Folliculin controls lung alveolar enlargement and epithelial cell survival through E-cadherin, LKB1, and AMPK

Abstract

Spontaneous pneumothoraces due to lung cyst rupture afflict patients with the rare disease Birt-Hogg-Dubé (BHD) syndrome, which is caused by mutations of the tumor suppressor gene folliculin (FLCN). The underlying mechanism of the lung manifestations in BHD is unclear. We show that BHD lungs exhibit increased alveolar epithelial cell apoptosis and that Flcn deletion in mouse lung epithelium leads to cell apoptosis, alveolar enlargement, and an impairment of both epithelial barrier and overall lung function. We find that Flcn-null epithelial cell apoptosis is the result of impaired AMPK activation and increased cleaved caspase-3. AMPK activator LKB1 and E-cadherin are downregulated by Flcn loss and restored by its expression. Correspondingly, Flcn-null cell survival is rescued by the AMPK activator AICAR or constitutively active AMPK. AICAR also improves lung condition of Flcn(f/f):SP-C-Cre mice. Our data suggest that lung cysts in BHD may result from an underlying defect in alveolar epithelial cell survival, attributable to FLCN regulation of the E-cadherin-LKB1-AMPK axis.

Figure 1. Lung histology and FLCN and SP-C immunostaining

(A) H&E staining of normal human lung (Control) (n=3) and BHD lung (n=4). (B) FLCN-positive (red) AECs (SP-C, green) are seen in normal human lung (n=3) but not in BHD (n=4) lungs. DAPI (blue) stains nuclei. (C) Genotyping of homozygous Flcn^f/f:SP-C-Cre (1), wild-type Flcn^wt/wt:SP-C-Cre (2), and heterozygous Flcn^wt/f:SP-C-Cre mice (3). (D) Flcn levels in whole lung lysates from Flcn^f/f:SP-C-Cre mice on doxycycline (Dox+) for 6 weeks (n=3) or regular (Dox-) diet (n=3). (E) Loss of Flcn (red) in lung AECs (SP-C, green) in Flcn^f/f:SP-C-Cre mice treated as in (D). Scale bars, 20 μM. A – conducting airways; V- blood vessels. See also Figure S1.

Figure 2. Loss of FLCN increases pulmonary alveoli, impairs lung function and induces alveolar epithelial cell apoptosis

(A–C) Flcn loss results in alveolar enlargement in Flcn^f/f:SP-C-Cre mice treated as in (1D). Scale bars, 200 μM. (D–F) Enlarged alveoli in pups with FLCN deletion in lung epithelium. Scale bars, 200 μM. (G) FLCN deletion in Flcn^f/f:SP-C-Cre mice impairs lung function, n=8 per group. BL–baseline; S-saline. The mean is shown; error bars represent SE (n>3). Data for Dox- mice are taken as one fold. (H) Cleaved caspase-3-positive (red) cells in lung epithelium (SP-C, green) of BHD lung, (n=5) but not in control (n=3) lung. (I) Loss of Flcn in lung epithelium (SP-C, green) results in alveolar epithelial cell apoptosis (red) in lung from Flcn^f/f:SP-C-Cre mice treated as in (1D), n=3 per group. Scale bars, 20 μM. See also Figure S2.

Figure 3. FLCN loss impairs AMPK activation and upregulates cleaved caspase-3

(A) FLCN-null UOK257 and FLCN-expressing UOK257-2 epithelial cells were serum deprived (S.D.) in DMEM supplemented with 0.1% BSA. Data are mean ± SE, n=3. (B) Energy depletion of FLCN-null UOK257 cells upregulates cleaved caspase-3.

Figure 4. FLCN regulates LKB1 levels and AMPK phosphorylation

(A) Primary lung AECs from FLCN^f/f mice were infected with control (−) or Cre- recombinase-expressing adenovirus (AdCre,+) followed by immunoblot analysis. See also Figure S3. (B) Immunoblot analyses of mouse epithelial NMuMG cells transfected with Flcn siRNA (siFlcn) or control scrambled siRNA (−). (C) Re-expression of FLCN in human FLCN-null UOK-257 cells increases membrane localization of LKB1. Top: representative images. Bottom: statistical analyses. Protein ratio for control cells were taken as one fold. Data are mean ± SE, n=3.

