Fstl1 determines luminal size of mouse trachea through negative regulation of BMP4 signaling Authors: Yueyue Jin, Yue Zhang, Yingying Liu, Maolin Yao, Chenjie Zhang, Shuzi Li, Yibo Ding, Zhuan Niu, Yue Han, Yan Geng, Lian Li, Xue Liu, Wen Ning Published: 2026 ABSTRACT SummaryTrachea is a wide tube with smooth luminal epithelia that enables efficient ventilation. In this study, we provide the evidence that the mitotic angles of most tracheal epithelial cells exhibit ordered and transverse orientations, which allows the tracheal lumen to grow along the anterior-posterior (AP) axis, with its circumference being higher than length. We show that the BMP4 signaling antagonist follistatin-like 1 (Fstl1) is a critical regulator for tracheal morphogenesis. Mice with Fstl1 deletion exhibit wider trachea with enlarged lumen as a result of the increased proportion of mitotic cells with their spindles nearly vertical to the tracheal longitudinal axis. Mechanistically, Fstl1 regulates BMP4 signaling by antagonizing p-Smad1/5/8 in tracheal epithelial cells. Reducing BMP4 signaling activity rescues lumen enlargement and normalizes the spindle angle distribution in Fstl1-deficient tracheas. Therefore, we provide evidence to demonstrate that Fstl1 modulates the oriented cell division and tracheal morphogenesis, in part, through BMP4 signaling.Keywords: tracheal morphogenesis, luminal epithelial morphogenesis, spindle angle distribution, Follistatin-like 1, Fstl1, BMP4 signaling Introduction The trachea is an exclusive passage that connects the larynx to the bronchi to deliver external air to the lungs. To ensure ventilation efficiency, the trachea develops a wide tubular morphology with smooth luminal epithelium, surrounded by C-shaped cartilage rings and smooth muscle stripe (SM), which balance the rigidity and elasticity of the airway to prevent collapse.1,2 In mice, the tracheal morphology occurs following lineage specification (embryonic day 8.5 [E8.5]–E10.5) and undergoes sequential events of tracheoesophageal separation (E9.5–E12.5) and tube elongation/expansion (E10.5–E18.5).3 Tracheal morphology is determined during the elongation/expansion stage, which is initiated and regulated by the mesenchymal tissues, including the cartilage and SM.3,4,5,6 Anomalies in their formation can result in life-threatening malformation of the trachea, such as tracheomalacia and tracheostenosis.7,8 As the trachea grows in both length and width, its luminal epithelial surface enlarges whilst developing into pseudostratified epithelial layers at the lumen. The cellular basis determining the lumen size of developing mouse trachea has not yet been elucidated. Studies have revealed that the increase in lumen size mainly relies on cell proliferation during early development until E14.5 but on expansion of the cell apical surface and modest cell proliferation after E14.5.3,4 In the lungs, oriented division of proliferating cells by the RAS-ERK1/2 signaling has been reported to determine the lumen size of airway tubes.9 Tang et al.9 measured the mitotic spindle angle of airway epithelial cells and showed that longitudinal orientation allows the tube length to increase more than circumference. Although both the trachea and the lung airways elongate along the anterior-posterior (AP) axis in early stages, their lumen morphogenesis is quite different. The potential role of oriented cell division on tracheal lumen size determination and mechanisms underlying remain largely unknown. Follistatin-like 1 (Fstl1) is a small glycoprotein primarily synthesized by mesenchymal-derived cells and belongs to the SPARC (secreted protein acidic and rich in cysteine) family of matricellular proteins.10 Although functions and mechanisms of Fstl1 are not completely understood, a rapidly expanding body of literature has described the involvement of Fstl1 in embryonic development, tissue remodeling/repair, and inflammatory processes and has identified the regulatory role of Fstl1 in cell growth, differentiation, apoptosis, and migration. We have previously generated Fstl1−/− mice and observed the postnatal lethality as a result of respiratory failure. Analyses of the mutant phenotypes have revealed the important role of Fstl1 in the differentiation/maturation of alveolar epithelial cells and demonstrated that Fstl1 is a bone morphogenetic protein (BMP) 4 signaling antagonist that plays a role in controlling mouse lung development.11 Meanwhile, we have also observed malformed tracheal tubes with interrupted cartilages, narrowed SM, as well as enlarged lumen size in Fstl1−/− mice (E18.5).11,12 In this study, we measured the mitotic spindle angles of the