Submucosal Gland Myoepithelial Cells Are Reserve Stem Cells That Can Regenerate Mouse Tracheal Epithelium

Abstract

The mouse trachea is thought to contain two distinct stem cell compartments that contribute to airway repair-basal cells in the surface airway epithelium (SAE) and an unknown submucosal gland (SMG) cell type. Whether a lineage relationship exists between these two stem cell compartments remains unclear. Using lineage tracing of glandular myoepithelial cells (MECs), we demonstrate that MECs can give rise to seven cell types of the SAE and SMGs following severe airway injury. MECs progressively adopted a basal cell phenotype on the SAE and established lasting progenitors capable of further regeneration following reinjury. MECs activate Wnt-regulated transcription factors (Lef-1/TCF7) following injury and Lef-1 induction in cultured MECs promoted transition to a basal cell phenotype. Surprisingly, dose-dependent MEC conditional activation of Lef-1 in vivo promoted self-limited airway regeneration in the absence of injury. Thus, modulating the Lef-1 transcriptional program in MEC-derived progenitors may have regenerative medicine applications for lung diseases.

Keywords: LEF1; airway basal cell; injury repair; lineage-trace; lymphoid enhancer-binding factor-1; multipotency; myoepithelial cell; plasticity.

Figure 1. MEC-derived cells emerge from SMGs and adopt a basal cell-like phenotype in the SAE following injury

(A) Timeline of lineage-tracing of myoepithelial cells with tamoxifen (Tmx) and airway injury with naphthalene (Naph). ACTA2-Cre^ERT2:ROSA-TG mice were treated with tamoxifen daily for 5 days, rested for five days, and then treated with either vehicle or naphthalene (300mg/kg) and euthanized 14 or 21 days post-injury (DPI). (B–H) Tracheal sections at the indicated time points are oriented with the proximal region to the left and were stained for nuclei, Tomato, GFP, and the indicated phenotypic markers: (B) αSMA expression at 14 DPI (b: enlarged image of the boxed region shown in B); (C) Krt5 expression at 14 DPI (c: enlarged image of the boxed region in C); (D) Trop2 expression in a gland duct at 21 DPI; (E,F) NGFR expression at 21 DPI (fi and fii: enlarged images of the boxed regions shown in F); (G) Krt8 expression at 21 DPI; (H) Krt14 expression at 21 DPI; and (I) Control trachea at 21 days post-induction in the absence of epithelial injury. (J–K) Quantitation of lineage-traced cells in the SAE at 21 DPI as (J) % of total cells that are GFP⁺ in the C0–C4 region of the SAE, and (K) distribution of total GFP⁺ cells at various cartilage ring segments. Dotted line in J marks background level of signal close to the basal lamina in uninjured controls. (L) Distribution of total GFP⁺ αSMA⁺ and total GFP⁺ Krt8⁺ cells at various cartilage ring segments in the SAE as the % of total C0–C4 GFP⁺ cells expressing each marker at 21 DPI. Graphs show means +/− SEM for N=6 mice. P-values indicate significance of (J) unpaired one-tailed Student’s t-test, and (K,L) one-way ANOVA followed by posttest for linear trend. See supplemental Figures S1–S3.

Figure 2. MEC-derived cells produce ciliated cells on the SAE following injury

(A) Experimental design and summary of results: ACTA2-Cre^ERT2:ROSA-TG mice were induced with tamoxifen as in Figure 1A, treated with vehicle or naphthalene to induce injury, and harvested at the indicated time points. Circles indicate the presence (closed) or absence (open) of lineage-traced ciliated cells on the SAE. Tracheal sections were stained for nuclei, Tomato, GFP, and acetylated αtubulin. (B) Vehicle treated mice harvested at 60 days post mock (DPM). (C–E) Naphthalene treated mice harvested at 14, 21, or 60 days post injury (DPI). (F) Color key for lineage detection of ciliated cells in mice shown in panels G–I. The cilia of MEC-derived (GFP⁺) cells appear either white (left or top panel) or cyan (right or bottom panel); those of cells lacking the lineage-marker GFP (Tomato⁺) are either magenta (left or top panel) or white (right or bottom panel). (G–I) Enlarged, two-channel images of boxed regions in C–E show GFP (green)/αtubulin (magenta) and Tomato (red)/αtubulin (cyan). Traced (GFP⁺) ciliated cells are marked by closed yellow arrowheads, whereas their non-traced (Tomato⁺) counterparts are marked by open yellow arrowheads. (J,K) Quantitation of (J) % of total SAE cells that are GFP⁺ at various DPI and (K) % of total GFP⁺ cells that also express αtubulin at various DPI. Graphs show means +/− SEM for (N) mice as depicted on graphs. (L) Uninjured ACTA2-Cre^ERT2:ROSA-TG mouse at 1.5 years following tamoxifen-induction. P-values indicate significance of one-way ANOVA followed by posttest for linear trend. See supplemental Figures S1 and S3.

