Plasticity of button-like junctions in the endothelium of airway lymphatics in development and inflammation

Abstract

Endothelial cells of initial lymphatics have discontinuous button-like junctions (buttons), unlike continuous zipper-like junctions (zippers) of collecting lymphatics and blood vessels. Buttons are thought to act as primary valves for fluid and cell entry into lymphatics. To learn when and how buttons form during development and whether they change in disease, we examined the appearance of buttons in mouse embryos and their plasticity in sustained inflammation. We found that endothelial cells of lymph sacs at embryonic day (E)12.5 and tracheal lymphatics at E16.5 were joined by zippers, not buttons. However, zippers in initial lymphatics decreased rapidly just before birth, as buttons appeared. The proportion of buttons increased from only 6% at E17.5 and 12% at E18.5 to 35% at birth, 50% at postnatal day (P)7, 90% at P28, and 100% at P70. In inflammation, zippers replaced buttons in airway lymphatics at 14 and 28 days after Mycoplasma pulmonis infection of the respiratory tract. The change in lymphatic junctions was reversed by dexamethasone but not by inhibition of vascular endothelial growth factor receptor-3 signaling by antibody mF4-31C1. Dexamethasone also promoted button formation during early postnatal development through a direct effect involving glucocorticoid receptor phosphorylation in lymphatic endothelial cells. These findings demonstrate the plasticity of intercellular junctions in lymphatics during development and inflammation and show that button formation can be promoted by glucocorticoid receptor signaling in lymphatic endothelial cells.

Figure 1

Zipper-like junctions in primitive lymphatic endothelium at E12.5. A: Low-magnification view of Prox1-immunoreactive endothelial cell nuclei (green) in a cross section of embryo at E12.5 showing the anterior cardinal vein (CV), jugular lymph sac (JLS), and tissue near the JLS. Boxed regions in A are enlarged in B–D. B–D: Endothelial cells joined by zippers (arrows) in lymphatic structures stained for Prox1 (green) and podoplanin (blue) are shown next to the cardinal vein (B), within the JLS (C), and in adjacent tissue (D). E and F: Zippers (arrows) in a cluster of Prox1-positive endothelial cells. G: Zippers (arrow) in the endothelium of a blood vessel (Prox1 negative) at E12.5. H–J: Whole mount of trachea and lungs at E12.5 showing zippers in the endothelium (VE-cadherin, red) of blood vessels and lymphatics and nuclei of lymphatic endothelial cells (Prox1, green). Arrows, main stem bronchi. The boxed region in H is enlarged in I to show the primitive lymphatic plexus in the trachea. The boxed region in I is enlarged in J, as well as a small region outside the field of view, to show zippers in the endothelium of lymphatics at E12.5 (J: arrows). Scale bars: 200 μm (A); 50 μm (B, D, and I); 20 μm (C, E–G, and J); 400 μm (H).

Figure 2

Development of button-like junctions in the endothelium of lymphatics. A: Lymphatic vascular plexus (LYVE-1, green) in a tracheal whole mount at E16.5, showing tips of initial lymphatics (long arrows). Inset: Zipper-like junctions (VE-cadherin, red) in lymphatic endothelium (short arrows). Most of the lymphatics are located between cartilage rings (asterisks). B: Time course of development of button-like junctions in lymphatics of trachea (red) and diaphragm (blue) from E16.5 to P70. Buttons are expressed as number of adult-like VE-cadherin–stained segments. Buttons that appear before birth represent only approximately 35% and 20% of the eventual number present in adult (P70) initial lymphatics of trachea and diaphragm, respectively. *P < 0.05, values at P0 (arrow) are significantly different from those at E18.5. C–L: Lymphatic endothelial cell junctions are shown as inverted gray scale images (C–G: VE-cadherin) and in color (H–L: VE-cadherin, red; LYVE-1, green) to illustrate the sequence of changes in zipper-to-button transformation from E16.5 to P70 (diaphragm). Zippers are also shown in a blood vessel at E16.5 (C and H: long arrow). At P70, a distinctive structure at the junction of three lymphatic endothelial cells with button-like junctions is indicated by circles (G and L). M: Changes of tracheal lymphatic junction phenotype from E16.5 to P70, including zippers (green), intermediate (light blue), and buttons (red). Scale bars: 100 μm (A); 10 μm (C–L).

Figure 3

Colocalization of β-catenin and p120-catenin with VE-cadherin at lymphatic endothelial cell junctions at E17.5 and P0. Confocal images of tracheal lymphatics showing colocalization of β-catenin (A–D) or p120-catenin (E–H) (red) and VE-cadherin (green) at zippers at E17.5 (A, B, E, and F) and at primitive buttons at P0 (C, D, G, and H). Junctions in tracheal lymphatics had clear VE-cadherin immunoreactivity at E17.5 and P0, albeit weaker than in blood vessels. Yellow staining in B, D, F, and H reflects the colocalization of VE-cadherin (green) and β-catenin or p120-catenin (red) at junctions in lymphatic vessels. Scale bar = 10 μm.