dividing cells and found that most tracheal epithelial cells are transversely oriented with their spindles nearly vertical to the tracheal longitudinal axis. We report that the deletion of Fstl1 in mice leads to lumen enlargement due to an increased proportion of mitotic epithelial cells with transverse orientation. Reducing BMP4 signaling activity by dorsomorphin or Noggin normalized spindle angle distribution and rescued lumen enlargement in Fstl1−/− tracheas. Our study highlights that the Fstl1-downregulated Bmp4 axis plays an important role in the maintenance of the biased spindle angle distribution for lumen size determination of the developing mouse trachea. Results Lumen enlargement in Fstl1−/− tracheas We have previously shown that Fstl1−/− neonates have unusually enlarged tracheal tubes with irregular inner margins.11 To examine the defects of luminal epithelial morphogenesis in Fstl1−/− tracheas in detail, we collected embryos from the elongation/expansion stage at indicated developmental days and performed histological analysis on the transverse sections from the middle level of tracheas. We found that the enlarged lumen size in Fstl1−/− tracheas was evident as early as E11.5 and that this enlargement continued to increase significantly along with the tracheal development, until postnatal day 0 (P0) (Figures 1A and S1A). Quantitative analysis confirmed the larger lumen size in Fstl1−/− tracheas than that of age-matched wild type (WT, Fstl1+/+) littermates at all indicated developmental days, as determined by tracheal lumen circumference and area (Figures 1B and 1C; Figures S1B and S1C). We also isolated embryonic tracheas with lung tissues (E11.5–E18.5) and performed whole-mount immunofluorescence staining for the epithelial marker E-cadherin. This staining highlighted epithelial cells lining the tracheal lumina, providing a clear visualization of the lumens (Figure 1D). Consistent with the above histological findings, Fstl1−/− tracheas displayed enlargement of the lumens compared with those of the age-matched Fstl1+/+ littermates, as determined by tracheal lumen diameters (Figures 1D and 1E). We also measured their length and found that it was comparable between Fstl1+/+ and Fstl1−/− tracheas at all stages examined (Figure 1F). Enlarged tracheal lumen in Fstl1-deficient embryos (A) Representative images of H&E staining on tracheal transverse sections from Fstl1+/+ and Fstl1−/− mouse embryos at different developmental stages. Scale bars: 100 μm. (B and C) Tracheal lumen circumference (B) and area (C) of embryos at different developmental stages were quantified by bar plots. E11.5: Fstl1+/+, n = 11, Fstl1−/−, n = 8; E12.5: n = 5; E13.5: n = 3; E15.5: n = 5; E18.5: Fstl1+/+, n = 4, Fstl1−/−, n = 8. ∗, p < 0.05, ∗∗, p < 0.01, and ∗∗∗, p < 0.001. (D) Representative images of whole-mount staining of embryonic tracheas and lungs for the epithelial marker E-cadherin to visualize the tracheal lumens. Scale bars: 500 μm. (E) Quantification of tracheal lumen diameters after whole-mount staining. E11.5: Fstl1+/+, n = 8, Fstl1−/−, n = 3; E12.5: Fstl1+/+, n = 13, Fstl1−/−, n = 9; E13.5: Fstl1+/+, n = 21, Fstl1−/−, n = 6; E15.5: Fstl1+/+, n = 11, Fstl1−/−, n = 5; E18.5: Fstl1+/+, n = 5, Fstl1−/−, n = 3. ∗, p < 0.05; ∗∗∗∗, p < 0.0001. (F) Quantification of tracheal tube length after whole-mount staining. E11.5: Fstl1+/+, n = 6, Fstl1−/−, n = 4; E12.5: Fstl1+/+, n = 11, Fstl1−/−, n = 7; E13.5: Fstl1+/+, n = 16, Fstl1−/−, n = 5; E15.5: Fstl1+/+, n = 7, Fstl1−/−, n = 4; E18.5: Fstl1+/+, n = 4, Fstl1−/−, n = 3. Data are expressed as the mean ± SEM. Differences between the groups were assessed using Student’s t tests. Of note, both Fstl1+/+ and Fstl1−/− tracheas (E11.5–E13.5) exhibited a round tubular shape lined by a smooth epithelial lumen. However, starting from E15.5, we observed that the tracheal tubes of Fstl1−/− embryos were deformed. At E18.5 and P0, the tracheal tubes were severely deformed, and their luminal epithelial layers were irregular with folds and protuberances (Figures 1A and S1A), as previously reported.11 These findings indicate an atelectatic phenotype of Fstl1−/− tracheas at the late stage of development due to the insufficient support from truncated cartilage rings and impaired SM.12 This atelectatic phenotype became more severe at birth (P0), as indicated by the increased lumen circumference but comparable lumen area in Fstl1−/− tracheas (Figures S1A–S1C). We also examined the lumen size of the main bronchi. Histological analysis was conducted on transverse sections of the trachea, bifurcation sites, and main bronchi from E12.5 embryos (Figures S1D and S1E). Consistently, the