A role for mesenchyme dynamics in mouse lung branching morphogenesis

Abstract

Mammalian airways are highly ramified tree-like structures that develop by the repetitive branching of the lung epithelium into the surrounding mesenchyme through reciprocal interactions. Based on a morphometric analysis of the epithelial tree, it has been recently proposed that the complete branching scheme is specified early in each lineage by a programme using elementary patterning routines at specific sites and times in the developing lung. However, the coupled dynamics of both the epithelium and mesenchyme have been overlooked in this process. Using a qualitative and quantitative in vivo morphometric analysis of the E11.25 to E13.5 mouse whole right cranial lobe structure, we show that beyond the first generations, the branching stereotypy relaxes and both spatial and temporal variations are common. The branching pattern and branching rate are sensitive to the dynamic changes of the mesoderm shape that is in turn mainly dependent upon the volume and shape of the surrounding intrathoracic organs. Spatial and temporal variations of the tree architecture are related to local and subtle modifications of the mesoderm growth. Remarkably, buds never meet after suffering branching variations and continue to homogenously fill the opening spaces in the mesenchyme. Moreover despite inter-specimen variations, the growth of the epithelial tree and the mesenchyme remains highly correlated over time at the whole lobe level, implying a long-range regulation of the lung lobe morphogenesis. Together, these findings indicate that the lung epithelial tree is likely to adapt in real time to fill the available space in the mesenchyme, rather than being rigidly specified and predefined by a global programme. Our results strongly support the idea that a comprehensive understanding of lung branching mechanisms cannot be inferred from the branching pattern or behavior alone. Rather it needs to be elaborated upon with the reconsideration of mesenchyme-epithelium coupled growth and lung tissues mechanics.

Figure 1. Three-dimensional reconstruction of right cranial lobe full structure.

(A) Dorsal view of a whole mount mouse lungs at E13.5 immunostained for E-cadherin (red) and counterstained with DAPI (bleu), showing both the airway epithelium architecture and the shape of the surrounding mesenchyme. Dotted lines show the trachea (Tr), the right (Rmb) and left (Lmb) main bronchi, the right cranial (RCr), right middle (RMd), right accessory (RAc), right caudal (RCd) and left (L) lobes. (B) Ventral view of a whole mount mouse lung immunostained for E-cadherin at E11.5 to show the epithelial tree. Cr, Md, Ac, Cd and L bronchi give rise to the related lobes. The unwanted background around the buds is a limiting factor to perform quantitative analysis. The procedure we developed (see Material and methods) allowed a highly precise 3D visualization of the bronchial tree (C) and the surrounding mesenchymal cell mass (D), with respect to their in vivo relationships. A, anterior; P, posterior; M, medial; L, lateral; D, dorsal; V, ventral; Scale bar: 200 µm.

Figure 2. Lung lobe packing in the mouse fetus thorax.

Transversal sections of the thorax are performed at the embryonic day indicated and stained with HPS to show the in vivo relationships of the mouse lung. (A) During the pseudoglandular stage, the size and shape of the lung anatomical cavity is mainly constrained by dense tissues (chest wall, liver) or cavities under pressure (heart and larges vessels). RCr develops competing interface with RMd lobe (arrow heads) and slightly imprint the looser tissues of the chest wall (arrows). (B–F) Magnifications show that the RCr lobe surface fits the shape of the surrounding tissues, even if the parietal and visceral mesotheliums are not in strict abutment. The RCr lateral edge is tightly embedded between the chest wall and the heart/RMd. Ep: lung epithelium, Mes: lung mesenchyme, Pl: pleural cavity (outlined by the visceral and parietal mesothelium), Fl: luminal fluid; scale bar: 100 µm.

Figure 3. Local bud-induced deformations on the lobe surface.

(A) Transverse section of RCr lobe at E13.25 showing bumps (red arrows) and grooves (red arrow head) on the lobe surface, (B) The characteristic Fgf10 expression pattern at E13.25 in the sub-mesothelial mesenchyme outlines regular curves along the lobe edge. (C–E) Three-dimensional reconstructions of RCr lobe at E12.5 to show slight bumps in front of the enlarging buds. (E) Larger and transient grooves appear around group of branches (red arrow heads) whereas small bumps face bud tips or branching sites (red arrows). (F–H) As bud grows toward the lobe surface and reaches the sub-mesothelial area the nucleus density increases specifically in the mesenchyme located between the tip and the mesothelium. In the same time window, the bud tip progressively enlarges and the hump on the lobe surface appears. Scale bar 100 µm.

