Mural cell-deficient larval brain vasculature displays abnormal patterning.

a–b Maximum intensity projections of mural cells (TgBAC(pdgfrb:EGFP)uq15bh) and brain vasculature (Tg(kdrl:Hsa.HRAS-mCherry)s916) in siblings and pdgfrb mutants at 7 (a) and 14 dpf (b). (a′) shows enlarged view of the white box in (a) highlighting the midbrain central arteries. c, Representative images of vascular tracing at 7 dpf using Imaris 10.1. Lumenized blood vessels branching from the middle mesencephalic central arteries in the midbrain were traced. d–e, Quantification of midbrain branching points (d) and vessel length (e). Data are mean±SEM, n=8 per group, unpaired t-tests. f, Maximum intensity projections of mural cells (TgBAC(pdgfrb:EGFP)uq15bh) and brain vasculature (Tg(kdrl:Hsa.HRAS-mCherry)s916) in siblings and pdgfrb mutants at 3-5 dpf. g, Quantification of branching points of midbrain central arteries demonstrating the first detectable vessel patterning phenotype at 5 dpf. Data are mean±SEM, n=10 per group at 3 dpf, n=15 sibling and 9 pdgfrb-/- at 4 dpf, n=15 sibling and 11 pdgfrb-/- at 5 dpf, two-way analysis of variance (ANOVA) followed by multiple comparisons with Tukey’s correction. a, a′, b, c, f, Dorsal views, anterior up, scale bars: 100 μm.

BBB integrity is maintained in pdgfrb mutants during larval stages.

a, c, e, Fluorescent tracer leakage assays in the midbrain of zebrafish larvae using 70 kDa Dextran–Fluorescein or 10 kDa Dextran–Alexa Fluor™ 647, visualised with vascular reporters (gray) Tg(kdrl:Hsa.HRAS-mCherry)s916 and Tg(kdrl:EGFP)s843. Siblings and pdgfrb mutants at 7 dpf (a) and 14 dpf (c) show no differences in tracer extravasation. Wild-type larvae at 3 and 7 dpf (e) serve as a positive control, illustrating reduced tracer leakage by 7 dpf. Dorsal views, anterior up, scale bars: 100 μm. b, d, f, Quantification of midbrain parenchymal 70- and 10-kDa dextran fluorescence intensity normalized to vascular kdrl reporter intensity. Data are mean±SEM. At 7 dpf (b): n=14 siblings and 8 pdgfrb-/- for 70 kDa, and n=14 siblings and 10 pdgfrb-/- for 10 kDa. At 14 dpf (d): n=9 siblings and 8 pdgfrb-/- for 70 kDa, and n=8 siblings and 9 pdgfrb-/- for 10 kDa. For the positive control at 3 and 7 dpf (f): n=12 (3 dpf) and 14 (7 dpf) for 70 kDa, and n=10 (3 dpf) and 14 (7 dpf) for 10 kDa. Unpaired t-tests for 10 kDa dextran comparisons at 3 vs 7 dpf and for sibling vs pdgfrb-/- comparisons at 7 and 14 dpf, and for 70-kDa dextran at 14 dpf. Mann–Whitney U tests for 70-kDa dextran comparisons at 3 vs 7 dpf and for sibling vs pdgfrb-/- comparisons at 7 dpf due to non-normal data distribution.