lumens of the Fstl1−/− trachea, bifurcation, and both left and right bronchi were significantly enlarged. This lumen enlargement was further validated through quantitative measurements of the lumen circumference and area (Figures S1F and S1G). Increased cell proliferation in Fstl1−/− tracheal epithelium During early development of the trachea, the luminal epithelial morphogenesis mainly relies on cell proliferation.3 To examine whether the enlarged lumen in Fstl1−/− tracheas is associated with the negative regulation of Fstl1 on epithelial proliferation,11 we conducted bromodeoxyuridine (BrdU) incorporation assay. As shown in Figures S2A and S2B, cell proliferation increased along with the tube elongation/expansion, reaching the peak at E13.5 and decreasing at E15.5. This is in agreement with the recent report that tracheal epithelial proliferation decreases after E14.5.3 As expected, Fstl1 deletion increased cell proliferation during all developmental stages in Fstl1−/− tracheal epithelium, as quantitated by the percentages of BrdU+ cells at indicated developmental days (Figures S2A and S2B). After E14.5, when cell proliferation decreased, the luminal size increased by cell rearrangement, expansion of the apical surface, and modest cell proliferation.3,4 To examine whether changed cell rearrangement also contributed to the lumen enlargement of Fstl1−/− tracheas after E14.5, we measured the apical surface of luminal epithelial cells (E18.5). Integration of quantitative values confirmed the modest increases in cell numbers and significant increases in apical surface in Fstl1−/− tracheal epithelium (Figures S2C–S2E). Transverse orientation in dividing tracheal epithelial cells The morphogenesis and maintenance of tubular organs rely on the proper orientation of the mitotic spindle during cell division.13 To identify oriented cell division of the tracheal epithelial cells and to address whether changes in their mitotic spindle angles also account for the enlarged lumen in Fstl1−/− trachea, we adapted a geometric model based on a previously reported approach9 and measured the spindle angles in individual mitotic cells. In this model, the mitotic angles of tracheal epithelial cells were measured by the distances between the mitotic spindle poles both perpendicular and parallel to the tracheal longitudinal axis. Epithelial cells were identified via E-cadherin staining, and mitotic spindle poles were marked using γ-tubulin. z stacks of spindle poles in dividing cells at anaphase and telophase were acquired using Leica TCS SP5 and Zeiss LSM710 confocal microscopes. The distance parallel to the tracheal longitudinal axis (Δz) was calculated from the two z stacks, and the confocal images from these z stacks were merged to calculate the distance perpendicular to the longitudinal axis (Δx − y), using ImageJ software. The mitotic angle (θ) was then computed using the formula: θ = ATAN[(Δx − y)/Δz] × 180/π (Figure 2A). Smaller mitotic angles (θ) were associated with biased contribution to tracheal length, while larger mitotic angles contributed to tracheal circumference (Figure 2B). Increased mitotic spindle angles in tracheal epithelial cells of the Fstl1-deficient embryos (A) Schematic graph of the mitotic spindle angles in the tracheal epithelial cells. (B) Representative cells with different mitotic angles. Immunostaining for centrosomes (γ-tubulin), epithelial cells (E-cadherin), and nuclei (DAPI) was performed to determine the mitotic spindle angles of the tracheal epithelial cells. Scale bars: 2.5 μm. (C) Data chart of the percentage change of mitotic spindle angles at E12.5, E13.5, and E15.5 in Fstl1+/+ tracheal epithelial cells. (D and E) Violin plots (D) and Rose plots (E) of mitotic spindle angles of tracheal epithelial cells at E12.5 in Fstl1+/+ and Fstl1−/− mouse embryos. Fstl1+/+, n = 114; Fstl1−/−, n = 99. ∗∗, p < 0.01. (F and G) Violin plots (F) and Rose plots (G) of mitotic spindle angles of tracheal epithelial cells at E13.5 in Fstl1+/+ and Fstl1−/− mouse embryos. Fstl1+/+, n = 163; Fstl1−/−, n = 225. ∗∗∗∗, p < 0.0001. (H and I) Violin plots (H) and Rose plots (I) of mitotic spindle angles of tracheal epithelial cells at E15.5 in Fstl1+/+ mouse embryos. Fstl1+/+, n = 35. Data are expressed as the mean ± SEM. Differences between the groups were assessed using Student’s t tests. We first measured the mitotic spindle orientation of dividing epithelial cells in Fstl1+/+ tracheas at indicated developmental days. Unlike the longitudinal orientation in lung epithelium,9 a large proportion of mitotic cells (E12.5, 74.6%; E13.5, 72.4%; E15.5, 91.4% of mitotic cells) in tracheal epithelium had their