Cell-fibronectin interactions propel vertebrate trunk elongation via tissue mechanics

Abstract

During embryonic development and tissue homeostasis, cells produce and remodel the extracellular matrix (ECM). The ECM maintains tissue integrity and can serve as a substrate for cell migration. Integrin α5 (Itgα5) and αV (ItgαV) are the α subunits of the integrins most responsible for both cell adhesion to the ECM protein fibronectin (FN) and FN matrix fibrillogenesis. We perform a systems-level analysis of cell motion in the zebrafish tail bud during trunk elongation in the presence and absence of normal cell-FN interactions. Itgα5 and ItgαV have well-described roles in cell migration in vitro. However, we find that concomitant loss of itgα5 and itgαV leads to a trunk elongation defect without substantive alteration of cell migration. Tissue-specific transgenic rescue experiments suggest that the FN matrix on the surface of the paraxial mesoderm is required for body elongation via its role in defining tissue mechanics and intertissue adhesion.

Figure 1

Axis elongation defects after loss of both itgα5 and itgαV. (A,B) Wild-type and (C,D) truncated itgα5^mo;αV^mo embryos at the end of trunk elongation, i.e. 16-somite stage embryos (A,C) and 24 hpf (B,D). At the 16-somite stage, we find that distance from the otic vesicle to the anterior of the head in itgα5^mo;αV^mo embryos (n=30) is 74% (std. 8%; p<0.05) of wild type (n=20) and that the distance from the otic vesicle to the tip of the tail is 71% of wild type (std. 7%; p<0.05). In A–D, anterior is left. (E–L) In situ hybridization of tailbud gene expression in 13 somite stage embryos. FN immunolocalization in 16 somite stage wild type (M) (n=10) and itgα5^mo;αV^mo embryos (N) (n=17). Note the reduction in FN matrix as well as the prominent medial-lateral fiber orientation in N. Scale bars are 50 µm. In E–N, anterior is up. See also Figure S1.

Figure 2

Quantitative analysis of cell motion in the tailbud. (A,B) Tailbud cell tracks were divided into four regions: ADM (magenta), DM (red), PZ (green), PSM (cyan). A is a dorsal view and B is a lateral view of 12–14 somite stage embryo. Anterior is left. See also Movie S1. (C,D) Mean track speeds. (E,F) The means of the coefficient of variation (C.V.). (G,H) Track straightness: length divided by displacement. In C–H, data are plotted for 3 wild-type embryos and 4 itgα5^mo;αV^mo embryos. Datasets averaged 163 minutes in length. P-values calculated via ANOVA with additional validation by permutation tests. (I,J) The vector displacement map averages cell motion in sectors. The heat map indicates mean speed with warmer colors indicating higher speeds and arrows signify averaged 3-D velocity vectors. (K,L) The top 10% of PZ tracks exhibiting the largest displacement in each direction (dorsal to ventral [green], medial to lateral [yellow], posterior to anterior [red] and ventral to dorsal [blue]). (M,N) 3-D Finite Element Method was used to measure the vorticity within the cell flow. Arrows slanting rightward indicate a dorsal to ventral curvature. The ADM is omitted from the vector map and FEM to better visualize motion in the PSM. See also Figure S2.

Figure 3

Characterizing relative cell motion: directed and diffusive movement, global order and local order. Data are plotted for 3 wild-type embryos and 4 itgα5^mo;αV^mo embryos. (A–D) A Bayesian analysis of the MSD determined that cell motion in the tailbud is best modeled using velocity magnitude (A,B) and diffusion coefficient (C,D) as parameters [34]. (E,F) Polarization (Φ) measures global order within each domain of the tailbud. (G,H) Neighbor similarity quantifies the local order of cell motion. The angles between all adjacent instantaneous velocity vectors for each timepoint are binned such that 1 represents parallel motion, −1 anti-parallel motion, 0 orthogonal motion and the intervening deciles denoting intermediate angles. The percentage of angles in each bin for the three wild-type and four itgα5^mo;αV^mo embryos are plotted to give the angle distribution. Overall, there is a switch in the characteristics of the cell motion as cells migrate from the DM to the PZ and from the PZ to the PSM. The pattern of cell motion is maintained in itgα5^mo;αV^mo embryos. P-values calculated via paired and unpaired t-tests.

Figure 4

itgα5 function in the paraxial mesoderm is sufficient to rescue body elongation in itgα5^mo;αV^mo embryos. (A) In wild-type embryos, cells on the surface of the paraxial mesoderm display low levels of blebbing (asterisks) as revealed by phalloidin staining. (B) In itgα5^mo;αV^mo embryos, cells along the medial surface of the paraxial mesoderm exhibit a dramatic increase in blebbing (brackets). See also Movie S2. (C–E) Phalloidin staining of the cell cortices shows the close alignment of the notochord and paraxial mesoderm in wild type embryos (C) (n=16), the loss of this inter-tissue adhesion and alignment in itgα5^mo;αV^mo embryos (D) (n=31) and the rescue of inter-tissue adhesion and organization in Tg(tbx6l:itgα5-RFP);itgα5^mo;αV^mo embryos (E) (n=10). Body elongation is rescued in Tg(tbx6l:itgα5-RFP);itgα5^mo;αV^mo embryos at 14 hpf (F) and 24 hpf (G). At the 16-somite stage, the trunks of transgenic rescue embryos (n=40) are 88% of wild type (std. 4%; p<0.05), and the heads are 83% of wild type (std. 5%; p<0.05). RFP fluorescence shows Itgα5 expression in the paraxial mesoderm. (H) Tg(tbx6l:itgα5-RFP) rescues FN matrix assembly and fiber orientation in itgα5^mo;αV^mo embryos (n=7). Anterior is left. Note that panels C–E and H are composites of anterior and posterior images. Scale bars equal 50 microns. See also Figure S3.