Figure 3. MEC-derived progenitors reestablish niches in the SAE that respond to repetitive injury and are multipotent for both SAE and SMG cell types

(A) Diagram of design for lineage tracing experiment comparing single injury (SI) and double injury (DI). Two other groups of control mice also included: uninduced/uninjured (UIND) and induced/uninjured (UI). (B–K) Tracheal sections of DI mice were stained for nuclei, Tomato, GFP, and the indicated phenotypic markers: (B) no marker, showing boundary of a lineage-traced group of cells (b: enlarged inset from B); (C) αtubulin (c: enlarged inset from C); (D) Scgb1a1; (E) Dolichos biflorus agglutinin (DBA) lectin; (F) Lysozyme (Lyz) in SMGs; (G, H) Ulex europaeus agglutinin I (UEA-1) lectin in (G) SMG and (H) SAE; (I,J) Muc5B in (I) SAE and (J) SMG; and (K) Scgb3a2. Arrows in c indicate ciliated cells that are GFP⁺ (yellow) or GFP⁻ (red). (L,M) Quantification of the % of total SAE cells that are GFP⁺ cells in the (L) SAE and (M) SMGs for all four groups of mice evaluated. (N,O) Quantitation of the (N) SAE and (O) SMG compartments showing the % of the total GFP⁺ cells (cyan) or % of the total Tomato⁺ cells (magenta) that also express the indicated phenotypic markers after single injury (solid bars) or double injury (checkered bars). Graphs show means +/− SEM for N=4–6 mice per group. P-values indicate significance of (L,M) one-way ANOVA and (N,O) two-way ANOVA followed by Sidak’s multiple comparisons test (ns=not significant, *P<0.05, **P<0.01, *** P<0.001, ****P<0.0001). See supplemental Figures S1 and S3.

Figure 4. Wnt/β-Catenin signaling is similarly activated in primordial gland stem cells during development and MECs following airway injury

(A,B) Glandular placodes (arrows) from newborn trachea localizing (A) TCF7 and (B) Lef-1 with Sox2. (C–J) Localization of αSMA with Sox2, Lef-1, TCF7, or β-Catenin in SMGs of (C–F) uninjured and (G–J) 24 hr post naphthalene (300mg/kg) injury. Panels to the right of F and J show (fi, ji) nuclear β-catenin (NβC) staining overlapping with DAPI (intensity of β-catenin staining is retained) and superimposed over an outline of αSMA staining (red lines) and (fii, jii) representative segmented images after multiwavelength cell scoring each nuclei showing αSMA⁻NβC⁻ (blue) and αSMA⁺NβC⁺ (yellow) cells. (K,L) Quantification of nuclear Sox2, Lef-1, TCF7, or nuclear β-Catenin staining as (K) the % of total SMG cells and (L) the % of αSMA⁺ MECs. Graphs show means +/− SEM for N=3–6 mice per group. All micron bars = 25 μm. Single daggers indicate significance of one-way ANOVA, and double daggers indicate posttest for linear trend. See supplemental Figure S1.

