By crossing the Tbx4-rtTA/TetO-Cre driver line to the floxed-Bmpr1a mice (Tbx4-rtTA/TetO-Cre/Bmpr1a^fx/fx), Bmpr1a was specifically knocked out in the lung mesenchyme with doxycycline (Dox) induction from the beginning of lung formation (E6.5, Suppl. Fig.1A). This was validated at the mRNA and protein levels (Suppl. Fig.1B and 1C). For example, in E15.5 mesenchyme-specific Bmpr1a conditional knockout (CKO), Bmpr1a immunostaining was absent in lung mesenchyme but present in the epithelia (Suppl. Fig.1C) while Bmpr1a was detected in both mesenchyme and epithelia of wildtype (WT) lungs. By gross view, early embryonic lung morphogenesis between Bmpr1a CKO lungs and WT controls was comparable prior to E13.5 (Fig.1A). Decreased epithelial branching and increased size of branching tips were observed from E14.5. As lung development progressed, the terminal airways in Bmpr1a CKO lungs displayed dilation and exhibited increasingly extensive cystic changes. However, the overall size of the lungs remained relatively unchanged. The dynamic change of the lung cysts in Bmpr1a CKO lungs was further analyzed in H&E-stained tissue sections (Fig.1B). The walls of the lung cysts gradually thinned, over the course of development, ultimately resulting in the formation of balloon-like cystic lesions by the end of gestation (E18.5). These findings were consistent with the overall morphology of the lungs. Interestingly, overall cell proliferation and apoptosis had no significant change between cystic Bmpr1a CKO and WT control lungs as measured by EdU labeling and TUNEL assay, respectively (Fig.1C-E).

To understand the molecular mechanisms underlying the aforementioned phenotypic changes, bulk RNA-seq was used to examine the differentially expressed genes (DEGs) between Bmpr1a CKO and WT lung tissues at E15.5. A total of 1,001 DEGs (FDR ≤ 0.05 and log₂FC ≥ 1) were identified, among which 547 genes were upregulated and 454 genes were downregulated in Bmpr1a CKO lungs as compared to the transcriptome of WT lungs (Fig.2A). The raw data are available in the Gene Expression Omnibus repository under accession number GSE97946 (token: kfofmkaqnpeflix). Gene Ontology (GO) enrichment analysis showed that lung mesenchymal Bmpr1a regulates a large number of genes involved in the Muscle System Process (P < 0.001, Fig. 2B).

As shown above in Fig.1, enlarged airways were initially seen in Bmpr1a CKO lungs from E14.5. Whole mount staining of smooth muscle actin and comprehensive 3D imaging revealed a specific absence of airway smooth muscle cells (SMCs) surrounding the dilated airway branching in Bmpr1a CKO lungs, while unaffected airway branches in Bmpr1a CKO lungs remained surrounded by SMCs, as compared to the WT lungs (Fig. 2C). As lung development proceeded to E15.5, the distal airway enlargement became increasingly severe and extensive, as indicated by the whole mount staining of Cdh1 (E-Cadherin), which outlines the airway epithelia. Reduction and even lack of airway SMCs was consistently found around the enlarged airways of E15.5 Bmpr1a CKO lungs (Fig. 2D). In contrast, vascular SMCs in Bmpr1a CKO lungs appeared to remain unaffected (Fig. 2E).

The Tbx4 driver line also targets lung pericytes and endothelial cells with early induction ¹⁶. However, Bmpr1a deletion in these cell lineages had no effects on these cells or the associated pulmonary vasculature, as measured by the relevant cellular markers (Cspg4 and Pecam1) at both the mRNA and protein levels (Fig.3A and B). Emerging evidence has suggested that the extracellular matrix (ECM), including the airway basement membrane and matrix fibril collagen, is critical to lung development and maturation ^8,17. Several major components of the ECM in the Bmpr1a CKO lungs, such as laminin and type III collagen, were examined at both the mRNA and protein levels. No significant changes were found, even in the areas surrounding the dilated airways and lung cysts when compared to the WT lungs (Fig.3A-3B).

However, the mesenchymal knockout of Bmpr1a caused significant reduction of elastin expression, as detected by real-time PCR and western blot (Fig3.A and D). In the Bmpr1a CKO lungs, co-immunostaining of epithelia (Cdh1), smooth muscle cells (Acta2), and elastin (Eln) showed that this decrease in elastin expression occurred primarily in the areas adjacent to the epithelia of enlarged airways with the deficiency of airway smooth muscle. This suggests that the elastin abnormality may be caused by airway SMC deficiency or by reduced elastin expression in adjacent epithelial cells (Fig.3C). To determine whether BMP signaling can directly regulate elastin expression in fetal lung mesenchymal cells, alteration of elastin protein expression upon BMP4 treatment in primary fetal lung mesenchymal cells was analyzed. A significant increase in elastin protein was observed in the cells treated with BMP4 (Fig.3E-3F). However, inclusion of the BMP type 1 receptor (BMPR1) inhibitor LDN193189 (LDN) fully blocked the BMP4-stimulated elastin upregulation.

