with primary BECs derived from healthy donors (n = 13) after inhibition of BEC secretion of FSTL3 via Dox-inducible lentiviral shRNA knockdown. After knockdown of FSTL3 production by healthy BECs, collagen I expression was 92% greater in HLFs cocultured with BECs exposed to Dox compared with BECs not exposed to Dox (P = 0.001). α-SMA expression was 88% greater in HLFs cocultured with BECs exposed to Dox compared with BECs not exposed to Dox (P = 0.02). (B) Expression of collagen I and α-SMA by HLFs conditioned with primary BECs derived from donors with asthma (n = 12) after inhibition of BEC secretion of activin A via Dox-inducible lentiviral shRNA knockdown. After knockdown of activin A secretion by asthmatic BECs, collagen I expression was not significantly lower in HLFs cocultured with BECs exposed to Dox compared with BECs not exposed to Dox (P = 0.1). α-SMA expression was 22% lower in HLFs cocultured with BECs exposed to Dox compared with BECs not exposed to Dox (P = 0.02). (C and D) Representative immunostaining for collagen I (red), α-SMA (green), and DAPI (blue) in a representative healthy BEC-HLF coculture FSTL3 shRNA knockdown experiment (C, Dox-negative; D, Dox-positive).

Figure 6.

(A) In healthy BEC-HLF cocultures (n = 5), shRNA knockdown of FSTL3 in BECs increased the percentage of HLFs that stained for cytoskeletal α-SMA by flow cytometry as a myofibroblast marker from 5.7% (± SD 4.6) to 14.7% (± SD 9.1) (P = 0.04). (B) In asthmatic BEC-HLF cocultures (n = 6), shRNA knockdown of activin A in BECs decreased the percentage of HLFs staining for cytoskeletal α-SMA by flow cytometry from 15.9% (± SD 12) to 8.5% (± SD 6.5) (P = 0.05).