spindles tending to be vertical to the tracheal longitudinal axis (θ = 60°–90°) (Figures 2C–2I). These data indicated that the transverse orientation determines the wide lumen morphogenesis of tracheal tube while elongating along the AP axis. Abnormal spindle angle distribution in proliferating cells of Fstl1−/− tracheal epithelium To address whether changes of mitotic spindle angle in the epithelial cells account for the enlarged lumen in Fstl1−/− trachea, we next measured spindle angles in Fstl1−/− tracheas. Because of the deformed tubular shapes at late stages (after E15.5), E12.5 and E13.5 Fstl1−/− tracheas with round tube shape were used to study changes in the oriented cell division. In E12.5 Fstl1−/− tracheal epithelium, more mitotic cells had their spindles vertical to the tracheal longitudinal axis than those in Fstl1+/+ (θ peaks at 70°–90°; Fstl1+/+, 57.9%; Fstl1−/−, 69.7%) (Figures 2D and 2E). This bias was more pronounced in E13.5 Fstl1−/− tracheal epithelium, with a higher proportion of mitotic cells and more vertical spindle angles (Fstl1+/+, θ peaks at 60°–80°, 54.3%; Fstl1−/−, θ peaks at 70°–90°, 85.8%) (Figures 2F and 2G). These data highlighted a clear shift in spindle orientation distribution associated with Fstl1 deficiency. Comparable ERK1/2 signaling in Fstl1−/− tracheal epithelial cells RAS-regulated ERK1/2 signaling has been reported to control the mitotic spindle angles of airway epithelial cells in the lungs.9 To investigate the molecular mechanisms whereby the deficiency of Fstl1 results in lumen enlargement of mouse trachea described above, we first examined the ERK1/2 signaling activity. Immunostaining for p-ERK1/2 on the transverse sections of E12.5 and E13.5 embryos showed comparable ERK1/2 activity in Fstl1+/+ and Fstl1−/− tracheal epithelial cells (Figure S3A). In addition, in vivo inhibition of the ERK1/2 pathway using the specific inhibitor PD0325901 failed to rescue the enlarged tracheal lumen in E12.5 and E13.5 Fstl1−/− embryos (Figures S3B and S3C). Ectopic activation of BMP4 signaling in Fstl1−/− tracheal epithelium Because Fstl1 acts as a BMP4 antagonist in lung and ureter development,11,14 we next examined BMP4 signaling activity. To our surprise, Fstl1−/− tracheal epithelial cells, particularly the proliferating cells, exhibited significantly elevated immunostaining signals of p-Smad1/5/8, a reliable marker of BMP4 canonical signaling activity, in both E12.5 (Figure 3A) and E13.5 (Figure 3B) embryos. Western blot analysis using total protein extracts from E16.5 (Figure 3C) and E18.5 (Figure 3D) tracheal tissues further confirmed higher levels of p-Smad1/5/8, whereas p-ERK1/2 levels remained unchanged (Figures 3C and 3D) in Fstl1−/− tracheas. We also identified proximal airway epithelial cells (Epcam+Sox2+) from our previously published single-cell RNA sequencing (scRNA-seq) dataset on E18.5 mouse lungs.15 Bar plots further confirmed that the expression of BMP4 signaling downstream targets (Id1, Id2, and Id3) were markedly upregulated in the proximal airway epithelial cells from Fstl1−/− lungs compared with those from Fstl1+/+ lungs (Figure 3E). These findings indicated that BMP4 signaling, rather than ERK1/2 signaling, is aberrantly activated in Fstl1−/− tracheal epithelial cells, potentially contributing to abnormal orientated cell division and the enlarged tracheal lumen phenotype. Ectopic activation of BMP4 signaling in tracheal epithelial cells of the Fstl1-deficient embryos (A and B) Immunostaining for p-Smad1/5/8 was performed on E12.5 (A) and E13.5 (B) in Fstl1+/+ and Fstl1−/− mouse tracheal sections. n = 3. Arrows indicate the proliferating epithelial cells. Scale bars: 10 μm. (C and D) Western blot for p-Smad1/5/8 and p-ERK1/2 in E16.5 (C) and E18.5 (D) Fstl1+/+ and Fstl1−/− mouse tracheal tissues. Data are representative of two independent experiments. (E) Bar plots showing the expression of canonical BMP4 target genes Id1, Id2, and Id3 in Epcam+Sox2+ proximal airway epithelial cells from scRNA-seq in E18.5 Fstl1+/+ and Fstl1−/− lungs (GSE225463). Data are expressed as the mean ± SEM. Differences between the groups were assessed using Student’s t tests. Comparable Wnt5a-Ror2 signaling pathway in Fstl1−/− tracheas Multiple signaling pathways, including the mesenchymal Wnt5a-Ror2 axis,4 have been implicated in tracheal epithelial polarization and tube morphogenesis. To determine whether this pathway contributes to the defects observed in Fstl1−/− tracheas, we examined its activity at both protein and transcript levels. Western blot analysis of E16.5 and E18.5 tracheal tissues (Figures S4A and S4B) showed no significant changes in Wnt5a or Ror2 