Convective tissue movements play a major role in avian endocardial morphogenesis

Abstract

Endocardial cells play a critical role in cardiac development and function, forming the innermost layer of the early (tubular) heart, separated from the myocardium by extracellular matrix (ECM). However, knowledge is limited regarding the interactions of cardiac progenitors and surrounding ECM during dramatic tissue rearrangements and concomitant cellular repositioning events that underlie endocardial morphogenesis. By analyzing the movements of immunolabeled ECM components (fibronectin, fibrillin-2) and TIE1 positive endocardial progenitors in time-lapse recordings of quail embryonic development, we demonstrate that the transformation of the primary heart field within the anterior lateral plate mesoderm (LPM) into a tubular heart involves the precise co-movement of primordial endocardial cells with the surrounding ECM. Thus, the ECM of the tubular heart contains filaments that were associated with the anterior LPM at earlier developmental stages. Moreover, endocardial cells exhibit surprisingly little directed active motility, that is, sustained directed movements relative to the surrounding ECM microenvironment. These findings point to the importance of large-scale tissue movements that convect cells to the appropriate positions during cardiac organogenesis.

Figure 1

Endocardial tube assembly takes place in a microenvironment abundant in fibronectin and fibrillin-2 ECM. A–D: maximum intensity dorso-ventral projections of confocal stacks obtained at HH8- (A), HH8 (B), HH10 (C), HH11 (D) stages of development. Whole mount specimens were immunostained with monoclonal antibodies QH1 (red), B3D6 anti-fibronectin (green), and JB3 anti-fibrillin-2 (blue). A′ – D′: single, 2.5–4 μm thick optical plane images selected from the regions marked with rectangles in panels A–D. In addition to the multicolor merged image, distinct panels show each immunofluorescence label inverted for better contrast. Scale bars: A–D — 100 μm, A′, C′ — 40 μm, B′, D′ — 20 μm. Fibronectin fibrils demonstrate an increasing degree of association with endocardial cell surfaces at progressively later developmental stages (red arrows). Fibronectin is also associated with the myocardium at HH11 (blue arrow). Fibrillin-2 fibrils are somewhat larger in diameter and appear to be aligned with the QH1-positive cell surfaces (A–C, A′–C′, red arrows). Both types of ECM surround the endocardial layer in the assembled heart tube (D, D′).

Figure 2

Fibronectin (red) and fibrillin-2 (green) fibrils derived from the LPM at HH8- translocate to the heart by HH10. A: Corresponding epifluorescence (left) and DIC (right) image pairs, representing development between HH8- and HH10. ECM movements are indicated by the superimposed trajectories. Scale bars — 100 μm. B: Schematic fate map of ECM fibrils in a HH8- stage embryo. ECM from the colored areas incorporates into the cranial (red) or middle (yellow) portions of the heart tube, or to the caudal heart tube and the omphalomesenteric veins (green) by HH10. Blue dots indicate ECM fibrils that do not incorporate into the heart. The fate map shown is a representative example out of 3 specimens that were imaged and analyzed by identical methods. See Movies S1 and S2.

Figure 3

ECM positioned ventral to the pharyngeal endoderm is translocated towards the anterior portion of the tubular heart and contributes to its elongation after HH10. A: Schematic transverse section, adopted from Lillie, 1930. B: Corresponding fibrillin-2 immunofluorescence (left) and DIC (right) image pairs are shown. During the elapsed 4 hours, the length of the cardiac tube substantially increases (DIC panels). Superimposed trajectories of individual fibrils are color-coded: red, green and blue colors indicate progressively later trajectory segments. The movements of fibrillin-2 fibrils are indicated by the red arrows in panel A. Scale bar-100 μm.

Figure 4

Exogenous fibronectin, injected into the LPM, incorporates into the cardiac tube. Human fibronectin conjugated to Alexa555 (red) was injected into the anterior LPM at HH7, while fibronectin-Alexa647 (green) was delivered to the paraxial mesoderm and medial-most regions of the LPM. A: Corresponding epifluorescence (left) and DIC (right) image pairs, showing an embryo at HH8− (top), HH8+, HH10, and HH11 (bottom) stages of development. Some fibronectin fibrils derived from the lateral (red) pool are progressively displaced towards the forming heart tube, which begins to exhibit Alexa555 fluorescence at HH10 and 11 (white arrowheads). At the same time, a large portion of the laterally-injected fibronectin pool remains associated with mesodermal and endodermal tissues in the anterior embryo. The medially-delivered (green) fibronectin pool does not contribute to the cardiac jelly. Scale bars — 100 μm. B–D: Transverse cryosections through the heart tube of the same embryo shown in panel A, fixed at HH 11. Approximate planes of section are indicated in the bottom right frame of panel A. Red fluorescence signal corresponds to the exogenous fibronectin-Alexa555. The green signal shows the distribution of endogenous quail fibronectin, labeled with B3D6 antibody post-sectioning. DAPI staining of cell nuclei is shown in blue. Human fibronectin is present at the endocardial-cardiac jelly interface (asterisks). nc: notochord, e: endocardium, m: myocardium. Scale bars — 40 μm.