Activation of the BMP receptor complex can trigger intracellular Smad-dependent and Smad-independent pathways ¹⁰. In line with this report, the downregulation of Smad1/5-mediated BMP signaling was also observed in the Bmpr1a CKO lungs, evident from the reduced phosphorylation of Smad1/5 (p-Samd1/5 in Fig.4A). The role of Bmpr1a in mediating Smad1/5 activation in lung mesenchymal cells was further investigated in primary culture of fetal lung mesenchymal cells isolated from WT fetal lung tissues. Treatment of these cell with BMP4 activated Smad1/5 phosphorylation, while the addition of BMPR1 inhibitor LDN reduced the BMP4-induced Smad1/5 phosphorylation (Fig. 4E). To determine whether the Smad1/5-mediated BMP canonical pathway controls airway SMC growth in vivo, Smad1 and Smad5 were both deleted in lung mesenchyme by crossing the floxed-Smad1 and -Smad5 mice with the Tbx4-rtTA/TetO-Cre driver. Although abnormal fetal lungs with hypoplastic development were present in the Smad1/5 double CKO mice, airway dilation and cystic lesions were not detected (Fig. 4B-4C). Furthermore, in the Smad1/5 CKO lungs, airway SMC growth was not affected, as indicated by the expression of Acta2 around the airways (Fig.4D). This different phenotypic observation suggests that airway SMC growth may be independent of Smad1/5-mediated pathway.

We then examined other major Smad-independent pathways in the Bmpr1a CKO lungs. Reduced phosphorylation of p38 (P-p38) MAP Kinase (MAPK), but not Jnk and Erk, was detected in the Bmpr1a CKO lungs (Fig.4A). Bmp4 treatment of the cultured primary lung mesenchymal progenitor cells also specifically activated p38, a response that was specifically blocked by the addition of the BMPR1 inhibitor LDN (Fig.4E). Additionally, the role of the BMP4-Bmpr1a-p38 pathway in regulating SMC-related gene expression was assessed in cultured fetal lung mesenchymal cells. As previously reported ¹⁸, Myocd is a key transcription co-factor in controlling airway SMC differentiation. Myh11 and Acta2, major contractile proteins, are two prominent SMC markers. Treatment of lung mesenchymal progenitor cells with BMP4 induced a pronounced increase in the expression of these SMC-related genes at both the mRNA and protein levels (Fig.4F-4G). Inhibition of either Bmpr1a by LDN or p38 by SB203580 (SB) significantly attenuated the promoting effect of BMP4 on SMC differentiation.

As shown in Fig. 2E, Bmpr1a CKO caused defective growth of airway SMCs in vivo, while vascular SMCs remained unaffected. To investigate the distinct effects of the Bmpr1a-mediated pathway on non-vascular SMCs vs. vascular SMCs, primary fetal lung mesenchymal cells were isolated from Tagln-YFP/Cspg4-DsRed double reporter mice, in which airway SMCs/myofibroblasts were solely labeled with YFP expression while vascular SMCs and pericytes were marked by both YFP and DsRed expression (Fig.5A). Following fluorescence-activated cell sorting (FACS) of the single cell suspension obtained from dissociated E15.5 lung tissues (Fig.5B), the vascular SMCs (YFP⁺/DsRed⁺) were separated from non-vascular SMCs (YFP⁺/DsRed^-). These two distinct SMC populations were then cultured for further analysis ¹⁹. As shown in Fig. 5C, BMP4 treatment could promote SMC-associated gene expression in lung SMCs of non-vascular origin rather than those of vascular origin, as assessed by their Myh11 expression. This BMP4-induced upregulation of Myh11 expression in non-vascular SMCs was effectively blocked by concurrent treatment with the BMP type 1 receptor-specific inhibitor LDN.

