Mural lymphatic endothelial cells regulate meningeal angiogenesis in the zebrafish

Journal name:
Nature Neuroscience
Volume:
20,
Pages:
774–783
Year published:
DOI:
doi:10.1038/nn.4558
Received
Accepted
Published online

Abstract

Mural cells of the vertebrate brain maintain vascular integrity and function, play roles in stroke and are involved in maintenance of neural stem cells. However, the origins, diversity and roles of mural cells remain to be fully understood. Using transgenic zebrafish, we identified a population of isolated mural lymphatic endothelial cells surrounding meningeal blood vessels. These meningeal mural lymphatic endothelial cells (muLECs) express lymphatic endothelial cell markers and form by sprouting from blood vessels. In larvae, muLECs develop from a lymphatic endothelial loop in the midbrain into a dispersed, nonlumenized mural lineage. muLEC development requires normal signaling through the Vegfc–Vegfd–Ccbe1–Vegfr3 pathway. Mature muLECs produce vascular growth factors and accumulate low-density lipoproteins from the bloodstream. We find that muLECs are essential for normal meningeal vascularization. Together, these data identify an unexpected lymphatic lineage and developmental mechanism necessary for establishing normal meningeal blood vasculature.

At a glance

Figures

  1. Mural lymphatic endothelial cells are present at the adult meninx.
    Figure 1: Mural lymphatic endothelial cells are present at the adult meninx.

    (a) Dorsal confocal image of a 12-month-old adult Tg(-5.2lyve1b:DsRed)nz101;Tg(kdrl:EGFP)s843 zebrafish diencephalon, showing the presence of lyve1-expressing cells (red) adjacent to kdrl-EGFP-expressing blood vessels (green). Representative image of n = 6 adult male and female brains analyzed. Scale bar, 200 μm. (b) Higher magnification of dashed outline in a, showing lyve1-expressing cells adjacent to kdrl-EGFP-expressing blood vessels. Scale bar, 100 μm. (c) Higher magnification of region equivalent to b (inset: boxed region), showing that lyve1-expressing cells do not form lumenized vessels and appear to contain vacuoles. Scale bar, 25 μm. (df) Confocal images of a cross-section, indicated by the horizontal bar in d, of a 12-month-old adult zebrafish diencephalon showing lyve1-expressing cells present in the meninx (arrowheads). Representative images of n = 3 sectioned male and female adult brains analyzed. Scale bar, 200 μm. (g) Magnification of boxed region in f showing lyve1-expressing cells (arrowheads) present only adjacent to blood vessels in the meninx. Representative image of n = 3 sectioned adult brains analyzed. Scale bar, 20 μm. (h) Schematic diagram showing the sagittal sections (vertical lines) in i and j. (i) Confocal image of a sagittal section of a 12-month-old adult zebrafish brain showing lyve1-expressing cells (muLECs) present in the forebrain, midbrain and hindbrain (arrowheads). Representative image of n = 3 sectioned male and female adult brains analyzed; scale bar represents 200 μm. (j) Magnification of region shown in the inset, showing confocal image of a sagittal section of muLECs adjacent to the major arteries entering the brain ventrally in an adult zebrafish brain (arrowheads). Scale bar, 200 μm.

  2. muLECs and meningeal lymphatics are present in the zebrafish larval brain.
    Figure 2: muLECs and meningeal lymphatics are present in the zebrafish larval brain.

    (a) Dorsal and (b) lateral confocal images of 7-mm Tg(-5.2lyve1b:DsRed)nz101;Tg(prox1a:KalTA4uq3bh;10xUAS:Venus) midbrain and hindbrain. White arrows indicate the presence of lymphatic vessels co-expressing prox1a and lyve1 at the level of the brain connected to the otolithic lymphatic vessel (arrowheads; n = 6 larvae). muLECs are indicated by a yellow arrow. Expression of prox1a is mosaic due to use of the KaltA4–UAS system25. See Supplementary Figure 1e for quantification. Scale bars, 100 μm. (c) Dorsal confocal image of 7-mm embryo with a schematic overlay of the brain showing the location of cells co-expressing prox1a and lyve1. FB, forebrain; MB, midbrain; HB, hindbrain; scale bar, 100 μm. (d) Dorsal confocal image of Tg(-5.2lyve1b:DsRed)nz101;Tg(pdgfrβ:EGFP)uq15bh 7-mm larvae showing lyve1-expressing muLECs forming a distinct population to the pdgfrβ-expressing pericytes (196 cells from n = 3 larvae). NC, neural-crest-derived cells; scale bar, 100 μm. (e) Confocal image of the surface of the brain, showing muLECs present in a more distal mural niche than perivascular pericytes (n = 2 adult brains). Scale bar, 100 μm. (f) Dorsal confocal image of Tg(-5.2lyve1b:BFPCaax)uq18bh;Tg(pdgfrβ:EGFP)uq15bh;Tg(kdrl:Cherry)s916 showing the three distinct cell types: muLECs (blue), pericytes (green) and endothelial cells (red; 172 cells from n = 2 larvae). Scale bar, 100 μm. (g) Dorsal confocal image of Tg(mpeg1:mcherry);Tg(prox1a:KalTA4uq3bh;10xUAS:Venus) 7-mm larvae28 showing prox1a-expressing muLECs forming a population distinct from mpeg1-expressing macrophages (n = 11 larvae; scale bar, 100 μm). See Supplementary Figure 1f for quantification.

  3. muLECs form by sprouting lymphangiogenesis and disperse throughout the larval meninx.
    Figure 3: muLECs form by sprouting lymphangiogenesis and disperse throughout the larval meninx.

    (ac) Dorsal confocal images of cells over the right side of the midbrain in Tg(-5.2lyve1b:DsRed)nz101;Tg(kdrl:EGFP)s843 from 5 to 15 dpf showing lyve1-expressing cells of the vascular loop becoming progressively mesenchymal in appearance (arrowheads; n = 5 larvae per timepoint; scale bar, 50 μm). (d) lyve1-expressing cells are adjacent to kdrl-expressing blood vessels in 7-mm embryos (n = 6 larvae; scale bar, 50 μm). (e) Schematic diagram showing experimental design for Kaede experiments. (f,g) Confocal images of Tg(prox1a:KaltA4)uq3bh;Tg(10xUAS:Kaede)s1999t (f) before (left) and after (middle) photoconversion at 5 dpf and (right) reimaged at 10 dpf; or (g) before (left) and after (middle) photoconversion at 15 dpf and (right) reimaged at 20 dpf (Supplementary Fig. 3). Positive lineage-tracing over a serial 5-d period revealed that a single cell population expanded (5–10 dpf, n = 18 cells counted from 3 embryos; 15–20 dpf, n = 53 cells from 6 embryos). Additional data and quantification of cell expansion are provided in Supplementary Figure 5a–c. Scale bar, 50 μm. (hk) Time-lapse imaging (lateral view) from 30 hpf to 94 hpf shows cells that give rise to the lymphatic vascular loop, sprout from the choroidal vascular plexus that develops parallel to the primary head sinus (arrowheads). The timestamp corresponds to time elapsed since 30 hpf (Supplementary Movie 2); n = 4 independent movies confirmed this cellular origin. Scale bar, 50 μm; ik show magnified view of the region outlined in h. (l) Lateral confocal images of a wild-type 5-dpf Tg(-5.2lyve1b:DsRed)nz101;Tg(fli1a:nEGFP)y7 zebrafish embryo showing the putative lymphatic loop. Representative image of n = 7 embryos analyzed. Scale bar, 100 μm. (mp) Confocal images of the boxed area in l, showing that the loop that forms in (m) wild-type embryos (arrowhead; n = 7 embryos) is reduced in (n) vegfchu5055 (arrowheads; n = 6 embryos) and (o) vegfduq7bh (arrowhead; n = 6 embryos) single mutants but is absent in (p) vegfchu5055vegfduq7bh double mutants (asterisk; n = 9 embryos). Additional data and quantification of cell numbers are shown in Supplementary Figure 7a–i. Scale bar, 50 μm. (qs) Dorsal confocal images showing the muLECs present in 3-month-old (q) wild-type adult fish (n = 4 adult brains) are absent in (r) vegfchu5055vegfduq7bh double mutants (n = 4 adult brains) and (s) vegfr3hu4062 mutants (n = 4 adult brains). Scale bar, 200 μm. (t) Quantification of lyve1-expressing cells in 500 μm2 in 3-month-old male and female wild-type (n = 4 adult brains), vegfr3hu4062 single-mutant (n = 4 adult brains) and vegfchu5055vegfduq7bh double-mutant (n = 4 adult brains) fish. Error bars represent mean ± s.e.m.; **P < 0.01, from one way ANOVA (F2,10 = 14.69).

  4. muLECs are immediately adjacent to endothelium and take up LDL.
    Figure 4: muLECs are immediately adjacent to endothelium and take up LDL.

