TRβ activation confers AT2-to-AT1 cell differentiation and anti-fibrosis during lung repair via KLF2 and CEBPA

Abstract

Aberrant repair underlies the pathogenesis of pulmonary fibrosis while effective strategies to convert fibrosis to normal regeneration are scarce. Here, we found that thyroid hormone is decreased in multiple models of lung injury but is essential for lung regeneration. Moreover, thyroid hormone receptor α (TRα) promotes cell proliferation, while TRβ fuels cell maturation in lung regeneration. Using a specific TRβ agonist, sobetirome, we demonstrate that the anti-fibrotic effects of thyroid hormone mainly rely on TRβ in mice. Cellularly, TRβ activation enhances alveolar type-2 (AT2) cell differentiation into AT1 cell and constrains AT2 cell hyperplasia. Molecularly, TRβ activation directly regulates the expression of KLF2 and CEBPA, both of which further synergistically drive the differentiation program of AT1 cells and benefit regeneration and anti-fibrosis. Our findings elucidate the modulation function of the TRβ-KLF2/CEBPA axis on AT2 cell fate and provide a potential treatment strategy to facilitate lung regeneration and anti-fibrosis.

Fig. 1. Thyroid hormone is decreased in serum after lung injury but is essential for lung regeneration.

a–c The levels of serum T₃, T₄, and TSH in mice after lung challenge with LPS (n = 6), BLM (n = 10, TSH n = 6), and SiO₂ (n = 7, 9; 8, 8; 8, 8) at 24-hour, 14-day, and 56-day, individually. d Dio2 mRNA level of lung homogenates (n = 4). e, f Illustration of the experiment design and line graph show the right lung Dio2 mRNA dynamic change after left lung PNX (n = 4). g, h Representative gross pictures of the right lung and the internal volume at day 21 after PNX (n = 3). Hypothyroidism (Hypo) was induced by PTU in drinking water and TDX. i, j Representative H&E staining of lung sections and mean liner intercept of alveolar (n = 9, from 3 mouse lung). Arrows indicated the thickened alveolar walls. Scale bar, 100 μm. k–n AT2 lineage tracing mice treated with T₃ after hypothyroidism and PNX construction (k), mRNA levels of proliferation and growth marker in lung homogenates (l, n = 4), and immunofluorescence images (m, the insets are shown on the right with individual fluorescence channels; tdT, tdTomato) and quantification (n, n = 5, from 3 mouse lung) of lung sections show the KI67-positive cells. The value of n indicates biologically independent samples (a–d, f, h, l). Data of different sections from three biologically independent mice (j, n). Similar results were repeated in two biologically independent experiments. Data are presented as mean ± SEM. P-values were obtained by two-tailed unpaired Student’s t test (a–d) and one-way ANOVA with Turkey’s multiple comparisons (h, j, l, and n).

Fig. 2. TRα facilitates cell proliferation, while TRβ promotes AT2 cells differentiation into AT1 cells.

a The heatmap of indicated gene expression in the HLCA and MCA data. b EdU assay in MRC5 after THRA and THRB overexpression for 24 h with 10 nM T₃ for another 24 h (n = 3). Scale bar, 100 μm. c EdU assay in MRC5 with or without 10 nM T₃ or GC-1 treatment (n = 5). Scale bar, 100 μm. d Workflow of the intervention with hypothyroidism and GC-1 in the differentiation phase of the PNX model. e qPCR analysis of AT1 cell markers (Ager, Hopx, and Pdpn) in lung homogenates (n = 3, 3, 4, 4). f, g Representative IF images and quantification show AT1 cells differentiated from AT2 cells (indicated by arrows) (n = 3). Scale bars, 50 μm. h–k Experimental workflow for 3D-organoid culture and GC-1 treatment. Cont., continued GC-1 from day 0. Late, GC-1 for the last 4 days (h). Representative images of colonies in wells, H&E staining, and IF. Scale bars, top 2 mm, middle and bottom, 50 μm (i). Colony average area and colony-forming efficiency of 3D-organoids (n = 4) (j). Quantification of AT1 cell markers by qPCR (n = 4) (k). The value of n indicates biologically independent samples. Similar results were repeated in two biologically independent experiments. Data are presented as mean ± SEM. Statistical significance among groups was determined using one-way ANOVA.

