Glypican-3 is a key tuner of the Hedgehog pathway in COPD

Abstract

Hedgehog (HH) pathway is involved in pulmonary development and lung homeostasis. It orchestrates airway epithelial cell (AEC) differentiation and contributes to respiratory pathogenesis. The core elements Gli2, Smo, and Shh were found altered in the bronchial epithelium of patients with chronic obstructive pulmonary disease (COPD). Here, we investigated the co-receptors to fully decipher the complex machinery of airway HH pathway activation in health and COPD. The core elements and co-receptors of HH signalling were investigated in lung cell populations using single-cell RNAseq analysis. The transcript levels of the principal co-receptor GPC3 were investigated on public RNAseq datasets and by RT-qPCR. The localisation of GPC3 was evaluated through immunofluorescent stainings on isolated bronchial AEC and tissues from non-COPD and COPD patients. GPC3 pharmacological modulation was achieved with Codrituzumab during AEC differentiation. We demonstrated that the core elements were not abundant in pulmonary cell populations. Focusing on co-receptors, GPC3 was the most expressed transcript in tracheobronchial epithelial cells. The decrease in GPC3 transcript levels correlated with the severity of airway obstrution in COPD patients. Finally, interfering with GPC3 signalling during AEC differentiation induced downregulation of the HH pathway attested by a decrease of Gli2 leading to reduced ciliogenesis and altered mucin production. GPC3 appears as a crucial co-receptor for the HH pathway in the respiratory context. The modulation of GPC3 may represent a novel experimental strategy to tune HH signalling in therapeutic perspectives.

Keywords: Airway epithelial cells; COPD; GPC3; Hedgehog.

Fig. 1

Expression of GPC3 at a single-cell resolution in the human airways. (A) Bubble plot representing cell populations in y-axis and genes (regrouped in two categories: HH pathway core elements (DHH, GLI1, GLI2, GLI3, IHH, PTCH1, SHH, and SMO) and co-receptors (BOC, CDON, GAS1, GPC1, GPC3, GPC5, GPC6, HHIP, and PTCH2)) in x-axis. The size of the dots represents the proportion of expressing cells and the colour intensity represents the median of expression levels. Four categories of respiratory cell populations were selected: immune cells (n = 408371 cells), stromal cells (n = 120988 cells), pneumocytes (n = 123456 cells) and tracheobronchial epithelial cells (n = 40822 cells). (B) Left panel, UMAP of the HLCA pulmonary cell clusters. Right panel, UMAP of the GPC3 transcript levels for each cell population. The colour intensity represents the median of GPC3 transcript levels. The dashed lines highlight the population of fibroblasts, multiciliated cells, and basal cells.

Fig. 2

GPC3 transcript levels are decreased in COPD patients. (A) Dot plot with mean ± SD showing GPC3 expression analysis from RNAseq data (GSE57148). The whole-lung tissues were obtained from COPD (n = 98, red) and non-COPD (n = 91, black) patients. ∗p < 0.05; non-COPD vs COPD. (B) Dot plot with mean ± SD showing GPC3 expression analysis from RNAseq data (GSE137557). The airway epithelial cells were obtained from tracheobronchial segments of non-COPD and COPD patients and cultured in air-liquid interface conditions. The proliferative basal cells (n = 8) were collected for analysis in submerged conditions before the air switch and the fully differentiated cells (n = 6) were collected for analysis after 28 days of differentiation. ∗∗p < 0.01; non-COPD vs COPD. FPKM: Fragments per kilobase million.

Fig. 3

GPC3 transcript levels are associated with lung function in COPD patients. ALI cultures were performed from isolated AEC obtained from COPD patients (n = 7, red) and non-COPD (n = 7, black), and collected for transcript analysis after 7 days (basal cells) or 35 days (differentiated cells). (A) Dot plot with mean ± SD showing ΔCT GPC3 expression analysis on basal cells. ns: non-significant; non-COPD vs COPD. (B) Linear regression representing a correlation between GPC3 transcript levels and FEV₁ (forced expiratory volume in 1s) in ALI7 cultures of COPD patients (n = 7). (C) Linear regression representing a correlation between GPC3 transcript levels and FEV₁/FVC (forced vital capacity) in ALI7 cultures of COPD patients. (D) Dot plot with mean ± SD showing GPC3 expression analysis on differentiated cells. ns: non-significant; non-COPD vs COPD. (E) Linear regression representing a correlation between GPC3 transcript levels and FEV₁ in ALI35 cultures of COPD patients (n = 7). (F) Linear regression representing a correlation between GPC3 transcript levels and FEV₁/FVC in ALI35 cultures of COPD patients.

Fig. 4

Localisation of GPC3 in isolated AEC of non-COPD and COPD patients. (A) Representative micrographs taken from cytospun AEC of non-COPD and COPD patients showing cilia (Arl13b, green) and GPC3 (red). Nuclei are stained in blue (DAPI). Magnification corresponding to the selected area is shown. (B) Dot plot with mean ± SD representing GPC3 pixel mean grey values of non-COPD (n = 7, black) and COPD (n = 7, red) patients. ns: non-significant; non-COPD vs COPD.

Fig. 5

Localisation of GPC3 on FFPE lung tissues of non-COPD and COPD patients. (A) Representative micrographs showing bronchial epithelia of non-COPD and COPD patients stained for GPC3 (red), basal cells (KP, green) and cell nuclei (DAPI, blue). Magnification corresponding to the selected area is shown. (B) Dot plot with mean ± SD representing GPC3 pixel mean grey values of non-COPD (n = 7, black) and COPD (n = 16, red) patients. ∗p < 0.05; non-COPD vs COPD. (C) Linear regression representing a correlation between GPC3 pixel levels and FEV₁ in bronchial epithelia of COPD patients (n = 16). (D) Linear regression representing a correlation between GPC3 pixel levels and FEV₁/FVC in bronchial epithelia of COPD patients (n = 14).

Fig. 6

GPC3 inhibition by Codrituzumab alters AEC differentiation. (A) Network of protein-protein interaction generated on the Search Tool for the Retrieval of Interacting Genes/Proteins (STRING). (B) Representative micrographs taken from AEC cultures codrituzumab-treated or untreated at ALI7 and ALI35 stained for Gli2 (red) and cell nuclei (blue). (C) Dot plot with mean ± SD showing relative pixel mean grey values of Gli2 at ALI7 and ALI35. ∗p < 0.05; Control vs Codrituzumab. (D) Representative micrographs of AEC cultures codrituzumab-treated or untreated at ALI7 and ALI35 showing cilia (Acetylated tubulin, red) and cell nuclei (blue). (E) Dot plot with mean ± SD showing relative pixel mean grey values of acetylated tubulin at ALI7 and ALI35. ∗p < 0.05; Control vs Codrituzumab. (F) Representative micrographs of AEC cultures codrituzumab-treated or untreated at ALI35 stained for CC10 (red) and cell nuclei (blue). (G) Dot plot with mean ± SD showing relative pixel mean grey values of CC10 at ALI35 in Codrituzumab-treated conditions compared to untreated cultures. (H) Representative micrographs taken from AEC cultures codrituzumab-treated or untreated at ALI35 stained for mucins (Muc5ac or Muc5b, red) and cell nuclei (blue). (I) Dot plot with mean ± SD showing relative pixel mean grey values of Muc5ac or Muc5b at ALI35. ∗∗p < 0.01; Control vs Codrituzumab. (J) Dot plots with mean ± SD representing the relative pixel mean grey value ratios of Muc5ac/Muc5b in untreated (black) and Codrituzumab-treated cells (red). ∗p < 0.05; Control vs Codrituzumab.