    (a) Immunoelectron microscopy showing a muLEC in relation to the basement membrane (BM) of an endothelial cell and the zebrafish meninx. Large vacuoles or inclusions (asterisks) are present in the muLEC immediately adjacent to the vascular matrix. EC, endothelial cell; L, blood vessel lumen; Mn, meninx. Scale bar, 1 μm. (b) False-colored immunoelectron microscopy from a. The muLEC is proximal to but not embedded in the BM. Scale bar, 1 μm; arrowhead, BM; arrow, collagen fibers of the vascular matrix. (c) Dorsal confocal images of muLECs over the right side of the midbrain in Tg(-5.2lyve1b:DsRed)nz101 3 h after Alexa Fluor-488-labeled acetylated LDL was injected into the blood stream. LDL is present in the endothelial cells as well as 61% ± 11% (mean ± s.d.) of the muLECs (n = 72 muLECs scored from 3 embryos). Scale bar, 100 μm. (d) 24 h after LDL injection as described above, LDL accumulated intracellularly in 100% of muLECs. Arrows, inclusions containing labeled LDL; n = 73 muLEC cells from 3 embryos. Left, LDL; middle, muLECs; right, merged image. Scale bar, 10 μm.

  5. Mural lymphatic endothelial cells are transcriptionally distinct from other mural cell types and produce endothelial growth factors.
    Figure 5: Mural lymphatic endothelial cells are transcriptionally distinct from other mural cell types and produce endothelial growth factors.

    (a) Gene ontology analysis of positively regulated pathways for genes differentially expressed between FAC-sorted muLECs (n = 3 FAC sorts) and 3 independent postembryonic zebrafish tissues (adult head, adult tail and 5-d larvae). P values represent Bonferroni-adjusted values. Gene ontology codes: 0001525, 0001944, 0048514, 0051348, 0050679, 1902531, 0031400 and 0032269. (b) Three-dimensional principle component (PC) analysis comparing transcriptomes of pericytes, muLECs, macrophages, larval-stage ECs and embryonic ECs. Full details in Online Methods and Supplementary Figure 8. (c) Heat-map representation of relative gene expression levels (YuGene analysis; full details in Online Methods) for genes encoding selected markers: known macrophage marker genes csf1ra, stabilin1 and mafba; known blood endothelial marker genes kdrl, cdh5, sox7, sox18 and tek; known pericyte marker genes pdgfrβ and notch3; and known lymphatic endothelial marker genes lyve1b, vegfr3, stabilin1, stabilin2, prox1a, nrp2a and mafba. Scale bar indicates relative expression, from 1 (highest) to 0 (lowest). (d) Expression levels in reads per kilobase per million mapped reads (RPKM) of endothelial factors in triplicate FAC-sorted muLECs based on RNA-seq data. Y axis represents log2(RPKM + 1). Error bars represent mean ± s.e.m. (e) Expression levels (RPKM) of neurotrophic factors in triplicate FAC-sorted muLECs based on RNA-seq data. Y axis represents log2(RPKM + 1). Error bars represent mean ± s.e.m.

  6. muLECs are correlated with meningeal blood endothelial cell numbers.
    Figure 6: muLECs are correlated with meningeal blood endothelial cell numbers.

    (a) Quantification of meningeal BEC numbers in stage-matched 5.7-mm siblings with normal muLECs and vegfchu5055vegfduq7bh double-mutants lacking muLECs, showing a significant decrease in meningeal BECs. Error bars represent mean ± s.e.m.; ***P < 0.001, two-tailed Student's t test (t13 = 4.267; n = 7 wild-type vs. 8 mutant larvae). See Supplementary Figure 10. (b) Top: dorsal confocal image of 6.5-mm Tg(-5.2lyve1b:DsRed)nz101;Tg(fli1a:nEGFP)y7 larvae with symmetrical distribution of muLECs between left and right hemispheres. Bottom: surface render from z-stacks showing meningeal BECs and the DsRed signal corresponding to muLECs. The numbers correspond to meningeal BECs on the left and right hemispheres of the brain. Scale bar, 100 μm. (c) Top: dorsal confocal image of 6.5-mm Tg(-5.2lyve1b:DsRed)nz101;Tg(fli1a:nEGFP)y7 larvae with asymmetrical distribution of muLECs between left and right hemispheres. Bottom: surface render from a z-stack showing meningeal BECs and the DsRed signal corresponding to muLECs. The numbers correspond to meningeal BECs on the left and right hemispheres of the brain. Scale bar, 100 μm. (d) Quantification of the ratio of meningeal BECs on the left vs. right hemispheres of the brain in larvae with asymmetrical distributions of muLECs (left/right, where the left side corresponds to the side with more muLECs). There are more BECs on the side of the brain with the most muLECs, leading to a significant loss of BEC symmetry. Error bars represent mean ± s.e.m.; ***P < 0.001, two-tailed Student's t test (t11 = 5.042), n = 6 symmetric (wild-type) vs. 7 asymmetric (mutant) larvae.

  7. Asymmetric muLEC ablation leads to asymmetric meningeal angiogenesis.
    Figure 7: Asymmetric muLEC ablation leads to asymmetric meningeal angiogenesis.

    (a) Quantification of the increase in meningeal BEC number in the 96-h period from 10 dpf in the left and right hemispheres in control (wild-type) brains (n = 15 larvae). Y axis shows the ratio of cell numbers at 96 h/cell numbers at 0 h. There is no significant difference between the left and right hemispheres (n = 15 larvae, P = 0.8126, two-tailed Student's t test, t28 = 0.2394). (b,c) Dorsal confocal images of 10-dpf Tg(-5.2lyve1b:DsRed)nz101;Tg(fli1a:nEGFP)y7 (b) before and (c) after (0 hpa) multiphoton laser ablation. (d) Confocal image of the larvae in b and c showing BEC (green) and muLEC (red) distribution 96 hpa. (e) Quantification of BECs on the left and right hemispheres (y axis, ratio of numbers of left/numbers of right BECs) shows there is no change in the left–right symmetry immediately following ablation (n = 27 larvae, P = 0.3024, two-tailed Student's t test, t52 = 1.042). (f) Quantification of muLEC symmetry before and after ablation (y axis, ratio of numbers of left/numbers of right BECs). There is a significant difference in muLEC symmetry following ablation (***P < 0.0001, two-tailed Student's t test, n = 27 larvae, t52 = 51.47). Error bars represent mean ± s.e.m. (g,h) Dorsal confocal images of 10-dpf control (mock-ablation) embryos (g) before and (h) 96 hpa (Supplementary Fig. 11). Scale bar, 100 μm. (il) Surface render of the z-stacks in bd, g and h corresponding to (i) 10-dpf and (k) 96-hpa mock-ablated and (j) 10-dpf and (l) 96-hpa muLEC-ablated larvae, respectively, showing rendered meningeal BECs and the DsRed signal (white) corresponding to muLECs. The BECs on the muLEC-ablated and mock-ablated sides are rendered green, and BECs on the nonablated sides are rendered in magenta. Numbers indicate meningeal BECs on the left and right. Ablation experiments were performed to ablate cells adjacent to muLECs (mock ablation, i) or the muLECs (j). Scale bars, 100 μm. (m) Quantification of the vascularization index (number of meningeal BECs at 96 hpa vs. 0 hpa) of BECs on the left and right hemisphere of the brain in control (n = 15), mock-ablated (n = 15) and muLEC-ablated (n = 27) larvae. There was a significant reduction in the expansion in number of meningeal BECs on the muLEC-ablated side of the brain (***P < 0.001, **P < 0.01, *P < 0.05; one-way ANOVA, F5,114 = 6.961). Error bars represent mean ± s.e.m. (n) Quantification of the symmetry of BECs between left and right hemispheres of the brain (y axis, left/right BEC number ratio) in control larvae (n = 15), mock ablated larvae (n = 15) and larvae with muLECs ablated on one side of the brain (n = 27 larvae) showing a significant decrease in BEC symmetry 96 hpa. Error bars represent mean ± s.e.m. ****P < 0.0001; one-way ANOVA, F5,114 = 28.54. (o) Quantification of the recovery of the muLEC population on the ablated hemisphere expressed as a percentage of the muLECs on the nonablated hemisphere (n = 27 larvae). Error bars represent mean ± s.e.m. (p) Positive correlation between muLEC symmetry and BEC symmetry 96 hpa (Pearson correlation coefficient, P = 0.004; dashed lines represent 95% confidence interval, solid line represents best fit from linear regression, n = 27 larvae; F1,25 = 16.63).

  8. lyve1-expressing cells in the adult zebrafish meninx co-express the lymphatic marker genes prox1a and fli1a
    Supplementary Fig. 1: lyve1-expressing cells in the adult zebrafish meninx co-express the lymphatic marker genes prox1a and fli1a

    (a) Confocal images of adult zebrafish brain in Tg(-5.2lyve1b:DsRed)nz101; TgBAC(prox1a:KalTA4 uq3bh;10xUAS:Venus) animals showing cells that co-express the lymphatic markers prox1a and lyve1. Expression of prox1a is mosaic due to the Gal4 system. Full quantification of marker co-expression can be found in Supplementary Figure 1e. Scale bar represents 100μm.