Adult pdgfrb mutants lack mural cells and display dilated arteries.

a, b, Confocal projections of whole brain imaging in siblings and pdgfrb mutants at 5 months of age, showing brain vasculature using Tg(kdrl:Hsa.HRAS-mCherry)s916 and mural cells using TgBAC(pdgfrb:EGFP)uq15bh (a) as a global marker and TgBAC(acta2:EGFP)uq17bh (b) as a vSMC marker. Dorsal views, anterior up, scale bars: 250 μm. a′, b′, Subset of z-slices from the whole brain imaging in (a) and (b) (white boxes) indicating mural cell loss and abnormal capillary network patterning. 100-μm-thick maximum intensity projections (MIP) were generated using the continuation of the left middle mesencephalic central artery (MMCtA, arrow) as an anatomical landmark. Scale bars: 250 μm. c, Confocal projections from whole brain imaging in sibling and pdgfrb mutants at 21 dpf. Siblings show acta2:EGFP-positive vSMC coverage along MMCtAs (yellow arrows). pdgfrb mutants have dilated MMCtAs with absent vSMCs. Scale bar: 100 μm. d–f, Quantification of capillary branching points (d), artery diameter (e) and capillary diameter (f) using the area shown in (a′). Data are mean±SEM, n=3 animals per group. Artery diameter was quantified by averaging 5 points per animal, and capillary diameter was quantified by measuring 10 capillary diameter per animal. Each animal marked with a different color in f. Unpaired t-tests for branching points and artery diameter, nested t-test for capillary diameter.

Adult pdgfrb mutants display aneurysms and vascular integrity defects.

a, Representative images of coronal midbrain sections obtained from 3-month-old sibling and pdgfrb mutants. b, Fluorescent tracer leakage assays in 3-month-old siblings and pdgfrb mutants (2 hpi). 10 kDa Dextran–Cascade Blue (b) was used to assess extravasation, and Tg(kdrl:Hsa.HRAS-mCherry)s916 was used to label the brain vasculature (merge in b′). Scale bar: 250 μm. c–c′, Enlarged regions from (a) and (b′) (dotted box) displaying blood and tracer accumulation (arrows) in pdgfrb mutants. Scale bar: 250 μm. d, Enlarged regions from (b) (dashed box) displaying vasculature and tracer leakage from medial (arterial region) to lateral (capillary region) regions. Area was divided into 10 regions (∼100 μm intervals) for quantification. d′, Enlarged view of yellow box in (d) showing capillary beds with intravascular tracer.Scale bar: 50 μm. e, Quantification of parenchymal 10 kDa tracer intensity (medial to lateral) normalized to vascular kdrl intensity. Data are mean±SEM, each point represents average normalized tracer intensity per brain region, n=5 sibling and 4 pdgfrb-/-, two-sample Kolmogorov–Smirnov test.

Adult and juvenile pdgfrb mutants display vessel rupture and tracer accumulation at aneurysm hotspots.

a–a′, Fluorescent tracer leakage assays in 5-month-old sibling and pdgfrb mutants. 70 kDa Dextran–Fluorescein was detected outside the vessels marked by Tg(kdrl:Hsa.HRAS-mCherry)s916 at aneurysm hotspots (arrowheads) shown in white box (magnified in a′). b–b′, Fluorescent tracer leakage assays in 1-month-old sibling and pdgfrb mutants. 10 kDa Dextran–Alexa Fluor™ 647 was detected outside the vessels marked by Tg(kdrl:EGFP)s843 at aneurysm hotspots (arrowhead) shown in white box (magnified in (b′)). a–b, Brains were collected at 2 hpi, dorsal views, anterior is up, scale bars: 200 μm. c, Quantification of hotspot leakage sites per animal in the midbrain region showing increased numbers in older animals. Data are mean±SEM, n=4 per group, unpaired t-test. d–e, High resolution imaging of artery (d) and capillary zones (e) in brain regions of 10-week-old adult pdgfrb crispants and uninjected siblings after tracer injection (2 hpi). 10 kDa Dextran–Alexa Fluor™ 647 injected, blood vessels (Tg(kdrl:EGFP)s843) and red blood cells (Tg(gata1:DsRed)sd2) are shown in MIPs and single Z-sections. Arrowheads indicate examples of hotspots. Capillary zones shown in MIPs (all channels) and single 10-kDa Dextran channel lack hotspots. Scale bar: 100 μm. f–g, Fluorescent tracer leakage assays in 1-month-old sibling and pdgfrb mutants shown in confocal projections of coronal midbrain sections collected at 0.5 hpi (f) or 6 hpi (g). 10 kDa Dextran–Alexa Fluor™ 647 was used to assess extravasation, and Tg(kdrl:EGFP)s843 was used to label the brain vasculature. The area was divided into 10 regions (∼70 μm intervals) for quantification (see Extended Data Fig. 6 for full section image). Arrowheads indicate examples of hotspots. Scale bar: 200 μm. h–i, Quantification of parenchymal 10-kDa dextran intensity (medial to lateral) normalized to vascular kdrl intensity. Data are mean±SEM, each point represents average normalized tracer intensity per brain region, n=3 sibling and 5 pdgfrb-/- for 0.5 hpi, and n=3 per group for 6 hpi, two-sample Kolmogorov–Smirnov test.

