Asymmetric cell convergence-driven zebrafish fin bud initiation and pre-pattern requires Tbx5a control of a mesenchymal Fgf signal

Abstract

Tbx5 plays a pivotal role in vertebrate forelimb initiation, and loss-of-function experiments result in deformed or absent forelimbs in all taxa studied to date. Combining single-cell fate mapping and three-dimensional cell tracking in the zebrafish, we describe a Tbx5a-dependent cell convergence pattern that is both asymmetric and topological within the fin-field lateral plate mesoderm during early fin bud initiation. We further demonstrate that a mesodermal Fgf24 convergence cue controlled by Tbx5a underlies this asymmetric convergent motility. Partial reduction in Tbx5a or Fgf24 levels disrupts the normal fin-field cell motility gradient and results in anteriorly biased perturbations of fin-field cell convergence and truncations in the pectoral fin skeleton, resembling aspects of the forelimb skeletal defects that define individuals with Holt-Oram syndrome. This study provides a quantitative reference model for fin-field cell motility during vertebrate fin bud initiation and suggests that a pre-pattern of anteroposterior fate specification is already present in the fin-field before or during migration because perturbations to these early cell movements result in the alteration of specific fates.

Keywords: Cell migration; Chemoattractant; Fgf24; Holt–Oram syndrome; Lateral plate mesoderm; Limb bud (fin bud); Limb-field (fin-field); Tbx5a; Zebrafish.

Fig. 1.

A single-cell-resolution fate map of the pectoral fin-field in the zebrafish LPM. (Ai) Dorsal view of 14 hpf embryo previously injected with Kaede RNA in a single blastomere. Box shows a single cell photo-converted to express red fluorescence. Insets show the boxed region at higher magnification of green Kaede fluorescence (upper), red photo-converted cell (middle) and overlaid fluorescence with differential interference contrast microscopy (bottom). (Aii) Same embryo in lateral view at 28 hpf. Box shows resultant clone in cardiac region; insets are as in Ai. (Bi,Bii) Fin bud clone. (Ci,Cii) Peritoneum clone. (Di,Dii) Midline clone. Scale bars: 100 µm in overviews; 10 µm in insets. (E) Fate map of LPM cell fates at 14 hpf. x-axis: AP position relative to somite borders. y-axis: ML position from midline of embryo. Numbered circles denote somite positions. Red dashed lines delineate the fin-field at 14 hpf. dc, duct of Cuvier; ML, mediolateral; o, otic vesicle.

Fig. 2.

Time-lapse and cell tracking of the fin-field LPM during pectoral fin bud initiation in wild-type embryos and Tbx5a and Fgf24 morphants. (Ai-Diii′) Time-lapse of double transgenic Et(hand2:eGFP)ch2; Tg(h2afx:h2afv-mCherry)mw3 embryos with backgrounds of wild-type embryos (Ai-Di′), Tbx5a morphant (Aii- Dii′) and Fgf24 morphant (Aiii-Diii′). (Ai-Di,Aii-Dii,Aiii-Diii) Time-lapse stills showing convergence over time of eGFP-expressing fin-field cells. (Ai′-Di′,Aii′-Dii′,Aiii′-Diii′) Corresponding stills of mCherry-labeled nuclei overlaid with cell trajectories indexed by random colors. Dorsal view, anterior left. Scale bars: 50 µm. Arrowheads in Ai′-Di′ highlight the extent of the fin-field.

Fig. 3.

Quantitative analyses of cell tracks of the fin-field LPM during pectoral fin bud initiation in wild-type embryos and Tbx5a and Fgf24 morphants. (Ai-Fiii) Each row shows identical quantification from eight embryos of the wild-type (Ai-Fi), Tbx5a morphant (Aii-Fii) and Fgf24 (Aiii-Fiii) background. (Ai-Aiii) Composite cell tracks, in which the color-coded groups, red, orange, green, blue and purple, represent LPM cells from somite levels 1, 2, 3, 4 and 5+, respectively. This color code is the same throughout all subsequent panels. (Bi-Biii) Polar plot showing the net average orientation and magnitude of displacement of the five track groups. (Ci-Ciii) Average net AP trajectories of the five track groups, subtracting the average AP trajectories of all LPM track groups. Fine lines represent group average from individual embryos; bold lines represent composite average from each group. Dotted circles denote average AP extent of fin-field domains at 18 (red) and 23 hpf (green). (Di-Diii) Scatter value along the AP axis by time. Dashed lines indicate the peritoneum field group. Solid lines indicate the fin-field group. Lines represent eight-embryo average; shaded areas represent 95% confidence interval. In Dii and Diii, open arrowheads show scatter of track group 5+ and solid arrowheads show scatter of track groups 1-4. Black shows value of morphant, whereas gray shows value of wild type (reproduced from Di). (Ei-Eiii) Correlation of starting AP positions with ending AP positions of tracked LPM cells. Solid line represents a linear least-squares regression of all data points; r² is the coefficient of determination. (Fi-Fiii) AP displacement value for individual track groups. Dashed vertical line represents the rough boundary at somite 4-5 between the fin-field and the peritoneum field. Solid lines represent eight-embryo average; shaded areas represent 95% confidence interval. In Fii and Fiii, black arrowhead shows value of morphant, whereas gray shows value of wild type (reproduced from Fi).