    (b) Confocal images of adult zebrafish brain in Tg(-5.2lyve1b:DsRed)nz101; Tg(fli1a:EGFP)y1 showing co-expression (arrows) of the lymphatic markers fli1a and lyve1. Full quantification of marker co-expression can be found in Supplementary Figure 1f. Scale bar represents 100μm.

    (c) Quantification of muLEC density on the surface of the brain (meningeal surfaces n=6 adult brains. Error bars represent mean +/− sem, Statistical testing N/A.)

    (d) Quantification of the number of lyve1-positive meningeal cells (over the surface of the tectum, or adjacent to the central arteries (cTA)) compared with non-meningeal lyve1-positive cells from cross sections of the adult zebrafish brain. n=4 adult brain cross sections, p<0.0001, from two-tailed student t-test (t=8.926 df=6).

    (e) Quantification of co-expression of cells from Tg(prox1a:KalTA4uq3bh;10xUAS:Venus) and lyve1 (Tg(-5.2lyve1b:DsRed)nz101) and the macrophage marker mpeg1 (Tg(mpeg1:mcherry)). 215 cells from n=4 larvae (lyve1) and 255 cells from n=5 larvae (mpeg1) were scored from confocal Z-stack images. Error bars represent mean +/− sem, Statistical testing N/A.

    (f) Quantification of co-expression of cells in Tg(-5.2lyve1b:DsRed)nz101 and Tg(fli1a:EGFP)y1, TgBAC(pdgfrβ:EGFP)uq15bh, TgBAC(acta2:EGFP)uq17 and Tg(nkx2.2a:EGFP)vu16Tg. 100 cells from n=5 larvae were scored from confocal Z-stack images. Error bars represent mean +/− sem, Statistical testing N/A.

    (g) Quantification of the distance from the nucleus of muLECs to the center of the parenchyma (defined as the centre of the unlabeled space surrounded by a vascular loop) shows that muLECs are more closely associated with blood endothelial cells than the space between vessels. 77 muLEC nuclei from n=3 larvae, p<0.0001, from two-tailed student t-test (t=15.64 df=130). Error bars represent mean +/− sem.

    (h) Quantification of the distance from the nucleus of a given muLEC to the nearest vessel branch point, compared with the distance from the nucleus of a given muLEC to the midpoint between branches of the closest vessel. muLECs are quantitatively closer to vessel branch points than vessel midpoints. 77 branch and 77 midpoints points from n=3 larvae, p<0.0001, from two-tailed student t-test (t=4.74 df=151). Error bars represent mean +/− sem.

  9. High-molecular-weight dye can be taken up by meningeal lymphatics but not by muLECs.
    Supplementary Fig. 2: High-molecular-weight dye can be taken up by meningeal lymphatics but not by muLECs.

    (a) Dorsal confocal image of 7mm Tg(-5.2lyve1b:DsRed)nz101 zebrafish larvae following injection of FITC-labeled high molecular weight (2000kDa) dextran. Arrowheads indicate the presence of the dye in lymphatic vessels, but the dye is absent from muLECs. Images representative of data obtained from injection of n=6 larvae. Scale bar represents 100μm.

    (b) Magnification of boxed region in (a) showing muLECs do not absorb high MW dextran. Images representative of data obtained from injection of n=6 larvae. Scale bar represents 100μm.

    (c) Dorsal confocal image of a Tg(-5.2lyve1b:DsRed)nz101; Tg(fli1a:EGFP)y1 zebrafish larvae. Image representative of n=5 larvae. Scale bar represents 100μm.

    (d) Magnification of boxed region in (c) showing muLECs co-express lyve1 and fli1a (arrowheads). Image representative of n=6 larvae. Full quantification of marker co-expression can be found in Supplementary Figure 1f. Scale bar represents 100μm.

    (e) muLECs are closely associated with blood vessels in Tg(-5.2lyve1b:DsRed)nz101 larvae. Image representative of n=6 larvae. Scale bar represents 100μm.

    (f) Confocal image of a cross-section of adult (20mm stage, 3 month old) Tg(-5.2lyve1b:DsRed)nz101; TgBAC(pdgfrβ:EGFP)uq15bh brain showing pdgfrβ expressing pericytes are associated with blood vessels throughout the adult CNS. Scale bar represents 100μm. Full quantification of marker co-expression can be found in Supplementary Figure 1f. Image representative of data obtained from injection of n=3 male and female 20mm zebrafish adults.

  10. muLECs do not express markers of neuronal and/or oligodendrocyte cells and smooth muscle cells.
    Supplementary Fig. 3: muLECs do not express markers of neuronal and/or oligodendrocyte cells and smooth muscle cells.

    (a-d) Confocal images of (a,b) 7mm stage Tg(-5.2lyve1b:DsRed)nz101;Tg(nkx2.2a:EGFP)vu16Tg and (c,d) 7mm Tg(-5.2lyve1b:DsRed)nz101;TgBAC(acta2:EGFP)uq17 zebrafish larvae showing a lack of co-expression of these markers (quantified in Supplementary Figure 1f) corresponding to neuronal/oligodendrocyte and smooth muscle respectively. Smooth muscle cells are present around the dorsal aorta in the lateral view of the trunk (arrow in d). Scale bars represent 100μm.

  11. Cells of the putative lymphatic loop express lymphatic markers and are separate from the blood vasculature.
    Supplementary Fig. 4: Cells of the putative lymphatic loop express lymphatic markers and are separate from the blood vasculature.

    (a) Confocal images of a 5dpf Tg(-5.2lyve1b:DsRed)nz101; Tg(kdrl:EGFP)s843 zebrafish midbrain showing that cells of the lymphatic loop do not co-express lyve1 and the blood vessel marker kdrl. n=5 embryos. Full quantification of marker co-expression can be found in Supplementary Figure 1f.

    (b) Confocal images of lyve1/prox1a positive cells of the lymphatic loop at 5dpf in Tg(-5.2lyve1b:DsRed)nz101; Tg(prox1a:KalTA4uq3bh;10xUAS:Venus) zebrafish. n=3 embryos. Full quantification of marker co-expression can be found in Supplementary Figure 1f.

    (c) Confocal images of a 5dpf zebrafish head showing cells of the lymphatic loop co-express the lymphatic markers lyve1 and fli1a in Tg(-5.2lyve1b:DsRed)nz101; Tg(fli1a:EGFP) (arrowheads). n=5 embryos. Full quantification of marker co-expression can be found in Supplementary Figure 1f.

    (d) Confocal images of 5dpf Tg(-5.2lyve1b:DsRed)nz101 embryo injected with dextran into the blood vasculature. The lyve1 positive cells do not contain dextran demonstrating they are separate from the blood vasculature. Scale bars represent 100μm. Representative of analysis from n=4 embryos.

  12. Kaede photoconversion demonstrates that the lymphatic loop gives rise to the meningeal muLEC population.
    Supplementary Fig. 5: Kaede photoconversion demonstrates that the lymphatic loop gives rise to the meningeal muLEC population.

    (a-b) Confocal images of Tg(prox1a:KalTA4)uq3bh; Tg(10xUAS:Kaede)s1999t photoconverted at 10dpf and reimaged at 15dpf (a) and photoconverted at 20dpf and reimaged at 25dpf (b). Scale bars represent 50μm. For the 10-15 dpf stages n=34 cells from 7 larvae were examined and for the 20-25dpf stages n=29 cells from 3 larvae were analysed. Positive lineage tracing was confirmed in all cases. (c) Quantification of the average increase in photo-converted cell number over 5 days (proliferation rate, cell number 5 days after conversion/0 days after conversion) data combined for 5-20 dpf experiments, n=19 larvae overall. Error bars represent mean +/− sem, Statistical testing N/A.

  13. Still images of a Tg(-5.2lyve1b:DsRed)nz101;Tg(fli1a:nEGFP)y7 zebrafish embryo from Supplementary Movie 2.
    Supplementary Fig. 6: Still images of a Tg(-5.2lyve1b:DsRed)nz101;Tg(fli1a:nEGFP)y7 zebrafish embryo from Supplementary Movie 2.

    (a-k) Images are representative of n= 5 movies.

    (a) Tilted view showing the labeled endothelial nucleus in the choroidal vascular plexus (arrowhead).

    (b,c) First cell (grey render, arrow) exits the choroidal plexus vessel and migrates dorsally along another pre-existing vessel.

    (d-f) Second cell (blue render, arrow) exits vessel and migrates dorsally along pre-exiting vessel.

    (g-i) First cell divides (arrowhead) and one daughter cell divides again (arrow).

    (i,j) Second cell (blue) divides (arrowhead) and continues to migrate (k, arrowhead). Scale bar represents 50μm.

  14. muLEC sprouting is dependent upon Vegfc, Vegfd, Ccbe1 and Vegfr3 signaling.
    Supplementary Fig. 7: muLEC sprouting is dependent upon Vegfc, Vegfd, Ccbe1 and Vegfr3 signaling.