Adult pdgfrb mutants display structural endothelial cell defects and vessel rupture.

a–a′′, Transmission electron microscopy images of sectioned adult zebrafish brain vessels at 5 months of age. pdgfrb mutants show basement membrane thickening and breakdown (black arrow), serum accumulation outside the vessels (white arrow) (a′), increased abluminal endothelial caveolae (a′′). Pseudocolors shown are ECs (purple), basement membrane (cyan), caveolae (magenta) and vesicles larger than 100 nm (green). Scale bars: 2 μm. b, Magnified regions from (a) showing intact tight junctions (arrowheads) in both siblings and pdgfrb mutants. Scale bars: 1 μm. c, Quantification of endothelial caveolae. Caveolae were defined as uncoated spherical profiles <100 nm in diameter and scored as luminal or abluminal (see methods for details). Measurements were made for n=3 vessels for each group and a total of 8 different cellular regions measured for siblings and 19 for mutants. Data are mean±SEM and are not distributed normally, Mann–Whitney U test was applied. d, Quantification of basement membrane thickness, which was scored in 6 different regions per vessel with n=8 vessels in siblings 4 in pdgfrb mutants across n=3 animal per group. Data are mean±SEM, unpaired t-test. Data are mean±SEM and are not distributed normally, Mann–Whitney U test was applied.

uq30bh is a predicted null allele of pdgfrb causing brain pericyte deficiency.

a, Schematic of the CRISPR/Cas9-mediated mutation in the pdgfrb gene. The uq30bh allele carries a 39-bp deletion and a 5-bp insertion, resulting in a premature stop codon in the pdgfrb coding sequence, and is predicted to be a null allele. b, Confocal projections of pericytes (TgBAC(abcc9:abcc9-T2A-mCherry)um139) and brain vasculature (Tg(kdrl:Hsa.HRAS-mCherry)s916) in wild-type siblings and pdgfrbuq30bh mutants at 7 dpf. Dorsal view, anterior up, scale bar: 100 μm. c, Quantification of brain mural cell numbers using the data shown in Fig. 1a. pdgfrb:EGFP-positive cells associated with central arteries were counted. Data are mean±SEM, n=10 pdgfrb+/+, 18 pdgfrbuq30bh/+, 10 pdgfrbuq30bh/uq30bh, one-way ANOVA followed by multiple comparisons with Tukey’s correction. d, Body lengths measurements of developing pdgfrb mutants and siblings from tip of the head to tail base at larval to adult stages. Data are mean±SEM, n=13–17 pdgfrb+/+, 14–20 pdgfrbuq30bh/+, 7–14 pdgfrbuq30bh/uq30bh per timepoint, two-way ANOVA followed by multiple comparisons with Tukey’s correction. e, Kaplan–Meier survival curves of siblings (n=35) and pdgfrbuq30bh mutants (n=33), selected based on presence or absence of brain pericytes, using TgBAC(pdgfrb:EGFP)uq15bh. Log-rank (Mantel–Cox) test, p=0.0070. f, Gross anatomy of adult wild-type siblings and pdgfrbuq30bh mutants at 90 dpf. pdgfrbuq30bh mutants exhibit reduced overall body size and craniofacial defects consistent with previously described null mutant alleles.