Figure 5. Lef-1 expression activates lineage commitment of MECs and migration to the SAE

(A) Transgenic ROSA26 knock-in construct (Lef-1KI) used to conditionally activate Lef-1 expression in MECs. (B) Experimental design for evaluating how Lef-1 expression influences MEC fate in ACTA2-Cre^ERT2:Lef-1KI^+/+ vs. ACTA2-Cre^ERT2:ROSA-TG mice in (C–F). (C) Uninduced and uninjured ACTA2-Cre^ERT2:Lef-1KI^+/+ demonstrating GFP expression in the majority of cells. (D,E) Tamoxifen induced/uninjured mice labeled with EdU as in (B) and stained for the indicated markers (d and ei–eiii: enlarged insets from D,E). (F) Quantification of EdU⁺αSMA⁺ MECs (N=5 mice per group). (G) Experimental design for evaluating how Lef-1 dosage impacts MEC fate with and without naphthalene (300mg/kg) injury in (H–U). (H–S) ACTA2-Cre^ERT2:Lef-KI^+/− and ACTA2-Cre^ERT2:Lef-1KI^+/+ mice were treated under the various conditions as marked and sections stained for (H–L) GFP and αSMA or (M–S) GFP and Lef-1. Insets of the SAE in (N,O) are from different animals at C6. (T,U) Quantification of lineage-traced cells (GFP⁻) as a % of the total cells in the (T) SMGs and (U) SAE of the C0–C4 tracheal region (N=3–4 mice per group). Graphs show means +/− SEM. Micron bars: (C–E, H–O) 100 μm; (P–S and insets N, O, i–iv) 50 μm. Asterisks indicate significance of (F) unpaired two-tailed Student’s t-test and (T,U) Newman-Keuls multiple comparisons testing (*P<0.05, **P<0.01, *** P<0.001). See supplemental Figure S4.

Figure 6. Lef-1 overexpression in MEC SCs promotes terminal differentiation toward multipotent progenitors in the absence of self-renewal

(A–J) ACTA2-Cre^ERT2:Lef-KI^+/− mice were subjected to the injury protocol in Figure 5G and tracheal sections immunostained for the indicated antigens (B&W inset of boxed regions in G and I show a Trop2⁺ SMG duct and αtubulin⁺ ciliated ducts, respectively). Arrows mark duct openings at the SAE. All images are from the C0–C4 region, except for (B) which is at C6. (K) Quantification of lineage-traced (GFP⁻) and untraced (GFP⁺) club and ciliated cells as a % of total cells in the SAE (C0–C4) from experiments in (A–J). (L) Experimental design for evaluating how Lef-1 expression in ACTA2-Cre^ERT2:Lef-KI^+/+ mice impacts self-renewal of MECs following sequential SO₂ (600 ppm) injury in (M,N,O). (M,N) GFP and Lef-1 immunostaining of tracheas from two (M) induced/uninjured and (N) induced/2x SO₂ injured ACTA2-Cre^ERT2:Lef-KI^+/+ mice. (O) Quantification of GFP⁻ and Lef-1⁺ cells as a % of total cells in the SAE and SMGs from experiments outlined in (L). Graphs show means +/− SEM for N=5 mice in (K) and N=3 mice per group in (O). Micron bars: 50 μm. Asterisks indicate significance of (K) paired two-tailed Student’s t-tests and (O) two-way ANOVA followed by Sidak’s multiple comparisons test (*P<0.05 and ****P<0.0001). See supplemental Figure S4.

Figure 7. Lef-1 expression in MECs induces regenerative and basal cell transcriptional programs

MECs were isolated from tamoxifen-induced ACTA2-Cre^ERT2:ROSA-TG (N=4) and ACTA2-Cre^ERT2:Lef-KI^+/+ (N=5) mice and purified by FACS at P1 in culture for RNAseq. (A) Heat map of 360 differentially expressed genes following unsupervised hierarchical clustering (Benjamini-Hochberg adjusted P<0.05). (B) Lef-1 expression levels in the two genotypes. (C) Principle component analysis (PCA) of the 13,337 genes expressed in the two groups. (D) IPA biological processes and functions defined by gene expression patterns showing p-values and z-scores. (E–I) Heat maps of the indicated IPA gene sets following unsupervised hierarchical clustering. (J,K) Motility assays on purified MEC^WT and MEC^Lef-1KI+/+ in culture showing (J) migration plot and (K) distance traveled with time. Measurements were taken of N=16 randomly selected cells traced from N=3 cultures for each genotype from a single experiment; mean +/− SEM is shown in K with the P-value indicating the comparison between nonlinear models fitting MEC^WT and MEC^Lef-1KI cells. (L,M) Heat maps of (L) differentially expressed transcription factors and (M) basal cell specific genes determined in Figure S5. In all heat maps, red indicates positive enrichment while blue indicates negative enrichment. See supplemental Figures S5–S7 and Tables S1–S3.