expression in Fstl1−/− samples. Consistently, expression analysis of E18.5 tracheas (Figure S4C) revealed no significant differences in Wnt5a, Ror2, or the downstream target Axin2 between Fstl1+/+ and Fstl1−/− tracheas. These results suggest that the Wnt5a-Ror2 signaling pathway is not significantly altered in the absence of Fstl1 and is unlikely to account for the observed phenotypes. Blocking BMP4 signaling activity rescues lumen enlargement in Fstl1−/− tracheas To test if the activated BMP4/Smad signaling plays a causative role in the development of lumen enlargement in Fstl1−/− tracheas, we injected pregnant mice with different doses of dorsomorphin, a selective BMP4 signaling inhibitor, at the indicated time points, and isolated E13.5 embryonic tracheas for whole-mount E-cadherin immunostaining. As shown in Figures 4A and 4B, Fstl1−/− tracheas displayed the phenotype of lumen enlargement with the tube diameter 97.5% wider than that of Fstl1+/+ controls. By contrast, dorsomorphin injections decreased the tracheal lumen size of Fstl1−/− embryos, with the tracheal lumen being 61.4% wider than those of Fstl1+/+ controls at 3 doses from E10.5 to E13.5 and 23.4% wider at 5 doses from E8.5 to E13.5, while 8 doses from E5.5 to E13.5 almost restored the tracheal lumen to the normal size. These findings indicated that the abnormal lumen enlargement seen in Fstl1−/− embryos was partially rescued by BMP4/Smad signaling inhibition. Histological examination of E13.5 embryos treated with 8 doses of dorsomorphin confirmed the partially rescued lumen enlargement in Fstl1−/− embryos, as determined by the decreased tracheal lumen circumference and area (Figures 4C–4E). Violin plots and Rose plots further confirmed a smaller proportion of mitotic cells with smaller vertical spindle angles (Fstl1−/−, 69.3% cells with θ peaking at 70°–90°; Fstl1−/− + dorsomorphin, 56.3% cells with θ peaking at 60°–80°) in tracheal epithelial cells of E13.5 Fstl1−/− embryos administered with 8 doses of dorsomorphin compared with Fstl1+/+ controls (44.1% cells with θ peaking at 60°–80°) (Figures 4F and 4G; Figure S5A). Blocking BMP4 signaling rescued enlarged tracheal lumen in Fstl1-deficient embryos (A and B) Whole-mount staining of E-cadherin (A) and quantification of the tracheal lumen diameters (B) in Fstl1+/+ and Fstl1−/− mouse tracheas after different doses of DMSO or dorsomorphin treatment in vivo. DMSO: Fstl1+/+, n = 7, Fstl1−/−, n = 4; dorsomorphin: E10.5–E13.5, Fstl1+/+, n = 5, Fstl1−/−, n = 4; E8.5–E13.5, Fstl1+/+, n = 4, Fstl1−/−, n = 4; E5.5–E13.5, Fstl1+/+, n = 4, Fstl1−/−, n = 3. Scale bars: 500 μm ∗, p < 0.05; ∗∗∗, p < 0.001. Differences between the groups were assessed using Student’s t tests. (C–E) Representative images of H&E staining (C), and the quantification of tracheal lumen circumference (D) and area (E) at E13.5 in Fstl1+/+ and Fstl1−/− mouse embryos with daily DMSO or dorsomorphin treatment in vivo from E5.5 to E13.5. n = 6. Scale bars, 100 μm ∗, p < 0.05; ∗∗, p < 0.01; ∗∗∗, p < 0.001; ∗∗∗∗, p < 0.0001. Differences between the groups were assessed using one-way ANOVA. (F and G) Violin plots (F) and Rose plots (G) of mitotic spindle angles of tracheal epithelial cells at E13.5 in Fstl1+/+ and Fstl1−/− mouse embryos with daily DMSO or dorsomorphin treatment in vivo from E5.5 to E13.5. ∗∗, p < 0.01. Differences between the groups were assessed using one-way ANOVA. Data are expressed as the mean ± SEM. In addition, we employed an explant culture model with E12.5 lung,16 and treated the cultures with dorsomorphin to block BMP4 signaling activity. After ex vivo culture for 48 h, dorsomorphin treatment decreased the diameter of Fstl1−/− tracheal tubes, resulting in comparable tracheal tube size between Fstl1−/− and Fstl1+/+ controls (Figures S5B and S5C). Collectively, Fstl1 negatively regulates BMP4/Smad signaling, which controls oriented cell division and consequently determines the lumen size of the tracheal tube. BMP4/Smad signaling influences spindle orientation in BEAS-2B cells To further examine the role of BMP4 signaling in regulating spindle orientation in tracheal epithelial cells, we treated BEAS-2B cells, a human proximal bronchial epithelial cell line, with a recombinant human BMP4 protein (20 ng/mL). BMP4 stimulation significantly activated BMP4 signaling but not ERK1/2 signaling, as determined by the increased levels of p-Smad1/5/8 but comparable p-ERK1/2 (Figure S6A), as well as elevation of BMP4 downstream genes (Figure S6B). We calculated the mitotic angle (α) of epithelial cells based on the vertical distance (Δz) and horizontal distance (10) between the mitotic spindle poles labeled with