Focal adhesions are essential to drive zebrafish heart valve morphogenesis

Abstract

Elucidating the morphogenetic events that shape vertebrate heart valves, complex structures that prevent retrograde blood flow, is critical to understanding valvular development and aberrations. Here, we used the zebrafish atrioventricular (AV) valve to investigate these events in real time and at single-cell resolution. We report the initial events of collective migration of AV endocardial cells (ECs) into the extracellular matrix (ECM), and their subsequent rearrangements to form the leaflets. We functionally characterize integrin-based focal adhesions (FAs), critical mediators of cell-ECM interactions, during valve morphogenesis. Using transgenes to block FA signaling specifically in AV ECs as well as loss-of-function approaches, we show that FA signaling mediated by Integrin α5β1 and Talin1 promotes AV EC migration and overall shaping of the valve leaflets. Altogether, our investigation reveals the critical processes driving cardiac valve morphogenesis in vivo and establishes the zebrafish AV valve as a vertebrate model to study FA-regulated tissue morphogenesis.

Figure 1.

Morphogenetic processes driving development of the superior and inferior leaflets of the AV valve. (A–F) AV ECs marked by Tg(nfatc1:GAL4); Tg(UAS:EGFP-CAAX) expression. (A′–F′) Superior AV ECs. (A″–F″) Inferior AV ECs. (A) At 48 hpf, single-layered ECs were found in the superior and inferior parts of the AVC. (B) At 55 hpf, leading ECs form protrusions in the superior AVC (B′), whereas inferior AV ECs in the majority of embryos examined (25/44) remained single layered (B″). (C) At 60 hpf, collective migration of superior AV ECs into the ECM in a ventricle-to-atrium direction (C′). Inferior AV ECs extended protrusions into the ECM but remain single layered (C″). (D) At 75 hpf, establishment of a folded multilayered structure in the superior part of the AVC (D′). Variability was observed in inferior leaflet formation, as inferior AV ECs appeared as a folded tissue in most larvae (D″; 18/27) and remained single layered in the others (9/27). (E and F) At 100 and 120 hpf, there was elongation of AV EC layers into the vascular lumen in both the superior (E′–F′) and inferior (E″–F″) parts of the AVC. Overviews are single planes (A–F); magnified images of AV ECs are 5-µm-thick maximum projections of five confocal planes (A′–F″). A, atrium; V, ventricle. In models A–F: blue, AV ECs; red, nonvalve ECs; brown, myocardium; yellow, vascular lumen; purple, ECM. Scale bars: (A–F) 20 µm; (A′–F″) 10 µm.

Figure 2.

Localization of factors involved in cytoskeletal organization in migrating AV ECs. (A–H) Immunohistochemistry of 58-hpf Tg(nfatc1:GAL4); Tg(UAS:EGFP-CAAX) (A–F) or Tg(nfatc1:GAL4); Tg(UAS:EGFP) (G and H) embryos. Green, GFP; cyan, DAPI; magenta (A–H) and white (A′–H′), specific factors of interest. (A–F) Localization of FA-associated factors. (A and B) Itgβ1 (A) and Itgα5 (B) appeared enriched in discrete puncta at the migrating cell protrusions. (C–E) Cytoplasmic FA adaptor molecules Talin1 (C), Vinculin (D), and p-Pax (E) appeared strongly localized in the leading edge of migrating cells. (F) The Itgα5β1 ligand Fn appeared distributed around the endocardial and myocardial cells at the AVC. (G and H) The tight junction–associated protein ZO-1 and adherens junction protein VE-cadherin were localized between the leading and follower AV ECs but appeared undetectable in the EC protrusions. Dashed red lines outline the leading edge of the migrating cells at a distance of 1–2 µm. Scale bars: (A–F) 20 µm; (A′–F′ and G–H′) 10 µm.

Figure 3.

Blocking FA signaling through DN Vinculin overexpression severely affects AV EC migration and valve tissue establishment. (A) Schematics of the UAS-driven DN Vinculin transgene, expressing the head domain of Vinculin fused to GFP, Tg(UAS:DN-Vcl-EGFP). (A–C′) 60 hpf: AV EC migration stage. (B) In WT embryos, AV ECs extended protrusions and formed a multilayered tissue. (C) In embryos overexpressing DN Vinculin, AV ECs remained single layered and extended microvilli-like protrusions. (D) Categorization of migration stages of superior AV ECs in WT (n = 12) and DN Vinculin–overexpressing (n = 21) embryos. (E) Number of protruding ECs in superior and inferior AVC was significantly reduced when DN Vinculin was overexpressed (averages of 2.25 and 1.32 cells per embryo in WT and DN Vinculin, respectively; P = 3.1^e−3, unpaired Student’s t test). WT, n = 47 embryos; DN Vinculin, n = 34 embryos. Error bars represent standard deviation; asterisks indicate statistical significance. (F–G′) 72 hpf: establishment of prevalvular structure. (F) WT AV ECs completed migration and established a folded multilayered prevalvular structure. (G) DN Vinculin–overexpressing AV ECs appeared disorganized and did not clearly establish a multilayered tissue. (H–I′) 96 hpf: establishment of functional AV valve leaflets. (H) In WT larvae, both superior and inferior leaflets formed and extensively elongated into the vascular lumen. (I) In DN Vinculin–overexpressing larvae, AV endocardial tissue remained single layered, without established leaflet structures. Scale bars: (B and C, and F–I) 20 µm; (B′ and C′ and E′–H′) 10 µm.