    (a-d) Dorsal confocal images of 5dpf Tg(-5.2lyve1b:DsRed)nz101; Tg(fli1a:nEGFP)y7 embryos showing the lymphatic loop that forms in wildtype embryos (a, arrows, representative image based on analysis of n=7 embryos) is reduced in vegfchu5055 (b, arrows, n=6 embryos) and vegfduq7bh (c, arrows, n=6 embryos) single mutants, but is absent in vegfchu5055/vegfduq7bh double mutants (d, asterisk, n=9 embryos). Scale bars represent 100μm.

    (e) Quantification of the number of muLECs present in a single putative lymphatic loop at 5dpf in wildtype (n=7 embryos), vegfchu5055 (n=6 embryos), vegfduq7bh (n=6 embryos) single mutants and vegfchu5055/vegfduq7bh double mutants (n=9 embryos). Error bars represent mean +/− sem; **** p<0.0001, from one way ANOVA (F (3, 25) = 166.6).

    (f-i) At 5 dpf, the lyve1/prox1a positive cells in Tg(-5.2lyve1b:DsRed)nz101; Tg(prox1a:KalTA4, 4xUAS:uncTagRFP)nim5; Tg(10xUAS:Venus) have formed a lymphatic loop in wildtype embryos (f, n=6 embryos), however in vegfchu5055 (g, n=6 embryos) and vegfd (h, n=6 embryos) mutants, the development of the lymphatic loop is reduced (arrows). In vegfchu5055 / vegfduq7bh double mutant embryos (i, n=6 embryos) the lymphatic loop is completely absent (asterisks). Scale bars represent 100μm.

    (j-l) The lyve1 lymphatic loop structure (control, j) is absent in vegfr3 morphant (n=15 embryos) (k) and (l) ccbe1 morphant (n=15 embryos) embryos at 5dpf (asterisks).

    (m) Quantification of the number of muLECs present in a single putative lymphatic loop at 5dpf in uninjected controls (n=15 embryos), vegfr3 morphant (n=15 embryos) and ccbe1 morphant (n=15 embryos). **** p<0.0001, from two-tailed student t-test (t=33.62 df=28). Scale bars represent 100μm.

    (n,o) Dorsal confocal images from 24 mm Tg(-5.2lyve1b:DsRed)nz101 vegfchu5055 (n) and vegfduq7bh (o) mutants showing the presence of either ligand is sufficient for development of the muLECs. Representative image based on analysis of n=4 adult brains. Scale bars represent 200μm.

  15. Mature muLECs display a distinctive ultrastructure and take up LDL.
    Supplementary Fig. 8: Mature muLECs display a distinctive ultrastructure and take up LDL.

    (a,b) Schematic diagram indicating the region used for immuno-electron microscopy.

    (c) Low magnification immuno-electron micrograph image showing overview of imaged region in Fig. 4a,b. Scale bar represents 1μm.

    (d) Higher magnification of image shown in (c). Scale bar represents 1μm. * = prominent vacuoles or inclusion bodies consistent with confocal imaging in Figure 1 and panel g below, L=Lumen, EC= endothelial cell, BM=basement membrane.

    (e) High magnification of muLEC cell showing positive immunostaining by anti-gfp antibody. Scale bar represents 1μm.

    (f) High molecular weight dextran is not observed in muLECs (asterisks) 3 hours post injection into the blood stream of Tg(-5.2lyve1b:DsRed)nz101 larvae. n=63 muLECs were analysed in detail by scoring confocal z-stacks from n= 3 larvae.

    (g) Alexa 488-labelled acetylated LDL is observed in the endothelial cells and the inclusion bodies (arrows) of 61 +/− 11 % of muLECs 3 hours post injection into the blood vasculature. Scale bars represent 20μm. n=72 muLECs were analysed in detail by scoring confocal z-stacks from 3 embryos.

    (h,i) Three hours post injection of high molecular weight dextran, there is no obvious vascular leakage in wildtype (h) or vegfchu5055 / vegfduq7bh double mutant larvae (i) which lack muLECs. Scale bar represents 50μm. n=3 larvae were analysed in detail by scoring confocal z-stacks.

  16. muLECs can be FAC-sorted and separated from blood endothelial cells and pericytes.
    Supplementary Fig. 9: muLECs can be FAC-sorted and separated from blood endothelial cells and pericytes.

    (a) Representative plots of FAC sorted cells from adult brains in unlabeled wildtype and Tg(-5.2lyve1b:DsRed)nz101 positive samples.

    (b) Representative plots of FAC sorted EGFP positive cells in unlabeled wildtype and Tg(-5.2lyve1b:DsRed)nz101; Tg(kdrl:EGFP)s843 positive samples.

    (c) Representative plots of FAC sorted EGFP positive cells in unlabeled wildtype and Tg(-5.2lyve1b:DsRed)nz101; TgBAC(pdgfrβ:EGFP)uq15bh positive samples.

    (d) Scree plot of variance (eigenvalue versus component number) and principle component comparisons from Fig. 5b shown in 2-dimensional plots (PC1 vs PC2, PC2 vs PC3 and PC1 vs PC3), clear separation of muLECs from control samples is observed.

    (e) Representative plots of FAC sorted EGFP positive cells in wildtype and Tg(pdgfrβBAC:EGFP)uq15bh; Tg(kdrl:mCherry)s843 positive samples.

  17. vegfchu5055vegfduq7bh double-mutant larvae have reduced meningeal BEC nuclei compared to wild-type larvae.
    Supplementary Fig. 10: vegfchu5055vegfduq7bh double-mutant larvae have reduced meningeal BEC nuclei compared to wild-type larvae.

    (a) Dorsal confocal image (a) and hyperstack (a’) of z-stacks from the Tg(-5.2lyve1b:DsRed)nz101;Tg(fli1a:nEGFP)y7 signal with darker shades representing z-stack slices closer to the objective, showing vegfchu5055/vegfduq7bh double mutant larvae (a, representative images of n=8 larvae analysed), which have reduced meningeal BEC number when compared to size matched (5.7mm) wildtype siblings (b, n=7 larvae). Scale bar represents 100μm.

  18. Visual confirmation of muLEC ablation and mock ablation.
    Supplementary Fig. 11: Visual confirmation of muLEC ablation and mock ablation.

    (a) Single z-slices from a dorsal confocal image of a 10dpf Tg(-5.2lyve1b:DsRed)nz101;Tg(fli1a:nEGFP)y7 larvae before and after targeted ablation of muLECs. Ablation can be confirmed by direct observation of cell death, with “bubbles” which disperse post ablation. Scale bar represents 50μm. Representative images of n=27 larvae analysed.

    (b) Second example of single z-slices from a dorsal confocal image of 10dpf Tg(-5.2lyve1b:DsRed)nz101;Tg(fli1a:nEGFP)y7 larvae before and after ablation of muLECs. Ablation can be confirmed by direct observation of cell death. The analysis in Fig. 7 used the same approach to verify cell ablations for n=27 embryos in total.

    (c) Example of single z-slices from a dorsal confocal image of 10dpf Tg(-5.2lyve1b:DsRed)nz101;Tg(fli1a:nEGFP)y7 larvae before and after mock ablation in which cells adjacent to the muLECs are targeted. Ablation confirmed by direct observation and this approach was used in Fig. 7 for analysis of n=15 larvae. This is the same larvae as in Fig. 7g,h. Scale bar represents 50μm.

    (d) Second example of single z-slices from a dorsal confocal image of 10dpf Tg(-5.2lyve1b:DsRed)nz101;Tg(fli1a:nEGFP)y7 larvae before and after mock ablation in which cells adjacent to the muLECs are targeted. This is the same larvae as above and in Fig. 7g,h.

    (e and f) 10 dpf Tg(-5.2lyve1b:DsRed)nz101;Tg(fli1a:nEGFP)y7 larvae before (e) and after (f) mock ablation. Representative of n=15 larvae analysed. Scale bar represents 100μm.

  19. Schematic summary of muLEC development and function
    Supplementary Fig. 12: Schematic summary of muLEC development and function

    a. Mural lymphatic endothelial cells (muLECs) emerge from the choroidal vascular plexus (CVP) at 54 hpf (a) and sprout to the periphery of the midbrain at 96 hpf (a’).

    b. The resulting lymphatic vascular loop in the midbrain (MB) begins to undergo a transition to a mesenchymal morphology between 5 (b’) and 10 dpf (b’’)

    c - d. muLECs (blue) are present on the midbrain, fore brain (FB) and hind brain (HB) in 7 mm larvae (c) and expand over the adult brain (d). c’ and d’ provide lateral views and indicate the meningeal lymphatics (green, c’) and muLEC location relative to blood vessels (red) (d’).

    e and f. Three dimensional (e) and cross sectional schematic representation of the blood vasculature and mural cells at the zebrafish meninx. muLECs (blue) secrete vascular growth factors and take up LDL from the bloodstream. PVM= perivascular macrophage.