γ-tubulin, using the formula α = ATAN(Δz/x) × 180/π (Figure S6C). As shown in Figures S6D and S6E, compared with a large proportion of BEAS-2B cells (92%) dividing parallel to the horizontal axis (α peaks at 0°–10°), BMP4 treatment significantly disrupted the bias, and spindle orientation distribution shifted to a higher angular range (α peaks at 0°–50°). To determine whether the role of BMP4 signaling-regulated spindle orientation is specific to the tracheal epithelial cells, we extended our analysis to MLE12 cells, a mouse alveolar epithelial cell line, following BMP4 treatment. As shown in Figures S6F and S6G, BMP4 did not induce any changes in the mitotic spindle orientation in MLE12 cells. These results suggest that the above mechanism is specific to the tracheal epithelium and does not extend to the lung alveolar epithelial cells. Fstl1-downregulated BMP4/Smad signaling influences spindle orientation in BEAS-2B cells Fstl1 is a mesenchymal-derived matricellular protein. Using a Fstl1 expression reporter mouse line (Fstl1-LacZ), we have previously shown a clear Fstl1 expression in subepithelial mesenchymal cells surrounding the tracheal epithelium (E18.5).12 To further characterize the expression pattern of Fstl1 during early lung development, we performed an integrative analysis of published scRNA-seq datasets from E8.0–E9.5 mouse foregut17 and E9.5–E11.518 and E12.519 mouse embryonic lungs. Comparative analysis revealed that Fstl1 exhibited a high but dynamic spatial-temporal expression pattern in the mesoderm/mesenchymal compartment. In contrast, several other BMP4 antagonists, including Nog, Fst, Chrd, Grem1, and Sostdc1, showed comparatively lower expression levels during these stages (Figure S7A), suggesting a predominant role for Fstl1 in modulating BMP4 signaling in lung mesenchyme during early development. We next compared the spatial and temporal expression patterns of Fstl1 and Bmp4. As shown in Figures 5A and 5B, Fstl1 expression is generalized in both mesoderm and endoderm of the foregut but becomes highly enriched in the mesoderm in later tracheal formation stages, whereas Bmp4 transcripts are localized mainly in the mesoderm, with little expression detected in the endoderm. Of note, after E11.5, we observed a significant increase in Fstl1 expression but a remarkable decrease in Bmp4 expression (Figures 5A and 5B). To further validate these observations in the trachea, RNAscope analysis of E13.5, E15.5, and E18.5 sections confirmed that Fstl1 is primarily localized to the mesenchymal cells surrounding the airway epithelium. Importantly, loss of Fstl1 led to a marked increase in Bmp4 transcript levels within the tracheal mesenchyme of Fstl1−/− embryos (Figure 5C), supporting a regulatory relationship between Fstl1 and Bmp4 signaling in vivo. BMP4 signaling is downregulated because of the expression of Fstl1 in the tracheal mesenchyme (A and B) Violin plots showing the expression of Fstl1 (A) and Bmp4 (B) in published scRNA-seq data on mouse E8.5/9.0/9.5 foregut (GSE136689), E9.5/10.5/11.5 (GSE87038), and E12.5 (GSE119228) lungs. (C) RNAscope in situ hybridization of Fstl1 and Bmp4 on tracheal transverse sections in Fstl1+/+ and Fstl1−/− mouse embryos at E13.5, E15.5, and E18.5. Dashed boxes mark regions displayed at a higher magnification, and dashed lines delineate the boundary between the epithelium and mesenchyme. Scale bars: 50 μm. To further validate the role of Fstl1 in regulating epithelial mitotic spindle orientation through the BMP4/SMAD signaling pathway, we isolated mouse embryonic fibroblasts (MEFs) from E13.5 Fstl1−/− embryos. These cells showed a marked reduction in Fstl1 expression and secretion (Figures S7B and S7C), accompanied by an increase in Bmp4 expression and secretion (Figures S7D–S7F). We then employed a transwell co-culture system with BEAS-2B epithelial cells in the upper chamber and MEFs in the lower plate (Figure 6A). To assess the specificity of BMP4 signaling, we also included Noggin, a well-characterized BMP4 antagonist, as an additional control. As expected, we found that co-culture with Fstl1−/− MEFs increased the level of p-Smad1/5/8 and expression of BMP4 target genes (ID1, ID2, and ID3) in BEAS-2B cells (Figures 6B and 6C), and subsequently increased the proportion of cells with mitotic angle distributing to the angular range (α peaks at 0°–60°). Treatment with Noggin decreased the levels of p-Smad1/5/8 and the proportion of cells skewed toward low angles (Figures 6D and 6E). These findings establish a functional role for Fstl1 in regulating epithelial BMP4/Smad1/5/8 signaling and mitotic angles in epithelial cells, thereby influencing