Workflow for extracting and defining the motion and deformation pattern of the myocardium during early heart morphogenesis.

The analysis pipeline consists of four main steps. Our dataset includes 16 specimens ranging from E7.75 to E8.25 (12 hours). A), Estimating Individual Time-Lapse Motion: A non-rigid algorithm extracts the transformation T needed to deform one frame, I(ti), into the next frame, I(ti+1). The transformation is computed for a set of control points of a regular 3D grid (blue grid). Control point displacement is updated (orange grid) until the two frames overlap. We collect the set of transformations {Ti} for each time-lapse image to compute the continuous description of HT morphogenesis. B), Integrating multiple time-lapse images into the Atlas: We incorporate the individual continuous descriptions of HT deformation into the Atlas. We define a staging system that aligns the individual time-lapse images on a common time reference and registers the related 3D mesh (Live-Shape) onto the spatial reference (Atlas). C), Quantifying Cardiac Spatial-Temporal Deformation Patterns: We extract the deformation patterns from each continuous description of HT deformation and compile them into the Atlas, generating a single stage-by-stage model (μdef, σdef). D), Creating an in-silico Fate Map: We concatenate a continuous description of HT deformation from the individual steps to recreate a unique and continuous description of early HT deformation patterns. Gray spots represent the initial positions of the pseudo-cells, while color lines, their displacement during development. Yellow spots indicate the tracked cells.

Deformation pattern analysis.

A), Mesh transformation from the initial state to the deformed state. The triangles of the mesh are transformed according to the MIRT tensor from time ti to time ti+1. The deformation pattern is calculated by applying the principles of continuum mechanics between the initial and deformed states. This pattern is then visualized on the deformed shape. The spatio-temporal registration onto the Atlas enables the mapping of the deformation into it. B), Description of the deformation parameters between the initial and deformed states. For each triangle, three different possible deformations (a, b, c) are represented, color-coded according to the parameter value. The growth rate (J) is defined as the ratio of the area of the same triangle before and after deformation. Anisotropy (θ) is defined as the ratio between the maximum and minimum deformation magnitudes. The maximum strain (ε) is characterized by the vector of maximum deformation. The agreement index reports the local coherence in the direction of tissue deformation. C), Difference between stepwise deformation pattern analysis (grey arrows) and cumulative deformation pattern analysis (orange arrows). D-E), Stepwise deformation pattern analysis in caudal view. The color maps represent the growth rate, anisotropy, strain magnitude and direction, and finally, their agreement. Here, the deformation maps are shown for the transition from stage 2 to stage 3 and from stage 7 to stage 8. The full arrowhead indicates a zone of low growth rate and high anisotropy in the most ventral-medial region. The empty arrowhead indicates discontinuity in the agreement of the strain direction. Deformations values are the mean values. The number of specimens averaged at stages 2–3 is 5, while at stages 7–8 it is 2.

Cumulative Deformation Pattern Analysis.

A), Concatenation of multiple time-lapses on a common timeline. Each time-lapse covers sub-windows of the Atlas temporal line. B), Schematic description of the concatenation pipeline. We fix a reference sample and for each equally staged sample, we select the centroids of the SurfaceMap closest to the reference SurfaceMap. C), Cumulative growth of HT. The color map refers to the mean growth rate of the tissue from stage 2. D), Cumulative anisotropy of HT. The color pattern indicates which zone of the myocardium accumulates more anisotropic deformation. The color map refers to the mean accumulated anisotropy of the tissue from stage 2. The color bars indicate the deformation magnitude. Grey zones correspond to the missing IFTs and arterial pole parts.

In-silico Fate Map.

A), Labelling of a CC discrete region by dye injection (t0, left) and 3D reconstruction of its contribution to the HT after 15 hours of embryo culture (t15h, right). A’), labelling on stage 2 virtual model of an equivalent region to that experimentally labelled in A) (left) and the resulting labelled region in the stage 6 virtual model (right). B), Labelling of a cardiac crescent discrete region (t0, left) by TAT-Cre-induced recombination and 3D reconstruction of its contribution to the HT after 20 hours of embryo culture (t20, right). B’), labelling on stage 2 virtual model of an equivalent region to that experimentally labelled in A) (left) and the resulting labelled region in stage 8 virtual model (right). C-F), different views of the virtual model showing different regions of the HT in stage 8 and their primordia in stage 2.

Fate maps of the cardiac crescent boundaries reveal tissue dynamics of heart tube formation.

