A Moving Source of Matrix Components Is Essential for De Novo Basement Membrane Formation

Abstract

The basement membrane (BM) is a thin layer of extracellular matrix (ECM) beneath nearly all epithelial cell types that is critical for cellular and tissue function. It is composed of numerous components conserved among all bilaterians [1]; however, it is unknown how all of these components are generated and subsequently constructed to form a fully mature BM in the living animal. Although BM formation is thought to simply involve a process of self-assembly [2], this concept suffers from a number of logistical issues when considering its construction in vivo. First, incorporation of BM components appears to be hierarchical [3-5], yet it is unclear whether their production during embryogenesis must also be regulated in a temporal fashion. Second, many BM proteins are produced not only by the cells residing on the BM but also by surrounding cell types [6-9], and it is unclear how large, possibly insoluble protein complexes [10] are delivered into the matrix. Here we exploit our ability to live image and genetically dissect de novo BM formation during Drosophila development. This reveals that there is a temporal hierarchy of BM protein production that is essential for proper component incorporation. Furthermore, we show that BM components require secretion by migrating macrophages (hemocytes) during their developmental dispersal, which is critical for embryogenesis. Indeed, hemocyte migration is essential to deliver a subset of ECM components evenly throughout the embryo. This reveals that de novo BM construction requires a combination of both production and distribution logistics allowing for the timely delivery of core components.

Keywords: Drosophila; basement membrane; cell migration; collagen IV; extracellular matrix; hemocyte; laminin; macrophage; morphogenesis; perlecan.

Figure 1

Temporal Hierarchy of BM Component Expression during Embryogenesis (A) Time-lapse imaging of Drosophila embryos expressing GFP-tagged BM components Laminin α (LanA), Collagen IV (Col IV), and Perlecan (Perl). Scale bar, 100 μm. (B) Quantification of embryonic fluorescence of BM components. Note that LanA is expressed first, followed by Col IV, and finally Perl. Mean ± SEM. (C) Quantification of embryonic fluorescence of BM components in control and srp^AS embryos. Mean ± SEM. Insets show quantification of the maximum fluorescence in each single embryo. Bars in the insets indicate median ± interquartile range. (D) Localization of GFP-tagged BM proteins on the ventral surface of the nerve cord in mutant backgrounds of opposite components. LanA-GFP and Col IV-GFP were imaged at stage 15, while Perl-GFP was imaged at stage 17 of development. Scale bar, 20 μm. (E) Quantification of embryonic fluorescence of GFP-tagged BM components in mutant backgrounds. Bars indicate median ± interquartile range. See also Figure S1, Table S1, and Movie S1.

Figure 2

During BM Deposition, LanA Is Diffusing in the Embryonic Hemocoel while Col IV and Perl Are Locally Deposited (A) Confocal imaging of GFP-tagged BM proteins (green) and hemocytes (magenta) during development. Scale bar, 50 μm. (B) High-magnification images of LanA-GFP (stage 15), secreted-GFP (SecrGFP) (stage 15), Col IV-GFP (stage 16), and Perl-GFP (stage 17). Scale bar, 10 μm. Graphs are line scans of BM component fluorescence with arrows highlighting the border of the hemocyte cell body. (C) FRAP analysis of secreted-GFP, LanA-GFP, and Col IV-GFP in stage 15 embryos at the midline of the VNC. Mean ± SEM. Scale bar, 10 μm. (D) TEM of the ventral hemocoel of a stage 15 embryo, with the hemocyte highlighted in blue. Note the hemocyte squeezing between the epithelium and VNC within the narrow hemocoel. Scale bar, 10 μm. (E) TEM of a stage 16 hemocyte migrating along the VNC reveals BM material (arrows) beneath migrating cells. Scale bar, 2 μm. (F and G) Time-lapse imaging by lattice light-sheet microscopy of Col IV deposition (green) and hemocytes (magenta) at stage 14 of embryogenesis, as hemocytes migrate laterally from the ventral midline. (F) Low magnification. (G) High-magnification view highlighting the correlation between the leading edges of hemocytes and Col IV deposition. Scale bars, 10 μm. See also Figure S2, Table S1, and Movies S2 and S3.

Figure 3

Hemocyte Dispersal Is Required for Even Delivery of Col IV throughout the Embryo (A) Time-lapse imaging of LanA distribution in control and Pvf2-overexpressing embryos, which leads to hemocyte clumping in the head. (B) High-magnification images from the regions highlighted by the rectangle in the control embryo in (A). A sheet-like structure containing Laminin (asterisk) extends laterally from the midline over time. (C) High-magnification images from the regions highlighted by the rectangle in the Pvf2-overexpressing embryo in (A). The Laminin sheet (asterisk) fails to extend, and the spaces not covered by Laminin (cross) enlarge over time. (D) Time-lapse imaging of Col IV distribution in control and Pvf2-overexpressing embryos. (E) Imaging of Col IV distribution at late stages revealing a diffuse distribution in the lateral hemocoel (arrows). Right panels are high-magnification images of the highlighted regions. (F) Hemocyte tracks or “footprints” were revealed by maximum-intensity projection of time-lapse images and correlated with Col IV localization. Right panels are high-magnification images of the highlighted regions showing Col IV localization (top panels) and Col IV localization with hemocyte footprints (bottom panels). Scale bars, 50 μm. Time points indicate time after the start of imaging (hr:min). See also Figure S3, Table S1, and Movie S4.

Figure 4

A Failure in Hemocyte Delivery of BM Leads to Morphogenetic Defects and Embryonic Lethality (A) Scanning electron microscopy of filleted embryonic VNC was performed to reveal the developing BM ensheathing the nerve cord. Lower panels show enlarged regions of the VNC highlighted in upper panels. Arrowheads, fibrils; arrows, BM holes; m, muscles; dn, dorsal nerves. Scale bars, 10 μm. (B) Scanning electron microscopy of the VNC from control and a Pvf2-overexpressing embryo. Enlarged images represent highlighted regions of the VNC from the head and tail regions, respectively. Arrow, BM hole. Scale bars, 20 μm. (C) Hemocyte migration was inhibited by overexpression of Pvf2, or hemocyte-specific expression of RacV12 or N17, and the VNC (brackets) was subsequently imaged using a glial-specific marker (RepoCherry). Scale bar, 100 μm. (D) Quantification of the distance the VNC condensed 8 hr after the start of the condensation process when hemocyte migration was inhibited by overexpression of Pvf2, or hemocyte-specific expression of RacV12 or N17. Bars indicate mean ± SD. (E) Quantification as in (D) in Col IV heterozygous mutants, Pvf2 overexpression, and embryos heterozygous for Col IV while simultaneously overexpressing Pvf2. Bars indicate mean ± SD. (F) The percentage of embryos that failed to hatch was quantified when causing aberrant hemocyte migration (RacN17), when hemocyte migration was combined with heterozygous BM mutants (ΔColIV and ΔLanB1), or in homozygous BM mutants. See also Figure S4, Table S1, and Movie S4.