Figure 4.

Real-time imaging shows WT AV ECs migrating and establishing a multilayered tissue, whereas DN Vinculin–overexpressing ECs failed to extend protrusions. (A–L) Maximum projections of the entire heart. (A–F) At the AVC, WT ECs originated as a single layer of cells (A), extended long protrusions, moved into the overlying ECM (B–D), and established a multilayered tissue (E and F; no more protrusions were observed at these stages). n = 4 animals. (G–L) DN Vinculin–overexpressing ECs did not extend prominent protrusions and remained single layered until 60 hpf (G–I). By 72 hpf, they appeared disorganized and did not adopt the folded tissue arrangement, and the AVC appeared collapsed (J–L). n = 2 animals. Scale bars: 20 µm.

Figure 5.

Overexpression of DN Vinculin leads to significantly fewer nfatc1:GAL4-expressing ECs in the AVC. (A–F) 3D images of heart valves in WT (A–C) and DN Vinculin–overexpressing (D–F) animals. Shown are sagittal views (left) and atrium-to-ventricle views (right). (G) Number of AV ECs, marked by nfatc1:GAL4; UAS:EGFP and kdrl:nls-mCherry expression, in WT and DN Vinculin–overexpressing animals imaged over a 48-h time window (P = 4.48^e−5 at 77 hpf, P = 1.6^e−5 at 105 hpf). Error bars represent standard deviation; asterisks indicate statistical significance. (H) Increase in the number of AV ECs between 56 and 77 hpf and 77 and 105 hpf in WT and DN Vinculin–overexpressing animals. WT, n = 12; DN Vinculin, n = 10; P = 2.24^e−5. An unpaired Student’s t test was used for G and H. (I and J) 3D images of 100-hpf heart valves. Scale bars: 10 µm.

Figure 6.

DN Vinculin overexpression impairs valve function. (A–D) Still images of beating hearts from WT (A and C) and DN Vinculin–overexpressing (B and D) larvae. Images were taken from videos that comprised one ventricular contraction (a duration of ∼40 ms). Larvae at 77 (A and B) and 100 (C and D) hpf are shown. WT valve leaflets closed the AVC for ∼30 ms per contraction, whereas AVC leaflets in DN Vinculin–overexpressing larvae did not close efficiently and did not prevent retrograde blood flow. Scale bars: 20 µm.

Figure 7.

DN Itgβ1 overexpression affects the protrusive activity of AV ECs. (A) Domain organization of the control Tg(UAS:mCherry) and the DN Itgβ1 Tg(UAS:mCherry-itgb1^DN transgenes. SP, signal peptide; TM, transmembrane; ICD, intracellular domain. (B–G) Representative images and schematics of the control (B–D) and DN Itgβ1–overexpressing (E–G) AV ECs at 56 and 75 hpf. (H and I) Proportions of cells expressing mCherry-tagged control or DN Itgβ1 in total AV ECs (H; control, n = 242 cells/12 embryos; DN Itgβ1, n = 361 cells/26 embryos) or leader AV ECs (I; control, n = 31 cells/12 embryos; DN Itgβ1, n = 36 cells/26 embryos). Whereas 100% of leader ECs expressed the control transgene, only 24% expressed DN Itgβ1 (P = 2.03^e−9). (J) Number of leader ECs in control (n = 12 embryos) and DN Itgβ1–overexpressing (n = 26 embryos) animals. The number of leader ECs in the superior and inferior AVC was significantly reduced when DN Itgβ1 was overexpressed (averages of 2.58 and 1.38 cells per embryo in control and DN Itgβ1, respectively; P = 4.5^e−4). An unpaired Student’s t test was used for I and J. Error bars represent standard deviation; asterisks indicate statistical significance. Scale bars: 20 µm.

Figure 8.

Loss of Itgα5 leads to delays in AV EC migration. (A–G) Phalloidin staining of itga5 WT, heterozygous, and mutant hearts at 58 (A–D) and 80 (E–H) hpf (magenta in overviews [left]; white in magnified AV areas [right]); green, ZO-1; cyan, DAPI. Representative images of hearts and superior AVCs of itga5 WT (A and E), heterozygous (B and F), and mutant (C and G) animals at 58 (A–C) and 80 (E–G) hpf. Categorization of migration stages at 58 hpf (D: +/+, n = 9; +/−, n = 13; −/−, n = 10) and formation of valvular structures at 80 hpf (H: +/+, n = 4; +/−, n = 17; −/−, n = 11). itga5 mutants exhibited delayed AV EC migration at 58 hpf and delayed establishment of valve structure at 80 hpf. Scale bars: (A–C and E–G, left) 20 µm; (right) 10 µm.