Controls for size-dependent tracer extravasation and distribution at larval and adult stages.

a, Schematic of fluorescent tracer leakage assay workflow in larvae. Tracers (1–2000 kDa) were intravenously injected, and fish were live imaged at ∼2 hpi. b, Confocal projections of tracer leakage assays in midbrain at 7 dpf, showing negative correlation between extravasation and tracer molecular weight. Larvae were imaged with kdrl reporter (gray) and intravenously injected 1 kDa NHS Ester–Alexa Fluor™ 405, 10 kDa Dextran–Alexa Fluor™ 647, 70 kDa Dextran–Fluorescein, or 2000 kDa Dextran–Tetramethylrhodamine. Masked regions show the quantified region in the midbrain. The vasculature was masked (yellow) with Tg(kdrl:EGFP)s843 in the 1-, 10- and 2000-kDa tracer injections and with Tg(kdrl:Hsa.HRAS-mCherry)s916 in the 70-kDa tracer injection. The midbrain was masked (gray) using the tracer accumulation in the skin for outer lining and using the metencephalic vessels (arrowheads) for posterior border. Dorsal views, anterior up, scale bars: 100 μm. c, Quantification of midbrain parenchymal tracer fluorescence intensity normalized to vascular tracer fluorescence intensity. Data are mean±SEM, n=9 for 1 kDa, 12 for 10 kDa and 70 kDa, 13 for 2000 kDa, one-way ANOVA followed by multiple comparisons with Tukey’s correction, p<0.0001 for all undisplayed pairwise comparisons. d–e, Confocal projections of cleared juvenile zebrafish at 30 dpf showing retro-orbitally injected 2000 kDa Dextran–Tetramethylrhodamine diffusing across the vasculature. (e) shows a subset of z-slices from (d) focused on the mid-sagittal region of the brain, highlighting tracer distribution throughout the brain vasculature. Lateral view, anterior to left, scale bars: 500 μm.

Tracer assays using an independent normalization method confirm that BBB integrity is maintained in pdgfrb mutant larvae.

a, c, Tracer leakage assays in the midbrain of zebrafish larvae in siblings and pdgfrb mutants at 7 dpf (a) and 14 dpf (c). 70 kDa Dextran–Fluorescein or 10 kDa Dextran–Cascade Blue was co-injected with 2000 kDa Dextran–Tetramethylrhodamine to detect parenchymal extravasation and injection volume. Dorsal views, anterior up, scale bars: 100 μm. b, d, Quantification of midbrain parenchymal 70- and 10-kDa dextran fluorescence intensity normalized to vascular 2000-kDa dextran fluorescence intensity. Data are mean±SEM. At 7 dpf (b): n=15 siblings and 9 pdgfrb-/- for 70 kDa, and n=8 siblings and 9 pdgfrb-/- for 10 kDa. At 14 dpf (d): n=10 siblings and 9 pdgfrb-/- for 70 kDa, and n=6 siblings and 4 pdgfrb-/- for 10 kDa. Unpaired t-tests for 10 kDa at 7 and 14 dpf, 70 kDa at 14 dpf. Mann–Whitney U test for 70 kDa at 7 dpf due to non-normal distribution.