Videos

  1. 3D reconstruction of Tg(prox1a:KalTA4 uq3bh;10xUAS:Venus) Z-stacks from Fig. 2a.
    Video 1: 3D reconstruction of Tg(prox1a:KalTA4 uq3bh;10xUAS:Venus) Z-stacks from Fig. 2a.
    The meningeal lymphatic vessels and muLECs surround the outer curvature of the brain in a 7mm larvae. Representative of n=6 larvae analysed.
  2. Time lapse imaging of muLECs sprouting from the choroidal vascular plexus.
    Video 2: Time lapse imaging of muLECs sprouting from the choroidal vascular plexus.
    Left side: Confocal time lapse imaging overview of a Tg(-5.2lyve1b:DsRed)nz101; Tg(fli1a:nEGFP)y7 zebrafish embryo from 30 hpf to 4 dpf. At approximately 54 hpf, cells migrated dorsally from a vein running parallel to the primary head sinus and subsequently give rise to the lymphatic loop observed at 4dpf. Representative of n=5 movies of individual embryos analysed.
    Right side: Zoomed and rendered version of left hand movie, showing confocal time lapse imaging of a Tg(-5.2lyve1b:DsRed)nz101; Tg(fli1a:nEGFP)y7 zebrafish embryo from 30 hpf to 4 dpf. Rendered nuclei (grey and blue) highlight cells sprouting at approximately 54 hpf, from a vein running parallel to the primary head sinus, subsequently dividing and up-regulating lyve1. Representative of n=5 movies of individual embryos analysed.
  3. 3D reconstruction showing processing used in Imaris to confirm the correct identification of meningeal BECs as represented in Fig. 6 and Fig.7.
    Video 3: 3D reconstruction showing processing used in Imaris to confirm the correct identification of meningeal BECs as represented in Fig. 6 and Fig.7.
    In this example, the deconvoluted z-stack was surface rendered (white) to detect the EGFP signal from Tg(-5.2lyve1b:DsRed)nz101; Tg(fli1a:nEGFP)y7 14 dpf mock ablated larvae. The EGFP signal corresponding to the meningeal BECs was rendered magenta on the left side and green on the right side. The white render corresponding to the total EGFP signal was removed to leave the meningeal BECs. The channel for DsRed corresponding to the muLECs was added (grey) followed by the surface render of the signal.

Accession codes

Primary accessions

Gene Expression Omnibus

Referenced accessions

Gene Expression Omnibus

Sequence Read Archive

References

  1. Zhao, Z., Nelson, A.R., Betsholtz, C. & Zlokovic, B.V. Establishment and dysfunction of the blood-brain barrier. Cell 163, 10641078 (2015).
  2. Chakraborty, A., de Wit, N.M., van der Flier, W.M. & de Vries, H.E. The blood brain barrier in Alzheimer's disease. Vascul. Pharmacol. 89, 1218 (2017).
  3. Sweeney, M.D., Ayyadurai, S. & Zlokovic, B.V. Pericytes of the neurovascular unit: key functions and signaling pathways. Nat. Neurosci. 19, 771783 (2016).
  4. Prinz, M., Erny, D. & Hagemeyer, N. Ontogeny and homeostasis of CNS myeloid cells. Nat. Immunol. 18, 385392 (2017).
  5. Abbott, N.J. Astrocyte-endothelial interactions and blood-brain barrier permeability. J. Anat. 200, 629638 (2002).
  6. Aspelund, A., Robciuc, M.R., Karaman, S., Makinen, T. & Alitalo, K. Lymphatic system in cardiovascular medicine. Circ. Res. 118, 515530 (2016).
  7. Li, J., Zhou, J. & Shi, Y. Scanning electron microscopy of human cerebral meningeal stomata. Ann. Anat. 178, 259261 (1996).
  8. Földi, M. et al. Lymphogenic haemangiopathy. “Prelymphatic” pathways in the wall of cerebral and cervical blood vessels. Angiologica 5, 250262 (1968).
  9. Louveau, A. et al. Structural and functional features of central nervous system lymphatic vessels. Nature 523, 337341 (2015).
  10. Aspelund, A. et al. A dural lymphatic vascular system that drains brain interstitial fluid and macromolecules. J. Exp. Med. 212, 991999 (2015).
  11. Carmeliet, P. & Jain, R.K. Molecular mechanisms and clinical applications of angiogenesis. Nature 473, 298307 (2011).
  12. Geudens, I. & Gerhardt, H. Coordinating cell behaviour during blood vessel formation. Development 138, 45694583 (2011).
  13. Holley, J.E., Newcombe, J., Whatmore, J.L. & Gutowski, N.J. Increased blood vessel density and endothelial cell proliferation in multiple sclerosis cerebral white matter. Neurosci. Lett. 470, 6570 (2010).
  14. Hashimoto, T. et al. Evidence of increased endothelial cell turnover in brain arteriovenous malformations. Neurosurgery 49, 124132 (2001).
  15. Arenillas, J.F., Sobrino, T., Castillo, J. & Dávalos, A. The role of angiogenesis in damage and recovery from ischemic stroke. Curr. Treat. Options Cardiovasc. Med. 9, 205212 (2007).
  16. Hogan, B.M. et al. Vegfc/Flt4 signalling is suppressed by Dll4 in developing zebrafish intersegmental arteries. Development 136, 40014009 (2009).
  17. Le Guen, L. et al. Ccbe1 regulates Vegfc-mediated induction of Vegfr3 signaling during embryonic lymphangiogenesis. Development 141, 12391249 (2014).
  18. Villefranc, J.A. et al. A truncation allele in vascular endothelial growth factor c reveals distinct modes of signaling during lymphatic and vascular development. Development 140, 14971506 (2013).
  19. Astin, J.W. et al. Vegfd can compensate for loss of Vegfc in zebrafish facial lymphatic sprouting. Development 141, 26802690 (2014).
  20. Bower, N.I. et al. Vegfd modulates both angiogenesis and lymphangiogenesis during zebrafish embryonic development. Development 144, 507518 (2017).
  21. Mato, M. & Ookawara, S. Influences of age and vasopressin on the uptake capacity of fluorescent granular perithelial cells (FGP) of small cerebral vessels of the rat. Am. J. Anat. 162, 4553 (1981).
  22. Mato, M. et al. Involvement of specific macrophage-lineage cells surrounding arterioles in barrier and scavenger function in brain cortex. Proc. Natl. Acad. Sci. USA 93, 32693274 (1996).
  23. Okuda, K.S. et al. Lyve1 expression reveals novel lymphatic vessels and new mechanisms for lymphatic vessel development in zebrafish. Development 139, 23812391 (2012).
  24. Jin, S.-W., Beis, D., Mitchell, T., Chen, J.-N. & Stainier, D.Y.R. Cellular and molecular analyses of vascular tube and lumen formation in zebrafish. Development 132, 51995209 (2005).
  25. Akitake, C.M., Macurak, M., Halpern, M.E. & Goll, M.G. Transgenerational analysis of transcriptional silencing in zebrafish. Dev. Biol. 352, 191201 (2011).
  26. Ando, K. et al. Clarification of mural cell coverage of vascular endothelial cells by live imaging of zebrafish. Development 143, 13281339 (2016).
  27. Xu, J. et al. Temporal-spatial resolution fate mapping reveals distinct origins for embryonic and adult microglia in zebrafish. Dev. Cell 34, 632641 (2015).
  28. Ellett, F., Pase, L., Hayman, J.W., Andrianopoulos, A. & Lieschke, G.J. mpeg1 promoter transgenes direct macrophage-lineage expression in zebrafish. Blood 117, e49e56 (2011).
  29. Secker, G.A. & Harvey, N.L. VEGFR signaling during lymphatic vascular development: From progenitor cells to functional vessels. Dev. Dyn. 244, 323331 (2015).
  30. Shimabukuro, M.K. et al. Lipid-laden cells differentially distributed in the aging brain are functionally active and correspond to distinct phenotypes. Sci. Rep. 6, 23795 (2016).
  31. Koltowska, K. et al. Mafba is a downstream transcriptional effector of Vegfc signaling essential for embryonic lymphangiogenesis in zebrafish. Genes Dev. 29, 16181630 (2015).
  32. Liu, C. et al. Macrophages mediate the repair of brain vascular rupture through direct physical adhesion and mechanical traction. Immunity 44, 11621176 (2016).
  33. Lê Cao, K.A., Rohart, F., McHugh, L., Korn, O. & Wells, C.A. YuGene: a simple approach to scale gene expression data derived from different platforms for integrated analyses. Genomics 103, 239251 (2014).
  34. Lojewski, X. et al. Perivascular mesenchymal stem cells from the adult human brain harbor no instrinsic neuroectodermal but high mesodermal differentiation potential. Stem Cells Transl. Med. 4, 12231233 (2015).
  35. Armulik, A. et al. Pericytes regulate the blood-brain barrier. Nature 468, 557561 (2010).
  36. Zonta, M. et al. Neuron-to-astrocyte signaling is central to the dynamic control of brain microcirculation. Nat. Neurosci. 6, 4350 (2003).
  37. Siekmann, A.F., Standley, C., Fogarty, K.E., Wolfe, S.A. & Lawson, N.D. Chemokine signaling guides regional patterning of the first embryonic artery. Genes Dev. 23, 22722277 (2009).
  38. Covassin, L.D., Villefranc, J.A., Kacergis, M.C., Weinstein, B.M. & Lawson, N.D. Distinct genetic interactions between multiple Vegf receptors are required for development of different blood vessel types in zebrafish. Proc. Natl. Acad. Sci. USA 103, 65546559 (2006).
  39. Vanhollebeke, B. et al. Tip cell-specific requirement for an atypical Gpr124- and Reck-dependent Wnt/β-catenin pathway during brain angiogenesis. Elife http://dx.doi.org/10.7554/eLife.06489 (2015).
  40. Keck, P.J. et al. Vascular permeability factor, an endothelial cell mitogen related to PDGF. Science 246, 13091312 (1989).
  41. Carmeliet, P. et al. Abnormal blood vessel development and lethality in embryos lacking a single VEGF allele. Nature 380, 435439 (1996).
  42. Bahary, N. et al. Duplicate vegfa genes and orthologues of the KDR receptor tyrosine kinase family mediate vascular development in the zebrafish. Blood 110, 36273636 (2007).
  43. Klotz, L. et al. Cardiac lymphatics are heterogeneous in origin and respond to injury. Nature 522, 6267 (2015).
  44. Wigle, J.T. & Oliver, G. Prox1 function is required for the development of the murine lymphatic system. Cell 98, 769778 (1999).
  45. Stanczuk, L. et al. cKit lineage hemogenic endothelium-derived cells contribute to esenteric lymphatic vessels. Cell Rep. http://dx.doi.org/10.1016/j.celrep.2015.02.026 (2015).
  46. Martinez-Corral, I. et al. Nonvenous origin of dermal lymphatic vasculature. Circ. Res. https://doi.org/10.1161/CIRCRESAHA.116.306170 (2015).
  47. Nicenboim, J. et al. Lymphatic vessels arise from specialized angioblasts within a venous niche. Nature 522, 5661 (2015).
  48. Dissing-Olesen, L., Hong, S. & Stevens, B. New brain lymphatic vessels drain old concepts. EBioMedicine 2, 776777 (2015).
  49. Bakker, E.N. et al. Lymphatic clearance of the brain: perivascular, paravascular and significance for neurodegenerative diseases. Cell. Mol. Neurobiol. 36, 181194 (2016).
  50. Roman, B.L. et al. Disruption of acvrl1 increases endothelial cell number in zebrafish cranial vessels. Development 129, 30093019 (2002).
  51. Lawson, N.D. & Weinstein, B.M. In vivo imaging of embryonic vascular development using transgenic zebrafish. Dev. Biol. 248, 307318 (2002).
  52. Koltowska, K. et al. Vegfc regulates bipotential precursor division and Prox1 expression to promote lymphatic identity in zebrafish. Cell Rep. 13, 18281841 (2015).
  53. Hogan, B.M. et al. Ccbe1 is required for embryonic lymphangiogenesis and venous sprouting. Nat. Genet. 41, 396398 (2009).
  54. Dunworth, W.P. et al. Bone morphogenetic protein 2 signaling negatively modulates lymphatic development in vertebrate embryos. Circ. Res. 114, 5666 (2014).
  55. van Impel, A. et al. Divergence of zebrafish and mouse lymphatic cell fate specification pathways. Development 141, 12281238 (2014).
  56. Ng, A.N. et al. Formation of the digestive system in zebrafish: III. Intestinal epithelium morphogenesis. Dev. Biol. 286, 114135 (2005).
  57. Scott, E.K. et al. Targeting neural circuitry in zebrafish using GAL4 enhancer trapping. Nat. Methods 4, 323326 (2007).
  58. Hartley, J.L., Temple, G.F. & Brasch, M.A. DNA cloning using in vitro site-specific recombination. Genome Res. 10, 17881795 (2000).
  59. Dahlem, T.J. et al. Simple methods for generating and detecting locus-specific mutations induced with TALENs in the zebrafish genome. PLoS Genet. 8, e1002861 (2012).
  60. Coxam, B. et al. Pkd1 regulates lymphatic vascular morphogenesis during development. Cell Rep. 7, 623633 (2014).
  61. Picelli, S. et al. Full-length RNA-seq from single cells using Smart-seq2. Nat. Protoc. 9, 171181 (2014).
  62. Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 1521 (2013).
  63. Anders, S., Pyl, P.T. & Huber, W. HTSeq – A Python framework to work with high-throughput sequencing data. Bioinformatics 31, 166169 (2015).
  64. Love, M.I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 15, 550 (2014).
  65. Heinz, S. et al. Simple combinations of lineage-determining transcription factors prime cis-regulatory elements required for macrophage and B cell identities. Mol. Cell 38, 576589 (2010).
  66. Jantzen, S.G., Sutherland, B.J., Minkley, D.R. & Koop, B.F.G.O. GO Trimming: systematically reducing redundancy in large Gene Ontology datasets. BMC Res. Notes 4, 267 (2011).
  67. Liao, Y., Smyth, G.K. & Shi, W. The Subread aligner: fast, accurate and scalable read mapping by seed-and-vote. Nucleic Acids Res. 41, e108 (2013).
  68. McKinney, W. Data structures for statistical computing in Python. Proc. 9th Python Sci. Conf. 5156 (2010).
  69. Pedregosa, F. et al. Scikit-learn: machine learning in Python. J. Mach. Learn. Res. 12, 28252830 (2011).
  70. Hatta, K., Tsujii, H. & Omura, T. Cell tracking using a photoconvertible fluorescent protein. Nat. Protoc. 1, 960967 (2006).
  71. Oorschot, V.M., Sztal, T.E., Bryson-Richardson, R.J. & Ramm, G. Immuno correlative light and electron microscopy on Tokuyasu cryosections. Methods Cell Biol. 124, 241258 (2014).
  72. Slot, J.W. & Geuze, H.J. Cryosectioning and immunolabeling. Nat. Protoc. 2, 24802491 (2007).