Figure 5

Large-scale displacements of fibronectin and fibrillin-2 structures are identical. A, B: Fibronectin and fibrillin-2 immunostaining (inverted for better contrast) and their superimposed images (third column) are shown together with the corresponding PIV velocity vector map (fourth column). The embryos shown represent an early (ES1, panel A) and a later (ES7, panel B) stage of development. The velocity fields show fibronectin (red vectors) and fibrillin-2 (green vectors) movements in a selected frontal optical plane intersecting the forming heart. Vectors represent velocities as extrapolated displacements during one hour. The reliability of the vectors is indicated by color brightness. Overlap between the two colors appears as yellow. Dots indicate the absence of movement, i.e., velocities smaller than 10 μm/h. Scale bars — 100 μm. C: Mean speed of centripetal ECM movements along the AIP and the inflow region, measured within locations similar to the indicated circles in panels A and B. For each endocardial stage of development, the average speed of ECM displacements is shown for both fibronectin (red) and fibrillin-2 (green). The average magnitude of the difference between the two is presented in yellow. Error bars indicate SEM; distinct specimens were considered statistically independent.

Figure 6

TIE1+ endocardial progenitors are spatially segregated from the endothelial population and participate in coordinated midline-directed displacement along with the ECM in the adjacent tissue. A: Epifluorescent images of a Tie1::H2B-YFP transgenic embryo at early (HH8, left) and later (HH9, right) stages of development, with superimposed trajectories of endothelial and endocardial progenitors. Cell trajectories were obtained using the CIPA automated tracking software; the color-code conveys timing information as in Fig. 3. Scale bars — 100 μm. B: Trajectories of manually-traced Tie1::H2B-YFP pre-endocardial cells (green) are compared with those of fibrillin-2 filaments (red). C: Trajectories representing the total (green) and active (red) movements of tracked pre-endocardial cells. Trajectories shown in panel C are re-plotted in such a way that each cell starts from the origin (green trajectories). Active cell movements (red trajectories) were estimated by subtracting the local ECM movements at each frame from the total displacements. D: Mean displacement vs. elapsed time for the total (green symbols) and active (red symbols) cell movements. On a double logarithmic plot a line with slope 1 (green line) indicates persistent motion (displacement and elapsed time are proportional), while a line with slope 1/2 (red line) indicates random walk (displacement is proportional to sqrt(t)). Error bars represent standard deviation, obtained from a pool of all tracked cells.

Figure 7

Active and total movements of Tie1::H2B-YFP pre-endocardial cells. A representative specimen is shown at an early (panels A and A′, ES 1) and a later (panels B and B′, ES 7) stage of heart development. Images in panels A and B show velocity vectors (red for ECM and green for cells) in a selected frontal optical plane intersecting the forming heart. Velocities are indicated as extrapolated displacements over a 1h long time period, the yellow square indicates a linear size of 100 μm. The reliability of the vectors is estimated by the contrast of the local fluorescence, and indicated as color brightness. The locally prevalent direction and speed of active cell movements was determined by averaging the vectors of active cell velocities within a radius of 100 μm. These vectors, projected to a frontal plane, are shown in A′ and B′, and thus represent the local directional bias of active cell movements. The color assigned to the vectors indicates the local standard deviation of active cell movements: blue and red colors indicate locally ordered (small standard deviation — vectors pointing in similar directions) and disordered movements, respectively. Notice the tendency to move towards the heart in the vicinity of the forming heart tube (asterisks). C: Speeds of active (bars 1 and 2) and total (bars 3 and 4) cell movements, the corresponding speed of ECM displacements (bar 6), and for comparison the calculated difference in the movements of the two ECM components (bar 5) are shown for the seven stages of endocardial development (See Supplemental Fig. 5). Cell displacements are estimated by two methods, PIV (bars 1 and 3) and automated cell tracking (bars 2 and 4), and yield similar results.

Figure 8

Tissue motion transfers passive objects to the site of cardiogenesis. A: Microinjected latex beads (green) within a HH8- embryo, also immunolabeled for fibrillin-2 (red). Arrows a–c indicate groups of beads that were incorporated in the forming heart tube at HH12 (see movie S9). B: Confocal micrograph of the same embryo at HH12, immunostained with QH1 antibody. A 24 μm thick confocal stack through the heart tube was obtained along the dorso-ventral axis. The image shown is a maximum intensity projection, beads are shown in green, fibrillin-2 in red and QH1 in blue. As determined by tracking the beads through the time-lapse image sequence, arrows a-c indicate identical groups of beads in both panels. In agreement to our results shown in Figs. 2 and 3, the group of beads indicated by (a) entered the heart tube from the caudal end, and those indicated by (b) and (c) were included in the forming heart from the cranial pole.