Figure 5. Flcn loss reduces E-Cadherin levels, increases cellular permeability and promotes apoptosis of primary mouse lung AECs

(A–C) AdCre-induced Flcn deletion in AECs from Flcn^f/f mice was detected by RT-PCR (A) and immunoblot (B) with statistical analysis (C). E-cadherin/tubulin ratio for control cells is taken as one fold. (D) Loss of FLCN in AECs downregulates membrane localization of E-Cadherin (red, upper panel) but not ZO1 (red, lower panel). DAPI (blue) stains nuclei. Scale bars, 50 μM. See also Figure S4. (E) Cytoplasmic E-cadherin localization in TSC2-null epithelial cells. Scale bars, 25 μM. (F–G) FLCN expression (green) results in membrane localization of E-cadherin (red) in TSC2-null cells (F). Data (G) represent % of cells, > 60 cells/condition (F). (H) Flcn deletion increases lung AEC permeability. Cell permeability in control is taken as one fold. (I–J) Flcn deletion in lung AECs upregulates cleaved caspase-3. (K) Flcn deletion in AECs results in DNA fragmentation. Number of TUNEL-positive cells to total number of cells is taken as 100%. Data are mean ± SE, n>3.

Figure 6. Increased Flcn-null epithelial cell apoptosis is rescued by AICAR and constitutively active AMPK

(A) Flcn knockdown downregulates membrane localization of E-cadherin (red). Scale bars, 25 μM. See also Figure S5 and S6. (B) Flcn knockdown decreases trans-epithelial resistance (TER). TER of siContr-transfected NMuMG cells was taken as 100%. (C) Cleaved caspase-3 is upregulated by siFlcn. (D) Flcn knockdown induces DNA fragmentation (TUNEL assay) of NMuMG cells. (E) AICAR and constitutively active AMPK (AMPK-CA) rescue disruption of epithelial cell morphology caused by siFlcn. Cells were treated either with 100 mM AICAR or diluent, or were infected with AdAMPK-CA or control adenovirus. Scale bars, 100 μM. (F–G) Expression of AMPK-CA (F) and AICAR-induced AMPK and ACC (G) phosphorylation in epithelial NMuMG cells. (H–I) DNA fragmentation (H) and epithelial cell death (I) induced by Flcn loss is rescued by AICAR and AMPK-CA. Data represent percentage of TUNEL-positive (H) or dead (I) cells per total number of cells taken as 100%. Data are mean ± SE, n=3.

Figure 7. AICAR improves lung homeostasis of Flcn^f/f:SP-C-Cre mice with Flcn deletion in lung epithelium

(A) Abnormalities in pulmonary phospholipids in BAL resulting from Flcn deficiency are restored by AICAR. Flcn^f/f:SP-C-Cre mice on Dox- or Dox+ were treated with AICAR or diluent for 6 weeks. (B) Flcn-induced impairment of surfactant surface tension is rescued by AICAR. (C) Flcn loss induces DNA fragmentation of AECs. Number of TUNEL-positive cells to total number of cells was taken as 100%. (D) AICAR normalizes increased inflammatory cell influx. (E) H&E staining of Flcn^f/f:SP-C-Cre mouse lungs on Dox- or Dox+ treated with AICAR as in (A). (F, G) Morphometric analyses of Flcn^f/f:SP-C-Cre mouse lungs treated as in (A). The mean is shown; error bars represent SE (n>3). Data for Dox- mice are taken as one fold. (H, I) AICAR inhibits IL-6 (H) and MCP-1 (I) increased by Flcn loss. (J, K) Upregulation of MMP-3 and MMP-9 induced by Flcn loss in lung epithelium treated as in (A) are abrogated by AICAR. See also Figure S7. Data (A-K) are represented as mean ± SEM from two independent experiments, n = 5–7. (L) A proposed model for the role of FLCN in lung alveolar homeostasis. FLCN mutations in lung epithelium downregulate membrane localization of E-cadherin and LKB1 which impairs AMPK activation. This model proposes that FLCN plays an important physiological function to control alveolar epithelial cell survival and maintains alveolar surface tension. Loss of FLCN results in alveolar collapse and impairment of lung function.