Download references

Author information

Affiliations

  1. Division of Genomics of Development and Disease, Institute for Molecular Bioscience, The University of Queensland, St Lucia, Brisbane, Australia.

    • Neil I Bower,
    • Katarzyna Koltowska,
    • Cathy Pichol-Thievend,
    • Scott Paterson,
    • Anne K Lagendijk,
    • Weili Wang,
    • Stephen J Bent,
    • Sungmin Baek,
    • Maria Rondon-Galeano,
    • Cas Simons,
    • Mathias Francois &
    • Benjamin M Hogan
  2. Centre for Stem Cell Systems, Department of Anatomy and Neuroscience, Faculty of Medicine Dentistry and Health Sciences, The University of Melbourne, Victoria, Australia.

    • Isaac Virshup,
    • Daniel G Hurley &
    • Christine A Wells
  3. Australian Regenerative Medicine Institute, Monash University Clayton Campus, Clayton, Victoria, Australia.

    • Benjamin W Lindsey &
    • Jan Kaslin
  4. Department of Cell Biology, National Cerebral and Cardiovascular Center Research Institute, Suita, Osaka, Japan.

    • Naoki Mochizuki
  5. AMED-CREST, National Cerebral and Cardiovascular Center, Suita, Osaka, Japan.

    • Naoki Mochizuki

Contributions

N.I.B. conceived, designed, performed experiments, analyzed data and wrote the manuscript. K.K., C.P.-T., B.W.L., S.P., A.K.L., W.W., S.B., M.R.-G., N.M. and M.F. performed experiments and provided unpublished reagents. I.V., D.G.H., C.A.W., C.S. and S.J.B. performed computational experiments and data analysis. J.K. designed and performed experiments. B.M.H. conceived and designed experiments and wrote the manuscript.

Competing financial interests

The authors declare no competing financial interests.

Corresponding author

Correspondence to:

Author details

Supplementary information

Supplementary Figures

  1. Supplementary Figure 1: lyve1-expressing cells in the adult zebrafish meninx co-express the lymphatic marker genes prox1a and fli1a (541 KB)

    (a) Confocal images of adult zebrafish brain in Tg(-5.2lyve1b:DsRed)nz101; TgBAC(prox1a:KalTA4 uq3bh;10xUAS:Venus) animals showing cells that co-express the lymphatic markers prox1a and lyve1. Expression of prox1a is mosaic due to the Gal4 system. Full quantification of marker co-expression can be found in Supplementary Figure 1e. Scale bar represents 100μm.

    (b) Confocal images of adult zebrafish brain in Tg(-5.2lyve1b:DsRed)nz101; Tg(fli1a:EGFP)y1 showing co-expression (arrows) of the lymphatic markers fli1a and lyve1. Full quantification of marker co-expression can be found in Supplementary Figure 1f. Scale bar represents 100μm.

    (c) Quantification of muLEC density on the surface of the brain (meningeal surfaces n=6 adult brains. Error bars represent mean +/− sem, Statistical testing N/A.)

    (d) Quantification of the number of lyve1-positive meningeal cells (over the surface of the tectum, or adjacent to the central arteries (cTA)) compared with non-meningeal lyve1-positive cells from cross sections of the adult zebrafish brain. n=4 adult brain cross sections, p<0.0001, from two-tailed student t-test (t=8.926 df=6).