Previous studies using embryonic lung explant culture have suggested that airway SMCs play an important role in branching morphogenesis in developmental lungs ^20,21. Myocd encodes an important transcriptional coactivator of serum response factor (SRF), modulating the expression of smooth muscle-specific cytoskeletal and contractile proteins ²². In our Bmpr1a CKO lungs, expression of Myocd was substantially downregulated (Fig.6A). To determine whether Bmpr1a-mediated downregulation of Myocd expression and the resultant airway SMC defects directly caused fetal airway cystic lesions in vivo, Myocd was deleted in lung mesenchyme starting at E6.5 using the same Tbx4-rtTA/TetO-Cre driver (Suppl. Fig.2). As detected by Acta2 immunostaining, knockout of Myocd led to a severe deficiency in airway SMC development (Fig.6B) resembling the airway SMC deficiency observed in the Bmpr1a CKO lungs (Fig. 2 and 3C). However, no significant branching defects or cystic malformation were observed in the Myocd CKO lungs (Fig.6C-6D), which is consistent with the previous report ¹⁸. Despite severe disruption of airway SMC development in the Myocd CKO lungs, elastin fiber deposition in the airways was unaffected (Fig. 6B), which was a significant difference when compared to the findings in the Bmpr1a CKO airways, where a deficiency of elastin fibers adjacent to the epithelium was observed (Fig. 3C).

The coordinated airway epithelial differentiation that forms the proximal-distal axis is another important mechanism controlling lung morphogenesis ²³. The proximal epithelial cells, marked by the expression of several unique genes such as Sox2, serve as progenitors for neuroendocrine cells, ciliated epithelial cells, and Club cells in the bronchi/bronchioles, while the distal epithelial cells, defined by expression of Sox9, Id2, and Sftpc in fetal lungs, are the progenitors of peripheral alveolar epithelial cells ²⁴. The mechanisms by which mesenchymal cells regulate airway epithelial cell fate during lung development are largely unknown. Interestingly, in the Bmpr1a CKO lung, Sox2-positive epithelial cells were barely detected in the proximal airways. Similarly, lack of further differentiation of Foxj1-positive cells was also observed. In contrast, cells expressing Sox9 and Sftpc, which are normally confined to the distal epithelial cells during lung branching, were expanded to the proximal region of the airways in Bmpr1a CKO lungs. The analysis of tissue RNA-Seq data revealed that the mesenchymal Bmpr1a knockout downregulated multiple proximal airway epithelial marker genes, while various distal airway epithelial marker genes were upregulated (Fig. 7B). Ectopic expression of Bmp4 in fetal mouse lung has been reported to cause distal epithelial expansion ¹⁵. Increased expression of Bmp4 in E15.5 Bmpr1a CKO lungs was initially found by RNA-seq (LogFC>1, GEO#: GSE97946, access token: kfofmkaqnpeflix) and validated by RT-PCR (Fig.7C). The inverse relationship between Bmpr1a deletion (intracellular BMP signaling) and Bmp4 ligand expression in these fetal lung mesenchymal progenitor cells (MPCs) was further confirmed by comparing Bmp4 expression in isolated fetal lung mesenchymal cells with different Bmpr1a genotypes (Fig. 7D-7E). Alterations of other key pathways including Fgf, Wnt, and Shh were not detected by RNA-Seq and RT-PCR through analyzing the related targeted genes.

The Tbx4-rtTA/TetO-Cre transgenic line was generated in our lab ¹⁶. Floxed-Bmpr1a (Bmpr1a ^fx/fx) mice were provided by Dr. Yuji Mishina ⁴⁷. Floxed-Myocd (Myocd ^fx/fx) mice were provided by Dr. Michael S. Parmacek ^48,49. Floxed-Smad1 (Smad1^fx/fx) mice were originally generated by Dr. Anita Roberts ⁵⁰. Floxed-Smad5 (Smad5 ^fx/fx) mice were provided by An Zwijsen at KU Leuven ⁵¹. Cspg4-DsRed mice were purchased from The Jackson Laboratory (#008241). Tagln-YFP reporter mice were provided by Dr. Jeffrey Whitsett at Cincinnati Children’s Hospital. To generate the lung mesenchyme-specific conditional knockout (CKO) mice of genes studied in this work, Bmpr1a^fx/fx, Myocd ^fx/fx, and Smad1^fx/fx/Samd5^fx/fxmice were bred to the Tbx4-rtTA/TetO-Cre mice carrying Bmpr1a^fx/+, Myocd ^fx/+ or Smad1^fx/+/Samd5^fx/+ respectively. Doxycycline (625 mg/kg in food (TestDiet) and 0.5 mg/ml in drinking water (Sigma)) was given to the mothers from E6.5 to induce Cre-mediated floxed gene deletion. Tagln-YFP mice were crossed with Cspg4-DsRed mice to generate double fluorescence reporter mice. Timed mating was set up as previously described ¹². Mouse fetal lungs of different embryonic stages were collected from pregnant females. All mice were bred in C57BL/6 strain background. The mice used in this study were housed in pathogen-free facilities. All mouse experiments were conducted in accordance with NIH Guide and approved by the Institutional Animal Care and Use Committee of Children’s Hospital Los Angeles.