Fig. 4.

The mRNA expression pattern of fgf24 overlaps with the posterior domain of tbx5a expression. (A-H) mRNA expression pattern of tbx5a (A-D) and fgf24 (E-H) at 18, 21, 23 and 25 hpf. (I-L) Schematics of mRNA expression patterns of tbx5a (red dots) and fgf24 (green dots), highlighting the overlap between tbx5a and fgf24 mRNA expression at the posterior limb-field. White ovals denote somites 1-5. Dorsal view, anterior up.

Fig. 5.

Fin-field LPM cells in Fgf24 morphants can converge towards Fgf8b-coated beads. (A-H) Stills from time-lapse of Fgf24 morphant embryos implanted with either a BSA-coated bead (A-D) or an Fgf8b-coated bead (E-H). Scale bars: 50 μm. (I) Temporal progression of relative distances of tracked cells to implanted BSA- or Fgf8b-coated beads. Solid lines show eight-embryo average. Shaded areas show eight-embryo 95% confidence interval. n=20 tracks per embryo. (J,K) Average speed (J) and persistence (K) of tracked cells in Fgf24 morphant embryos with BSA- or Fgf8b-coated beads. Error bars show eight-embryo 95% confidence interval. (L-W) Dorsal views of fin bud region (brackets) in 36 hpf embryos showing pea3 (L-O), fgf10 (P-S) and tbx5a (T-W) expression in wild types, Fgf24 morphants, Fgf24 morphants with a BSA-coated bead and Fgf24 morphants with an Fgf8b-coated bead. Arrowheads point to implanted beads.

Fig. 6.

Incomplete cell convergence in the fin-field of Tbx5a partial morphants is correlated with anteriorly biased truncations in fin buds. (A,B) Maximal intensity projections of the pectoral cartilaginous disc in 4 dpf Tg(b-act2:Brainbow1.0L)^pd49 larvae of an otherwise wild-type embryo (A) and a Tbx5a partial morphant (B); asterisks highlight lesions on the anterior edge of the cartilaginous disc. (C) Polar plot showing net average orientation and magnitude of displacement of the five track groups in eight Tbx5a partial morphant embryos. (D) Average net AP trajectories of five track groups relative to the overall trajectories in all LPM tracks in eight Tbx5a partial morphant embryos. Dashed circles highlight the similar AP extent of fin-field domains between 18 (red) and 23 hpf (green). (E) Scatter values in eight wild-type embryos (blue arrowhead), eight Tbx5a full morphants (pink arrowhead) and eight Tbx5a partial morphants (black arrowhead). Solid arrowhead represents fin-field; open arrowhead represents peritoneum field. Lines represent eight-embryo average; shaded areas represent 95% confidence interval. (F-I) Expression patterns of hoxc6a (F,H) and hoxd12a (G,I) in the 36 hpf fin buds of wild-type (wt; F,G) and Tbx5a partial morphant (pMO) embryos (H,I; dorsal view, anterior up). (J) Quantification of expression areas of hoxc6a (red) and hoxd12a (blue) in wild-type embryos (dark color) and Tbx5a partial morphants (light color; n=40 each). (K) AP displacement plotted against starting AP positions in eight wild-type embryos (blue arrowhead), eight Tbx5a full morphants (red arrowhead) and eight Tbx5a partial morphants (black arrowhead). Dashed vertical line represents the rough boundary at somite 4-5 between the fin and peritoneum fields. Solid lines represent eight-embryo average; shaded areas represent eight-embryo 95% confidence interval. (L) Track ending AP positions of fin-field tracks in eight wild-type embryos (WT) and eight Tbx5a partial morphants (pMO). CoM, center of mass; CoM-75/CoM+75, anterior/posterior limits of the wild-type fin bud; CoM-34/CoM+34, anterior/posterior limits of the Tbx5a partial morphant fin bud. (M) Quantification of incorporation ratio of anterior (red, somites 1-2) and posterior (blue, somites 3-4) fin-field tracks in wild-type embryos (dark color) and Tbx5a partial morphants (light color; n=40 each). Wilcoxon test: ***P=9.1×10⁻⁴ (A tracks wild type-Tbx5a pMO); ***P=1.6×10⁻⁴ (P tracks wild type-Tbx5a pMO); ***P=2.7×10⁻³ (A tracks Tbx5a pMO-P tracks Tbx5a pMO). (N) Models of fin bud initiation in various conditions. First column: in wild types, fin-field cells undergo asymmetric and topological AP convergence. Fin-field positional information (rainbow color) translates into AP axis patterning in the fin. Second column: in Tbx5a or Fgf24 full morphants, all fin-field cells fail to converge, resulting in initiation failure and absence of fins. Third column: in Tbx5a or Fgf24 partial morphants, some fin-field cells fail to converge, with anterior fin-field cells preferentially mis-migrating. Given that positional information might be specified in the early fin-field, reduced incorporation of anterior progenitors results in the anteriorly biased truncation of the fin.