    (e) Quantification of co-expression of cells from Tg(prox1a:KalTA4uq3bh;10xUAS:Venus) and lyve1 (Tg(-5.2lyve1b:DsRed)nz101) and the macrophage marker mpeg1 (Tg(mpeg1:mcherry)). 215 cells from n=4 larvae (lyve1) and 255 cells from n=5 larvae (mpeg1) were scored from confocal Z-stack images. Error bars represent mean +/− sem, Statistical testing N/A.

    (f) Quantification of co-expression of cells in Tg(-5.2lyve1b:DsRed)nz101 and Tg(fli1a:EGFP)y1, TgBAC(pdgfrβ:EGFP)uq15bh, TgBAC(acta2:EGFP)uq17 and Tg(nkx2.2a:EGFP)vu16Tg. 100 cells from n=5 larvae were scored from confocal Z-stack images. Error bars represent mean +/− sem, Statistical testing N/A.

    (g) Quantification of the distance from the nucleus of muLECs to the center of the parenchyma (defined as the centre of the unlabeled space surrounded by a vascular loop) shows that muLECs are more closely associated with blood endothelial cells than the space between vessels. 77 muLEC nuclei from n=3 larvae, p<0.0001, from two-tailed student t-test (t=15.64 df=130). Error bars represent mean +/− sem.

    (h) Quantification of the distance from the nucleus of a given muLEC to the nearest vessel branch point, compared with the distance from the nucleus of a given muLEC to the midpoint between branches of the closest vessel. muLECs are quantitatively closer to vessel branch points than vessel midpoints. 77 branch and 77 midpoints points from n=3 larvae, p<0.0001, from two-tailed student t-test (t=4.74 df=151). Error bars represent mean +/− sem.

  2. Supplementary Figure 2: High-molecular-weight dye can be taken up by meningeal lymphatics but not by muLECs. (1,261 KB)

    (a) Dorsal confocal image of 7mm Tg(-5.2lyve1b:DsRed)nz101 zebrafish larvae following injection of FITC-labeled high molecular weight (2000kDa) dextran. Arrowheads indicate the presence of the dye in lymphatic vessels, but the dye is absent from muLECs. Images representative of data obtained from injection of n=6 larvae. Scale bar represents 100μm.

    (b) Magnification of boxed region in (a) showing muLECs do not absorb high MW dextran. Images representative of data obtained from injection of n=6 larvae. Scale bar represents 100μm.

    (c) Dorsal confocal image of a Tg(-5.2lyve1b:DsRed)nz101; Tg(fli1a:EGFP)y1 zebrafish larvae. Image representative of n=5 larvae. Scale bar represents 100μm.

    (d) Magnification of boxed region in (c) showing muLECs co-express lyve1 and fli1a (arrowheads). Image representative of n=6 larvae. Full quantification of marker co-expression can be found in Supplementary Figure 1f. Scale bar represents 100μm.

    (e) muLECs are closely associated with blood vessels in Tg(-5.2lyve1b:DsRed)nz101 larvae. Image representative of n=6 larvae. Scale bar represents 100μm.

    (f) Confocal image of a cross-section of adult (20mm stage, 3 month old) Tg(-5.2lyve1b:DsRed)nz101; TgBAC(pdgfrβ:EGFP)uq15bh brain showing pdgfrβ expressing pericytes are associated with blood vessels throughout the adult CNS. Scale bar represents 100μm. Full quantification of marker co-expression can be found in Supplementary Figure 1f. Image representative of data obtained from injection of n=3 male and female 20mm zebrafish adults.

  3. Supplementary Figure 3: muLECs do not express markers of neuronal and/or oligodendrocyte cells and smooth muscle cells. (965 KB)

    (a-d) Confocal images of (a,b) 7mm stage Tg(-5.2lyve1b:DsRed)nz101;Tg(nkx2.2a:EGFP)vu16Tg and (c,d) 7mm Tg(-5.2lyve1b:DsRed)nz101;TgBAC(acta2:EGFP)uq17 zebrafish larvae showing a lack of co-expression of these markers (quantified in Supplementary Figure 1f) corresponding to neuronal/oligodendrocyte and smooth muscle respectively. Smooth muscle cells are present around the dorsal aorta in the lateral view of the trunk (arrow in d). Scale bars represent 100μm.

  4. Supplementary Figure 4: Cells of the putative lymphatic loop express lymphatic markers and are separate from the blood vasculature. (941 KB)

    (a) Confocal images of a 5dpf Tg(-5.2lyve1b:DsRed)nz101; Tg(kdrl:EGFP)s843 zebrafish midbrain showing that cells of the lymphatic loop do not co-express lyve1 and the blood vessel marker kdrl. n=5 embryos. Full quantification of marker co-expression can be found in Supplementary Figure 1f.

    (b) Confocal images of lyve1/prox1a positive cells of the lymphatic loop at 5dpf in Tg(-5.2lyve1b:DsRed)nz101; Tg(prox1a:KalTA4uq3bh;10xUAS:Venus) zebrafish. n=3 embryos. Full quantification of marker co-expression can be found in Supplementary Figure 1f.

    (c) Confocal images of a 5dpf zebrafish head showing cells of the lymphatic loop co-express the lymphatic markers lyve1 and fli1a in Tg(-5.2lyve1b:DsRed)nz101; Tg(fli1a:EGFP) (arrowheads). n=5 embryos. Full quantification of marker co-expression can be found in Supplementary Figure 1f.

    (d) Confocal images of 5dpf Tg(-5.2lyve1b:DsRed)nz101 embryo injected with dextran into the blood vasculature. The lyve1 positive cells do not contain dextran demonstrating they are separate from the blood vasculature. Scale bars represent 100μm. Representative of analysis from n=4 embryos.

  5. Supplementary Figure 5: Kaede photoconversion demonstrates that the lymphatic loop gives rise to the meningeal muLEC population. (121 KB)

    (a-b) Confocal images of Tg(prox1a:KalTA4)uq3bh; Tg(10xUAS:Kaede)s1999t photoconverted at 10dpf and reimaged at 15dpf (a) and photoconverted at 20dpf and reimaged at 25dpf (b). Scale bars represent 50μm. For the 10-15 dpf stages n=34 cells from 7 larvae were examined and for the 20-25dpf stages n=29 cells from 3 larvae were analysed. Positive lineage tracing was confirmed in all cases. (c) Quantification of the average increase in photo-converted cell number over 5 days (proliferation rate, cell number 5 days after conversion/0 days after conversion) data combined for 5-20 dpf experiments, n=19 larvae overall. Error bars represent mean +/− sem, Statistical testing N/A.

  6. Supplementary Figure 6: Still images of a Tg(-5.2lyve1b:DsRed)nz101;Tg(fli1a:nEGFP)y7 zebrafish embryo from Supplementary Movie 2. (826 KB)

    (a-k) Images are representative of n= 5 movies.

    (a) Tilted view showing the labeled endothelial nucleus in the choroidal vascular plexus (arrowhead).

    (b,c) First cell (grey render, arrow) exits the choroidal plexus vessel and migrates dorsally along another pre-existing vessel.

    (d-f) Second cell (blue render, arrow) exits vessel and migrates dorsally along pre-exiting vessel.

    (g-i) First cell divides (arrowhead) and one daughter cell divides again (arrow).

    (i,j) Second cell (blue) divides (arrowhead) and continues to migrate (k, arrowhead). Scale bar represents 50μm.

  7. Supplementary Figure 7: muLEC sprouting is dependent upon Vegfc, Vegfd, Ccbe1 and Vegfr3 signaling. (1,256 KB)

    (a-d) Dorsal confocal images of 5dpf Tg(-5.2lyve1b:DsRed)nz101; Tg(fli1a:nEGFP)y7 embryos showing the lymphatic loop that forms in wildtype embryos (a, arrows, representative image based on analysis of n=7 embryos) is reduced in vegfchu5055 (b, arrows, n=6 embryos) and vegfduq7bh (c, arrows, n=6 embryos) single mutants, but is absent in vegfchu5055/vegfduq7bh double mutants (d, asterisk, n=9 embryos). Scale bars represent 100μm.

    (e) Quantification of the number of muLECs present in a single putative lymphatic loop at 5dpf in wildtype (n=7 embryos), vegfchu5055 (n=6 embryos), vegfduq7bh (n=6 embryos) single mutants and vegfchu5055/vegfduq7bh double mutants (n=9 embryos). Error bars represent mean +/− sem; **** p<0.0001, from one way ANOVA (F (3, 25) = 166.6).

    (f-i) At 5 dpf, the lyve1/prox1a positive cells in Tg(-5.2lyve1b:DsRed)nz101; Tg(prox1a:KalTA4, 4xUAS:uncTagRFP)nim5; Tg(10xUAS:Venus) have formed a lymphatic loop in wildtype embryos (f, n=6 embryos), however in vegfchu5055 (g, n=6 embryos) and vegfd (h, n=6 embryos) mutants, the development of the lymphatic loop is reduced (arrows). In vegfchu5055 / vegfduq7bh double mutant embryos (i, n=6 embryos) the lymphatic loop is completely absent (asterisks). Scale bars represent 100μm.