Development of acta2:EGFP-positive vSMCs in brain vasculature of siblings and pdgfrb mutants.

a–b′, Confocal projections of mural cells using TgBAC(pdgfrb:gal4FF)um140; Tg(5xUAS:RFP)nkuasrfp1a and TgBAC(acta2:EGFP)uq17bh, live imaged at 7 and 14 dpf (a–a′) or imaged at 21 dpf post-fixation and clearing (b–b′).acta2:EGFP colocalizes with RFP driven by pdgfrb:gal4FF along MMCtAs (arrows) and at the Circle of Willis (CoW) (dashed line) by 7 dpf, expanding to arterioles by 21 dpf but absent from capillaries. (a′) and (b′) are enlarged views of white boxes in (a) and (b), respectively. (a′′) shows the same fish reimaged at 14 dpf. c–d′, Confocal projections of developing smooth muscle cells (TgBAC(acta2:EGFP)uq17bh) and brain vasculature (Tg(kdrl:Hsa.HRAS-mCherry)s916) in siblings and pdgfrb mutants live imaged at 7 dpf (c–c′) or imaged at 21 dpf post-fixation and clearing (d). White boxes are magnified in (c′). a–d, Dorsal views, anterior up, scale bars: 100 μm (a–b′, d), 50 μm (c–c′).

Arterial aneurysms in juvenile pdgfrb mutants and additional control data.

a, Fluorescent tracer leakage assays in 2-month-old zebrafish midbrain (2 hpi). 70 kDa Dextran–Fluorescein and 2000 kDa Dextran–Tetramethylrhodamine were co-injected to detect parenchymal extravasation and vascular tracer, respectively. Arrowheads mark aneurysms. a′, Same brains reimaged at higher resolution targeting capillary regions. 30-μm-thick MIPs were displayed, starting 30 μm below the most superficial point of left hemisphere to exclude potential leakage from meningeal vessels. b, Examples of mild and severe aneurysms (arrowheads) in pdgfrb mutants. c, Quantification of the severity of aneurysms in sibling and pdgfrb mutants. Aneurysm phenotypes were qualitatively categorised based on severity; normal, mild and severe, n=11 per group. d, Quantification of parenchymal 70-kDa Dextran intensity normalized to vascular 2000 kDa Dextran intensity using the data in (a′). Data are mean±SEM, n=5 sibling and 4 pdgfrb-/-, unpaired t-test. e–f, Validation that crispants phenocopy pdgfrb mutants (supporting data in Fig. 5d–e). Maximum intensity projections of mural cells (TgBAC(pdgfrb:EGFP)uq15bh) and brain vasculature (Tg(kdrl:Hsa.HRAS-mCherry)s916) in uninjected sibling and pdgfrb crispants at larva (5 dpf) and adult (10-week-old) stage. pdgfrb crispants lack brain mural cells and show dilated arteries in adulthood, phenocopying pdgfrbuq30bh mutants. a–b, e, Dorsal views, anterior up. a–b, e–f, Scale bars: 200 μm.

Coronal midbrain sections corresponding to Fig. 5f-g.

a, b, Entire coronal midbrain sections from tracer leakage assays in 1-month-old sibling and pdgfrb mutants shown as confocal projections. Brains were collected after retro-orbital co-injection of 10 kDa Dextran–Alexa Fluor™ 647 and 2000 kDa Dextran–Tetramethylrhodamine at 0.5 hpi (a) or 6 hpi (b), and Tg(kdrl:EGFP)s843 was used to label vasculature. a′–a′′, b′–b′′, Enlarged views of regions indicated in (a) and (b). Panels (a′) and (b′) show medial regions (white boxes), whereas (a′′) and (b′′) show lateral regions (yellow boxes) in midbrain. Perfusion of tracer dyes and leakage hotspots are readily identifiable. Scale bars: 200 μm.

Adult pdgfrb mutants maintain normal endothelial ultrastructure in capillaries.

a, Transmission electron microscopy images of sectioned adult zebrafish brain vessels. b, Quantification of endothelial caveolae in capillaries of siblings and pdgfrb mutants. Measurements were made for n=11 vessels in siblings and 9 in pdgfrb mutants across n=3 animals per group. Data are mean±SEM and are not distributed normally, Mann–Whitney U test was applied. No consistent defects were observed in capillaries. Scale bars: 1 μm.