A), Clusters of pseudo-cells were labelled at regular spacing along the D2 line of the stage 2 CC model. The labeled clusters were then followed through stage 8. Caudal views are shown, which allow visualize the caudal side of the developing inflows and ventricle. A’) Left, representation of the degree of expansion (green) or contraction (red) of the different D2 segments defined between the labelled clusters at stage 2. Numbers indicate fold change from the initial to the final time points. Right, representation of the stage 8 view indicating the regions of D2 that contract (red) versus those that expand (green). Fold-change cranial ward expansion of each labelled cluster from their original craniocaudal extension is shown in green. The direction of the white arrows indicates the main direction of expansion. The angle of the main expansion direction with respect to the embryo mid-plane is indicated in blue. B), Contiguous segments of pseudo-cells were labelled at regular spacing along the D1 line of the stage 2 CC model. The labeled clusters were then followed through stage 8. Dorsal views are shown. B’) Left, representation of the labelled segments on stage 2 CC with indication of their fold-change expansion (green) or contraction (red). Right, Representation of the final extension at stage 9 of the D1 labelled segments. Arrows indicate the direction and extension of the expansion of the labelled pseudo-cells to colonize the dorsal parts of the HT. C), Schematic representation of the cardiac crescent at stage 2 indicating the four anchor points used for experimental validation: two along D2 (yellow and magenta dots) and two along D1 (purple and green dots). D) Experimental validation of D2 contraction. An embryo (e6_D2) was microinjected at the D2 anchor points (yellow and magenta) and imaged by multiphoton microscopy at t0 (left) and after 10 hours (right). The Euclidean distance between the two labeled anchor points was measured at both time points (right image). E), Experimental validation of D1 contraction. An embryo (e1_D1) was microinjected at the D1 anchor points (purple and green) and imaged at t0 (left) and after 10 hours (centre). Given the difficulty of imaging the arterial pole at high resolution by whole-mount microscopy, cryosections were performed (right, pink frame) to measure the geodesic distance between anchor points in a coronal plane (arterial pole in yellow). Multiphoton images were acquired at a voxel size of 0.57 × 0.57 × 2.5–6.0 µm. Cryosection images were acquired at a pixel size of 0.65 × 0.65 × 0.042 µm.

Fate map of the ventricle primordium shows the tissue dynamics underlying ventricle formation.

A), longitudinal rectangular groups of pseudo-cells were regularly labeled in the ventricle of the stage 8 model. The labelled regions were then tracked back through stage 2. B), Schematic representation of regional deformations at stage 2 and stage 8. C) Representation of the main growth and anisotropy areas in the myocardium. D) Representation of the main directions of tissue motion during heart tube formation.

Stepwise Deformation Pattern Analysis.

Deformation analysis for stage 3 and stage 4. The number of specimens averaged at stage 2 is 5, while at stage 4 it is 6.

Stepwise Deformation Pattern Analysis.

Deformation analysis for stage 5 and stage 6. The number of specimens averaged at stage 5 is 2, while at stage 6 it is 2.

Stepwise Deformation Pattern Analysis.

Deformation analysis for stage 7, stage 8 and stage 9. The number of specimens averaged at stage 7 is 1, at stage 8 is 2, and at stage 9 is 1.

Standard deviation of the growth rate and the anisotropy patterns.

Colormaps indicate the spatial distribution of the Standard Deviations for the growth rate and anisotropy magnitudes shown from Ventral, Dorsal and Caudal views. Stages 6,7,9 have only one staged Live-Shape and therefore variability could not be measured.

Mean growth rate derived from stepwise deformations between consecutive developmental stages for individual embryos, mapped onto the Atlas geometry (caudal view).

Values represent superficial tissue changes, with values < 1 indicating compression, values = 1 isochoric deformation, and values > 1 expansion. Partial IFTs and the arterial pole are not present in the Atlas. Developmental stages represented by more than one embryo are included.

Mean anisotropy rate derived from stepwise deformations between consecutive developmental stages for individual embryos, mapped onto the Atlas geometry (caudal view).

Values represent superficial tissue changes, with values = 1 isotropic deformation, and values > 1 anisotropic. Partial IFTs and the arterial pole are not present in the Atlas. Developmental stages represented by more than one embryo are included.

Dataset of TAT-Cre microinjection and cell Dye microinjection.

eCre(20h) image was captured with Flash procedure (Messal et al.,2021).

Dataset of cell dye microinjection experiments reporting labeling information for the right and left sides of the heart tube (HT).

Embryos e1–e3 were used to measure D1, computed as a geodesic distance. Embryos e4–e5 were used to measure D2, computed as a Euclidean point-to-point distance between fixed landmarks. Both the landmarks and the ROI are available in the Mendeley dataset (Figure 5). Fold changes are reported between t(0) and t(end) over an approximately 10-hour time window.