Fig. 3. GC-1 inhibits lung fibrosis in bleomycin-induced mouse model.

a Diagram of GC-1 delivery starting on day 11 after the bleomycin challenge. b Right lung hydroxyproline contents with gradient GC-1 treatment from 5 to 80 μg/kg (n = 5). c Right lung hydroxyproline contents (n = 8, 8, 8, 7) with or without GC-1 30 μg/kg. d Total lung internal volume under static pressure of 25 cm H₂O for 1 h (n = 4). e Total lung wet weight (n = 10, 10, 9, 10). f, g Protein content and white blood cell count in BALF (f, n = 11, 12, 8, 8; g, n = 8, 9, 8, 8). h Percentage of body weight change (n = 12, 11, 10, 12). i Percent survival of mice treated as indicated from d7 after 2U/kg BLM challenge (n = 10). j Immunoblotting of ECM proteins (COL1A1, FN1), mesenchymal cell markers (N-Cad, α-SMA, VIM), and the epithelial cell marker (E-Cad) in lung homogenates. k qPCR analysis of indicated genes in the lungs (n = 3). l Representative micro-CT images, H&E, and trichrome-stained lung sections at d21 after BLM. Scale bar of whole sections, 3 mm; Scale bar of 10 × images, 100 μm. The value of n indicates biologically independent samples. Similar results were repeated in three biologically independent experiments. Data are presented as mean ± SEM. P-values were obtained by one-way ANOVA with Turkey’s multiple comparison and log-rank (i), respectively.

Fig. 4. Activation of TRβ constrains the metaplasia and hyperplasia of AT2 cells in fibrosis.

a The uniform manifold approximation and projection (UMAP) plot displays cells colored by cell type identity in scRNA-seq. b UMAP visualization of Thrb regulon activity. c GO enrichment analysis of GC-1-regulated genes in differential expression between BLM and PBS groups in mouse lung bulk RNA-seq (n = 3). d Immunochemistry staining for SFTPC and KRT8 in lung sections of mice, same as Fig. 3. Arrows indicate normal AT2 cells, and arrowheads indicate elongated or hypertrophic AT2 cells. Scale bar, 100 μm. e Protein analysis in lung homogenates for M-AT2 cell marker CLDN4. f, g Workflow, representative IF images, and quantification of Ki67-positive AT2 cells in lineage-labeled mice lungs with or without GC-1 after BLM challenge. Arrowheads indicate basaloid cells from lineage-labeled cells (indicated by arrows) (n = 3). Scale bar, 50 μm. h Representative pictures and quantification of EdU assay (n = 12, images from four experiments) in A549. GC-1 15 nM for 36 h. Scale bar, 100 μm. i–l AT2 cells isolated from different groups were used to stimulate normal primary fibroblasts for 48 h. The activation of lung fibroblasts was assayed by the 3D-collagen gel contraction (j, n = 3), expression of a-SMA (k, Scale bar, 20 μm), COL1A1, and FN1 (l). The value of n indicates biologically independent samples (c, g, j). Data of different views from four biologically independent samples (h). Similar results were repeated in two biologically independent experiments. Significant differences were assessed using two-tailed unpaired Student’s t test (h) and one-way ANOVA with Tukey test (g, j). Results are presented as mean ± SEM.

Fig. 5. TRβ activation promotes M-AT2 cells differentiation into AT1 cells.

a RNA-velocity analysis of AT2 cells, Krt8⁺ cells, and AT1 cells in scRNA-seq data. The arrows indicate the predicted lineage trajectories. b Heatmap showing the average FPKM value of AT1 cell genes per group in mouse lung bulk RNA-seq (same as Fig. 3c, n = 3). c, d Protein and mRNA levels of AT1 cell markers in lung homogenates (n = 3), same as Fig. 3. e AT2-lineage tracing mice were used to detect AT2 differentiation with a time-lapse. f Representative IF images from lung cryosections stained with antibodies against KRT8 and PDPN. Arrows point to KRT8⁺ cells, arrowheads indicate AT1 from lineage-labeled AT2 differentiation, and unfilled arrows point to KRT8⁺ lineage-labeled cells with basaloid morphology (n = 3). Scale bars, 50 μm. g Quantification of indicated cells in (f) (left, n = 15; right, BLM n = 10, BLM + GC-1 n = 11). h The percentage of KT8⁺tdT⁺ cells were tested by flow cytometry at d14 after lung digestion (n = 3). i, j Immunoblotting and qPCR analysis of AT1 marker genes after GC-1 treatment with dose gradient in A549 (j, n = 3) and MLE12. The value of n indicates biologically independent samples (d, h, j). Data of different sections from three biologically independent mice (g). Similar results were repeated in two biologically independent experiments. Significant differences were obtained using one-way ANOVA with Turkey’s multiple comparison tests. Error bars, SEM.