tracheal shape formation in mice. Fstl1-downregulated BMP4/Smad signaling influences spindle orientation in BEAS-2B cells (A) Schematic graph of the BEAS-2B cells and MEF co-culture system and the mitotic spindle angles of BEAS-2B cells. (B) Western blot analysis of p-Smad1/5/8 and Smad5 in BEAS-2B cells co-cultured with Fstl1+/+ and Fstl1−/− MEFs with PBS or Noggin (25 ng/mL). Data are representative of three independent experiments. (C) BEAS-2B cells were co-cultured with Fstl1+/+ and Fstl1−/− MEFs and treated with PBS or Noggin (25 ng/mL), and relative expression of ID1, ID2, and ID3 was determined by RT-qPCR. n = 3. ∗, p < 0.05; ∗∗∗, p < 0.001. (D and E) Violin plots (D) and Rose plots (E) showing the mitotic spindle angles of BEAS-2B cells co-cultured with Fstl1+/+ and Fstl1−/− MEFs and treated with PBS or Noggin (25 ng/mL). PBS: Fstl1+/+, n = 89, Fstl1−/−, n = 91; Noggin: Fstl1+/+, n = 95, Fstl1−/−, n = 94. ∗∗∗∗, p < 0.0001. Data are expressed as the mean ± SEM. Differences between the groups were assessed using one-way ANOVA. Discussion Understanding the mechanisms underlying tube formation is a fundamental goal in developmental biology. The reported roles of epithelial cells in the development of the wide lumen of tracheal tubes correlate with their proliferation, shape change, and arrangement.3,4 The findings reported here add new insights into our understanding of tracheal tubulogenesis and demonstrate the oriented epithelial division may influence the size of the tracheal lumen. We provide data from knockout mouse models, in vivo pharmacological intervention, in vitro co-culture system, as well as single-cell transcriptome analysis to support the role of Fstl1-regulated Bmp4/Smad signaling axis in the determination of the oriented epithelial division and subsequent tracheal luminal growth. Although the luminal epithelial morphogenesis of the trachea and lungs mainly relies on cell proliferation, the regulatory mechanisms are not the same. In the lungs, the dividing cells are oriented longitudinally by default, with a large proportion of epithelial cells dividing parallel to the airway longitudinal axis, allowing the tube length to increase more than circumference as it grows along the AP axis at the early stage.9 In contrast, Kishimoto et al. observed that the epithelial cells of the trachea exhibit random orientations.3,4 Here, we adopted the geometric model as reported by Tang et al.9 and marked the mitotic spindle poles by using γ-tubulin, instead of using nuclei. We measured the mitotic spindle angle of the dividing tracheal epithelial cells at different developmental days (E12.5, E13.5, and E15.5) and showed that tracheal epithelial cells are oriented and that most mitotic cells divide with their spindles nearly vertical to the tracheal longitudinal axis. Their transverse orientation shapes the development of the wide lumen of the tracheal tube as it extends along the AP axis. We further showed that this process involves canonical BMP4/Smad1/5/8 signaling, distinguishing it from lung airway tube morphogenesis, in which RAS/ERK1/2 signaling determines longitudinal orientation.9 Altogether, our data provide evidence to support that lumen epithelial morphogenesis of the trachea and lungs is quite different. Fstl1, an antagonist for BMP4 signaling, has been reported in alveologenesis and regeneration of alveolar epithelium from alveolar type 2 (AT2) stem cells.11,20 Here, we offer the following lines of evidence to demonstrate the critical role of the BMP4 antagonist Fstl1 in determining luminal epithelial morphogenesis of trachea: targeting deletion of Fstl1 in mice leads to tracheal lumen size to grow wider; Fstl1 exerts its effect on luminal size through the regulation of oriented cell division, together with cell proliferation and cell shape. Fstl1-reduced BMP4/Smad signaling primes or allows tracheal epithelial cells to divide with transverse orientation, determining the development of the wide lumen of the tracheal tube. Pharmacological intervention on BMP4 signaling with dorsomorphin partially recovers the tracheal lumen size of Fstl1−/− embryos and mitotic spindle angle in vivo and ex vivo. BMP4 signaling has been implicated in the tubulogenesis of trachea, especially dorsal-ventral (DV) pattern establishment and tracheal lineage specification.3,7 In the current mouse model, we observed the upregulation of phospho-Smad1/5/8, together with the increased expressions of Id1, Id2 and Id3—established BMP/SMAD downstream targets—in the wider Fstl1−/− luminal epithelium. Reducing BMP4 signaling activity by Noggin partially rescued the mitotic spindle angle and reversed the increased