Feal lung mesenchymal cells from E15.5 lungs were isolated and cultured as previously described⁵². To study the role of BMP signaling in subpopulations of SMCs, lung mesenchymal cells from Tagln-YFP/Cspg4-DsRed reporter mice were further sorted into perivascular SMCs and non-perivascular SMCs based on their fluorescent reporter expression using FACS (BD, FACSAria I). The cells were then plated at low density (∼10⁴/100-mm dish) and cultured in αMEM medium supplemented with 20% FBS, 2 mM L-glutamine, 55 μM 2-mercaptoethanol and antibiotics (100 U/ml penicillin and 100 μg/ml streptomycin). The cells (<10 passages) were treated with recombinant mouse BMP4 protein (50 ng/ml), LDN193189 (200 nM) or p38 pathway inhibitor (SB203580, 1 µM) for two weeks. BMP4 protein was purchased from R&D Systems (5020-BP). LDN193189 was provided by Dr. Paul Yu at Harvard Medical School ⁵³, and SB203580 was purchased from Sigma (#S8307).

For preparation of regular paraffin tissue sections and 2-D histological examination, isolated fetal lungs were fixed in 4% paraformaldehyde, as described in our previous publication ¹². Hematoxylin and eosin (H&E) and immunofluorescent staining was performed as described previously ⁵. For whole mount immunofluorescence staining and 3-D imaging, fetal lungs were isolated and fixed in DMSO:methanol (1:4) overnight at 4 °C, following a protocol described before ⁸. Antibodies used for immunofluorescent staining were listed in the Suppl. Table 1.

EdU incorporation was used to assess the in vivo cell proliferation following instruction from the manufacture (#C10337, Thermo Fisher Scientific). Cell quantification was performed using Fiji imaging software (1.52p) as described previously ⁵. Apoptosis was evaluated by TUNEL assay (#S7110, MilliporeSigma). E15.5 lung specimens with additional treatment of DNase I served as the positive control.

E15.5 mouse lungs were dissected in RNase-free PBS and flash-frozen in liquid nitrogen. After genotyping, mRNA from WT and Bmpr1a CKO lungs was prepared by using the TRIzol RNA extraction reagent (ThermoFisher, #15596026). The RNeasy Micro Kit (Qiagen, #74004) was then used to clean up the extracted mRNA. Total RNA quality was measured with a bioanalyzer (Agilent) with RIN greater than 7.0 being acceptable. RNA-sequencing libraries were generated and sequenced by Quickbiology (Quickbiology Inc. Pasadena, CA), using the Illumina TruSeq v2 kit (Illumina, San Diego CA). Data processing was performed using the USC high performance-computing cluster (https://hpcc.usc.edu/). Roughly 50 million 100bp single-end sequences were aligned to the Gencode M8 annotation ⁵⁴ based on the Genome Reference Consortium mouse genome (GRCm38.p4) and using the STAR aligner ⁵⁵. Read counts per gene were calculated using the HTSeq-count software ⁵⁶. Differential gene expression was determined using the R/Bioconductor software ‘edgeR’ ⁵⁷. Genes were considered significantly different if their false discovery rate (FDR) corrected P-values were less than 0.05 and Log2-Fold change was greater than 1.0. The sequencing data was deposited to the NCBI GEO repository with accession number GSE97946 (access token: kfofmkaqnpeflix).

Lung mesenchymal cells and tissue specimens were harvested and flash-frozen in liquid nitrogen for the subsequent mRNA and protein analysis. Total RNA was isolated from the cultured cells and lung tissues using the RNeasy Kit (Qiagen, #74106). The cDNA synthesis and real-time PCR were performed as described in a previous publication ⁵. The related oligonucleotide primers are listed in Suppl. Table 2. Protein lysate was prepared from the cells and lung tissues (n>3 per group) and analyzed by western blot as previously described ⁵². The related antibodies used for western blot are listed in Suppl. Table 1.

The quantitative data are presented as means ± SD. All in vitro experiments were repeated at least three times, and data represent consistent results. At least 5 mice per experimental group were utilized and the statistical difference between any two groups was assessed by an independent-sample t-test. One-way analysis of variance (ANOVA) was used to examine differences between groups of three or more, followed by post-hoc comparisons to study differences among individual group. Chi-square test with Yates correction was used for GO enrichment analysis. P< 0.05 was considered statistically significant.