Figure 4

Transformation of buttons to zippers in lymphatics after M. pulmonis infection. Confocal images of lymphatics in tracheal whole mounts of pathogen-free or M. pulmonis–infected mice. A–C: Different distributions of lymphatics (LYVE-1, green) and blood vessels (VE-cadherin, red) in pathogen-free trachea. Regions of mucosa over cartilage rings (asterisks) are almost free of lymphatics (A). VE-cadherin (red) and LYVE-1 (green) have complementary distributions at buttons in initial lymphatics (B), but LYVE-1 is weak or absent in collecting lymphatics (C). Some extracellular leukocytes also have LYVE-1 immunoreactivity (C: short arrows). D–I: Zippers in lymphatic endothelium shown after M. pulmonis infection (14 days) by staining for VE-cadherin (red) and LYVE-1 (D, E, G, and I) or claudin-5 (F and H) (green). Network of new or remodeled lymphatics above a cartilage ring (D). The boxed regions in D are enlarged in E–H to show zippers. E–I: Zippers (long arrows) in tips of lymphatic sprouts (E and F), stalks of lymphatic sprouts (G and H), and other regions of remodeled initial lymphatics (I). Claudin-5 is weak in blood vessels (E and F: short arrow). Scale bars: 200 μm (A); 50 μm (D); 20 μm (B, C, and E–I).

Figure 5

Reversal of button-to-zipper conversion in inflamed airways. A and B: Lymphatic area density in mucosa overlying cartilage rings (A) and normalized bronchial lymph node and lung weights (B). *P < 0.05 versus pathogen-free mice; ^†P < 0.05 versus the 14-day infected baseline group. Dex, dexamethasone. C–E: Zippers in lymphatic endothelium shown after M. pulmonis infection and vehicle treatment by staining for VE-cadherin (red) and LYVE-1 (green). Enlargement of boxed regions in C shows zippers (arrows) in both new (D) and existing (E) lymphatics. F–H: Buttons in nonregressed lymphatic endothelium shown after M. pulmonis infection and Dex treatment. Enlargement of boxed regions in F shows buttons (arrows) in the oak leaf–shaped endothelial cells of new (G) and existing (H) lymphatics. I–K: Zippers (arrows) in lymphatic endothelium shown after M. pulmonis infection and anti-VEGFR-3 antibody treatment. Enlargement of boxed regions in I shows zippers (arrows) in both new (J) and existing (K) lymphatics. Scale bars: 50 μm (C, F, and I); 20 μm (D, E, G, H, J, and K).

Figure 6

Promotion of button formation in neonatal lymphatics by dexamethasone. A and B: Lymphatics (LYVE-1, green) in tracheal whole mounts of untreated (A) or dexamethasone-treated (B) P4 pups. Boxed regions in A and B are enlarged in C and D, and E, respectively. C and D: VE-cadherin (red)–stained zippers in tips of lymphatic sprouts (C: arrows) and primitive buttons in other portions (D: arrows) of initial lymphatics from untreated trachea. E: After dexamethasone, oak leaf–shaped endothelial cells have buttons (arrows) with a complementary distribution with LYVE-1 (green) at cell borders. F and G: Zippers (arrows) in collecting lymphatics containing intraluminal valves from untreated (F) and dexamethasone-treated (G) mice. H and I: Jagged zipper-like junctions (arrows) in thoracic ducts of untreated (H) and dexamethasone-treated (I) mice. Scale bars: 100 μm (A and B); 50 μm (F and G); 20 μm (C–E and H–I).

Figure 7

GR activation in lymphatics by dexamethasone at P4. A–F: Confocal images of lymphatics in tracheal whole mounts of untreated P4 pups (A, B, and E) or littermates treated with dexamethasone from P0 to P4 (C, D, and F). Endothelial cell junctions stained for VE-cadherin (green), nuclear phospho-GR Ser211 (red), and nuclear Prox1 (blue). A and B: Nuclear phospho-GR Ser211 is expressed in almost every cell in untreated trachea but is weak in lymphatic endothelial cells marked by Prox1 (red, arrows). C and D: After dexamethasone, nuclear phospho-GR Ser211 staining in lymphatic endothelial cells is stronger and more widespread, as shown by colocalization with Prox1 in nuclei (D: purple, arrows). E and F: Negative control (primary antibody for phospho-GR Ser211 was omitted). Scale bar = 20 μm.