ID1, ID2, and ID3 expressions in transwell-cultured BEAS-2B cells in vitro. Together with the in vivo pharmacological intervention data, the findings support our hypothesis that the BMP4 signaling axis is crucial for luminal epithelial morphogenesis and that Fstl1 affects oriented cell division and subsequent lumen growth in the developing trachea partially through modulating the BMP4 signaling axis. Although multiple signaling pathways have been implicated in the development of trachea, no evidence of the interaction between Fstl1 and ERK1/2 signaling pathway or Fstl1 and Wnt5a-Ror2 signaling pathway was found in our Fstl1−/− embryos. Fstl1 may orchestrate with other BMP antagonists via their individual spatiotemporal expression patterns and control BMP4 signaling in the developing trachea. BMP4 signaling is downregulated in the dorsal foregut because of the expression of Noggin in the dorsal endoderm foregut in early embryo.3 Fstl1 starts to be expressed at E7.21 Its expression level increases during the development of trachea, being ubiquitous until E14.5 and restricted to the mesenchyme surrounding the tracheal tube at later stages.3 Single-cell transcriptome analyses indicate a mutually exclusive spatiotemporal expression pattern between Noggin and Fstl1, indicating the balance of BMP4 signaling by different antagonists temporally, with Noggin being predominant in stage of DV patterning (E8.5–E10.5), whereas Fstl1 being pronounced in the tube elongation/expansion stage (E10.5–18.5). Almost no expression of other BMP4 antagonists, including Fst, Grem1, and Chrd, in trachea further suggests the importance of Fstl1 on BMP4 signaling regulation in the developing trachea. In summary, our findings reveal a previously unrecognized mechanism underlying tracheal tubulogenesis, demonstrating that oriented epithelial cell division plays a critical role in determining the luminal size. We show that, unlike in the lung, tracheal epithelial cells predominantly divide with a transverse orientation, contributing to lumen widening along the AP axis. Importantly, we identify the Fstl1-BMP4/Smad signaling axis as a key regulator of this process, integrating control of mitotic orientation with epithelial proliferation and cell shape. Through complementary in vivo and in vitro approaches, we establish that Fstl1 modulates BMP4 signaling to ensure proper tracheal epithelial morphogenesis and that the disruption of this balance leads to aberrant lumen expansion. Our work not only distinguishes the mechanisms governing tracheal versus lung development but also highlights a novel functional role for BMP4 antagonism in directing oriented cell division. These insights advance our understanding of early tracheal development and provide a conceptual framework for how signaling pathways coordinate epithelial architecture during organogenesis. Limitations of the study In this study, we report that Fstl1 negatively regulates BMP4/Smad1/5/8 signaling and that the deficiency of Fstl1 leads to abnormal lumen enlargement and disrupted tube morphology. Fstl1 is a matricellular protein produced mainly by cells of mesenchymal origin.12,21,22,23 We have previously shown the heterogeneity of Fstl1 expression in multiple mesenchymal subpopulations in E18.5 lungs by scRNA-seq analysis,15 but it remains to be clarified which mesenchymal subpopulations regulate the morphogenesis of trachea. We will use genetic lineage tracing, RNA-scope and scRNA-seq approaches to explore this issue. Resource availability Lead contact Requests for further information and resources should be directed to and will be fulfilled by the lead contact, Wen Ning (ningwen108@nankai.edu.cn). Materials availability This study did not generate new unique reagents. Data and code availability •This paper analyzes existing, publicly available data, accessible at GEO: GSE136689, GSE87038, GSE119228, and GSE225463. This paper analyzes existing, publicly available data, accessible at GEO: GSE136689, GSE87038, GSE119228, and GSE225463. •This paper does not report original code. This paper does not report original code. •Original western blotting images have been deposited at supplemental PDF and are publicly available as of the date of publication. Microscopy data reported in this paper will be shared by the lead contact upon request. Original western blotting images have been deposited at supplemental PDF and are publicly available as of the date of publication. Microscopy data reported in this paper will be shared by the lead contact upon request. •Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request. Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.