Fig. 6. TRβ modulates KLF2 and CEBPA to drive the expression of AT1 cell marker genes.

a Principal component analysis (PCA) of TFs in bulk RNA-Seq of mouse lung (same as Fig. 3c, n = 3). The lower and upper bounds of the boxplot correspond to the first and third quartiles (the 25th and 75th percentiles); whiskers represent minima/maxima or 1.5*IQR. b Histogram showing the contributions of TFs in PC.2 and PC.5. c, d Protein and mRNA levels (n = 3) of KLF2 and CEBPA in the mouse lung, as shown in Fig. 3a. e qPCR analysis of KLF2 and CEBPA expression in isolated AT2 cells from mice (n = 3). f, g Immunoblotting and qPCR analysis (n = 3) of KLF2 and CEBPA in A549 and MLE12 with a GC-1 gradient. h The mRNA level of Klf2 and Cebpa in primary AT2 cells after 10 nM GC-1 treatment for 6 h (n = 3). i–l Protein (i, k) and mRNA (j, l) tests of AT1 cell markers, KLF2, and CEBPA expression in A549 after KLF2 (i, j) and CEBPA (k, l) overexpression (n = 3). m Luciferase activity of KLF2, CEBPA, and AT1-marker promoters cloned in pGL3.0 after THRB transfection for 36 h with or without GC-1 10 nM for 12 h (n = 3). Values were normalized to the transfection vector. n, o Luciferase activity of indicated promoters after KLF2 (n) and CEBPA (o) transfection with 600 ng plasmid for 36 h (n = 3). p–r TRβ (p), KLF2 (q), and CEBPA (r) bind to the promoter regions (pro) of KLF2, CEBPA, and AT1 cell genes in ChIP q-PCR assays (n = 3). KRT5 promoter as negative control. Values were normalized to IgG. s Luciferase activity of AT1-maker promoters with KLF2 (300 ng) and CEBPA (300 ng) co-transfection for 36 h, as shown in (n, o) (n = 3). t IF image of KLF2 and CEBPA after co-transfection in A549. Scale bar, 10 μm. u CoIP of KLF2 and CEBPA overexpressed in A549. v Illustration of the regulatory processes of GC-1 on AT1 cell markers. The value of n indicates biologically independent samples. Similar results were repeated in two biologically independent experiments. The statistical tests used were one-way ANOVA (d, e, and g), two-tailed unpaired Student’s t test (h), and two-way ANOVA (j, l, m–s). Data are mean ± SEM.

Fig. 7. The pro-regeneration and anti-fibrotic effects of TRβ activation depended on KLF2 in mice.

a Schematic of the Klf2 Cre-loxp knockout site in AT2 cells and experimental design. b Hydroxyproline content in the right lung (n = 6, 6, 9, 6, 9, 6). c, d Protein concentration (n = 8, 7, 9, 7, 9, 7) and WBC numbers (n = 6, 6, 8, 7, 8, 7) in BALF. e Immunoblotting for COL1A1 and AT1 cell markers in lung homogenates. f Representative images of trichrome staining. Scale bar of whole sections, 2 mm; scale bar of 10 × images, 100 μm. g, h Representative IF staining and quantification of KRT8⁺ cells in AT2-KLF2-KO mice (n = 5). Arrows show the KRT8⁺ cells. Scale bars, 50 μm. i, j The mRNA levels of AT1 cell markers in the lung of AT2-Klf2-KO mice at the differentiation phase in the PNX model. The value of n indicates biologically independent samples (b–d, j). Data of different sections from three biologically independent mice (h). Similar results were repeated in two biologically independent experiments. Throughout, data are mean ± SEM. P-values were obtained by one-way ANOVA with Turkey’s multiple comparison tests.

Fig. 8. CEBPA is required for the benefits of TRβ activation in vivo.

a Workflow of AAV-shCebpa delivery, BLM challenge, and GC-1 treatment. b Hydroxyproline level (n = 6, 6, 7, 6, 7, 6). c Total lung wet weight (n = 7, 6, 5, 6, 6, 6). d WBC counts (n = 6, 6, 6, 5, 6, 6) and protein content (n = 7, 7, 7, 6, 7, 6) in BALF. e Line plots of the percent survival from 7 to 21 days (n = 10, 20, 20). f Protein levels of fibrotic markers and AT1 cell markers. g Representative images of trichrome staining. Scale bar of whole sections, 3 mm; Scale bar of 10 × images, 100 μm. h–j Representative images of AT2 cell hyperplasia and quantification of AT1 cell differentiation in AT2-lineage cells after shCebpa delivery in the PNX model (n = 5). The value of n indicates biologically independent samples (b–e). Data of different sections from three biologically independent mice (i). Similar results were repeated in two biologically independent experiments. Scale bar, 50 μm. Throughout bars, mean ± SEM. P-values were obtained by one-way ANOVA with Turkey’s multiple comparison test and log-rank (e), respectively.

Fig. 9. Work model of the pro-regeneration and anti-fibrotic effects of TRβ activation.

TRβ activation promotes M-AT2 cell differentiation into AT1 cells and inhibits M-AT2 cell accumulation through KLF2 and CEBPA. NCoA, Nuclear Receptor Coactivator. RXR, Retinoid X Receptor.