    (j-l) The lyve1 lymphatic loop structure (control, j) is absent in vegfr3 morphant (n=15 embryos) (k) and (l) ccbe1 morphant (n=15 embryos) embryos at 5dpf (asterisks).

    (m) Quantification of the number of muLECs present in a single putative lymphatic loop at 5dpf in uninjected controls (n=15 embryos), vegfr3 morphant (n=15 embryos) and ccbe1 morphant (n=15 embryos). **** p<0.0001, from two-tailed student t-test (t=33.62 df=28). Scale bars represent 100μm.

    (n,o) Dorsal confocal images from 24 mm Tg(-5.2lyve1b:DsRed)nz101 vegfchu5055 (n) and vegfduq7bh (o) mutants showing the presence of either ligand is sufficient for development of the muLECs. Representative image based on analysis of n=4 adult brains. Scale bars represent 200μm.

  8. Supplementary Figure 8: Mature muLECs display a distinctive ultrastructure and take up LDL. (1,287 KB)

    (a,b) Schematic diagram indicating the region used for immuno-electron microscopy.

    (c) Low magnification immuno-electron micrograph image showing overview of imaged region in Fig. 4a,b. Scale bar represents 1μm.

    (d) Higher magnification of image shown in (c). Scale bar represents 1μm. * = prominent vacuoles or inclusion bodies consistent with confocal imaging in Figure 1 and panel g below, L=Lumen, EC= endothelial cell, BM=basement membrane.

    (e) High magnification of muLEC cell showing positive immunostaining by anti-gfp antibody. Scale bar represents 1μm.

    (f) High molecular weight dextran is not observed in muLECs (asterisks) 3 hours post injection into the blood stream of Tg(-5.2lyve1b:DsRed)nz101 larvae. n=63 muLECs were analysed in detail by scoring confocal z-stacks from n= 3 larvae.

    (g) Alexa 488-labelled acetylated LDL is observed in the endothelial cells and the inclusion bodies (arrows) of 61 +/− 11 % of muLECs 3 hours post injection into the blood vasculature. Scale bars represent 20μm. n=72 muLECs were analysed in detail by scoring confocal z-stacks from 3 embryos.

    (h,i) Three hours post injection of high molecular weight dextran, there is no obvious vascular leakage in wildtype (h) or vegfchu5055 / vegfduq7bh double mutant larvae (i) which lack muLECs. Scale bar represents 50μm. n=3 larvae were analysed in detail by scoring confocal z-stacks.

  9. Supplementary Figure 9: muLECs can be FAC-sorted and separated from blood endothelial cells and pericytes. (318 KB)

    (a) Representative plots of FAC sorted cells from adult brains in unlabeled wildtype and Tg(-5.2lyve1b:DsRed)nz101 positive samples.

    (b) Representative plots of FAC sorted EGFP positive cells in unlabeled wildtype and Tg(-5.2lyve1b:DsRed)nz101; Tg(kdrl:EGFP)s843 positive samples.

    (c) Representative plots of FAC sorted EGFP positive cells in unlabeled wildtype and Tg(-5.2lyve1b:DsRed)nz101; TgBAC(pdgfrβ:EGFP)uq15bh positive samples.

    (d) Scree plot of variance (eigenvalue versus component number) and principle component comparisons from Fig. 5b shown in 2-dimensional plots (PC1 vs PC2, PC2 vs PC3 and PC1 vs PC3), clear separation of muLECs from control samples is observed.

    (e) Representative plots of FAC sorted EGFP positive cells in wildtype and Tg(pdgfrβBAC:EGFP)uq15bh; Tg(kdrl:mCherry)s843 positive samples.

  10. Supplementary Figure 10: vegfchu5055vegfduq7bh double-mutant larvae have reduced meningeal BEC nuclei compared to wild-type larvae. (848 KB)

    (a) Dorsal confocal image (a) and hyperstack (a’) of z-stacks from the Tg(-5.2lyve1b:DsRed)nz101;Tg(fli1a:nEGFP)y7 signal with darker shades representing z-stack slices closer to the objective, showing vegfchu5055/vegfduq7bh double mutant larvae (a, representative images of n=8 larvae analysed), which have reduced meningeal BEC number when compared to size matched (5.7mm) wildtype siblings (b, n=7 larvae). Scale bar represents 100μm.

  11. Supplementary Figure 11: Visual confirmation of muLEC ablation and mock ablation. (744 KB)

    (a) Single z-slices from a dorsal confocal image of a 10dpf Tg(-5.2lyve1b:DsRed)nz101;Tg(fli1a:nEGFP)y7 larvae before and after targeted ablation of muLECs. Ablation can be confirmed by direct observation of cell death, with “bubbles” which disperse post ablation. Scale bar represents 50μm. Representative images of n=27 larvae analysed.

    (b) Second example of single z-slices from a dorsal confocal image of 10dpf Tg(-5.2lyve1b:DsRed)nz101;Tg(fli1a:nEGFP)y7 larvae before and after ablation of muLECs. Ablation can be confirmed by direct observation of cell death. The analysis in Fig. 7 used the same approach to verify cell ablations for n=27 embryos in total.

    (c) Example of single z-slices from a dorsal confocal image of 10dpf Tg(-5.2lyve1b:DsRed)nz101;Tg(fli1a:nEGFP)y7 larvae before and after mock ablation in which cells adjacent to the muLECs are targeted. Ablation confirmed by direct observation and this approach was used in Fig. 7 for analysis of n=15 larvae. This is the same larvae as in Fig. 7g,h. Scale bar represents 50μm.

    (d) Second example of single z-slices from a dorsal confocal image of 10dpf Tg(-5.2lyve1b:DsRed)nz101;Tg(fli1a:nEGFP)y7 larvae before and after mock ablation in which cells adjacent to the muLECs are targeted. This is the same larvae as above and in Fig. 7g,h.

    (e and f) 10 dpf Tg(-5.2lyve1b:DsRed)nz101;Tg(fli1a:nEGFP)y7 larvae before (e) and after (f) mock ablation. Representative of n=15 larvae analysed. Scale bar represents 100μm.

  12. Supplementary Figure 12: Schematic summary of muLEC development and function (672 KB)

    a. Mural lymphatic endothelial cells (muLECs) emerge from the choroidal vascular plexus (CVP) at 54 hpf (a) and sprout to the periphery of the midbrain at 96 hpf (a’).

    b. The resulting lymphatic vascular loop in the midbrain (MB) begins to undergo a transition to a mesenchymal morphology between 5 (b’) and 10 dpf (b’’)

    c - d. muLECs (blue) are present on the midbrain, fore brain (FB) and hind brain (HB) in 7 mm larvae (c) and expand over the adult brain (d). c’ and d’ provide lateral views and indicate the meningeal lymphatics (green, c’) and muLEC location relative to blood vessels (red) (d’).

    e and f. Three dimensional (e) and cross sectional schematic representation of the blood vasculature and mural cells at the zebrafish meninx. muLECs (blue) secrete vascular growth factors and take up LDL from the bloodstream. PVM= perivascular macrophage.

Video

  1. Video 1: 3D reconstruction of Tg(prox1a:KalTA4 uq3bh;10xUAS:Venus) Z-stacks from Fig. 2a. (1.77 MB, Download)
    The meningeal lymphatic vessels and muLECs surround the outer curvature of the brain in a 7mm larvae. Representative of n=6 larvae analysed.
  2. Video 2: Time lapse imaging of muLECs sprouting from the choroidal vascular plexus. (19.08 MB, Download)
    Left side: Confocal time lapse imaging overview of a Tg(-5.2lyve1b:DsRed)nz101; Tg(fli1a:nEGFP)y7 zebrafish embryo from 30 hpf to 4 dpf. At approximately 54 hpf, cells migrated dorsally from a vein running parallel to the primary head sinus and subsequently give rise to the lymphatic loop observed at 4dpf. Representative of n=5 movies of individual embryos analysed.
    Right side: Zoomed and rendered version of left hand movie, showing confocal time lapse imaging of a Tg(-5.2lyve1b:DsRed)nz101; Tg(fli1a:nEGFP)y7 zebrafish embryo from 30 hpf to 4 dpf. Rendered nuclei (grey and blue) highlight cells sprouting at approximately 54 hpf, from a vein running parallel to the primary head sinus, subsequently dividing and up-regulating lyve1. Representative of n=5 movies of individual embryos analysed.
  3. Video 3: 3D reconstruction showing processing used in Imaris to confirm the correct identification of meningeal BECs as represented in Fig. 6 and Fig.7. (2.67 MB, Download)
    In this example, the deconvoluted z-stack was surface rendered (white) to detect the EGFP signal from Tg(-5.2lyve1b:DsRed)nz101; Tg(fli1a:nEGFP)y7 14 dpf mock ablated larvae. The EGFP signal corresponding to the meningeal BECs was rendered magenta on the left side and green on the right side. The white render corresponding to the total EGFP signal was removed to leave the meningeal BECs. The channel for DsRed corresponding to the muLECs was added (grey) followed by the surface render of the signal.

PDF files

  1. Supplementary Text and Figures (3,793 KB)

    Supplementary Figures 1–12

  2. Supplementary Methods Checklist (477 KB)

Additional data