Figure 4. Mesenchyme dynamics effects on the underlying branching architecture.

Series of E11.25-E12 RCr 3-D reconstructions, (A) Upper panel: lateral views, lower panel: dorsal views, (B) Upper panel: ventral views, lower panel: lateral views. The RCr lobe quickly elongates along the anterior-posterior axis, first inducing planar bifurcations. The first side-branches (SB1 and SB2) sprout latter, as the mesenchyme thickness increases on the medial and dorsal sides respectively (A). The mesenchyme growth dynamics (white crosses and white dashed arrows) also induces absolute orientation changes of the previously formed branches (black and white dashed lines), rotation at the branching site (rounded black dashed arrow) and differential rate of bud growth (black dashed arrows) (B). Scale bar 100 µm.

Figure 5. Branching stereotypy and variability.

(A) RCr lobe at E12.75. Planar bifurcations fill the angles and occur rigidly in the angle bisector plane (doted line). The side-branches sprout in front of the flat faces (first the medial/dorsal face). They grows in parallel directions and forms rows of bristle (dashed lines), even if they sprout from different parental branches. The side-branches are formed in a proximal-to-distal order (full circles and arrows) from the large perihilar region to the thin edges. Doted circle and arrow depict the next budding sites (see also Figure S1), where mesenchyme is going to enlarge. (B) Branching variations mostly occur in the more open spaces (here the rounded medial face), where a classical rosette (a) co-exist with poorly stereotyped bunches of sprouts (b, c). (C) The sprouting orientation of SB2 (side-branch 2) is highly variable (white arrow). SB2 also originate from variable sites: up, from (dashed white line) or down the PL branching fork. An optional side branch (SB2*) sprouts proximal to SB2 (white star) and modify the bifurcation plane of OB1 and OB2. (D) The PL branch exhibits polymorphic patterns of planar bifurcations or trifurcation (white crosses). The latter originate from large belly also corresponding to an optional-side branch site (white circle). Subtle differences in the mesoderm growth are associated with variable branching rate (white and black arrow). Of interest, the branching pattern also can raise nomenclature confusions: an apparent side-branch (underscored black cross) is indeed generated through end bifurcation (black cross). (E–H) E13.25 RCr lobes showing several morphological branching variants: (E) 3-D trifurcations, (F) rosette directly sprouting from the parental branch, or at the same site, a missing proximal branch (black star) leading to tripod, (G) elbow in the vicinity of another lineage and (H) variable rotation planes twisting the classical rosette (left panel) and the orthogonal bifurcations (right panel). Scale bar 100 µm.

Figure 6. Overall growth coupling of the bronchial tree and the mesenchyme.

Graphs plotting the length against the width (A), the width against the thickness (B) and the length against the thickness (C) of a series of E11.25-E13.5 RCr lobes. Using the same specimens, the surface and the volume of the bronchial tree were plotted against the respective surface and volume of the mesenchyme cell mass (E and D). The main dimensions are strongly correlated showing that slight inter-specimen differences occur while the overall shape of the RCr lobe is conserved. In comparison, the overall growth of the epithelial tree and the mesenchyme compartment are very strongly correlated at the lobe level. Number of analyzed specimens: 20. Distances are denominated in µm, surfaces in µm² and volumes in µm³.

Figure 7. Space filling properties of the developing bronchial tree.

(A) Spheres (blue) are allowed to grow at the bud tips until they reach one another and/or the lobe surface. Given these conditions, spheres show only slight differences of diameters. The minimal sphere radius within a lobe is used as in vivo space filling criterion (Kiv). Arrows indicate intricate spheres at site undergoing bifurcation. The gaps between the spheres are predictive of the next branching sites: dorsal side-branches (doted circles) and posterior side-branch (star). (B) Kiv is given as percentage of the optimal value approximated by the best random value BRV (see Material and Methods). Although the best random spacing is geometrically overrated (see Material and Methods), all the Kiv values are higher than 80% of BRV at E11 (red), E12 (purple) and E13 (green). (C) From E11 to E13, Kiv (blue bars) are comprised between the 99th BRV percentile (light grey bars) and the BRV value (dark grey bars), except for one E13 lobe, demonstrating that new bud tips are not distributed at random and tend to fill homogeneously the available space in the mesenchyme. Scale bar 100 µm.