Artery gene expression of endothelial cells (ECs) in circle of Willis (CoW) arteries

(A-D) Confocal live images of Tg(flt4:yfp, kdrl:hras-mcherry)hu4881/s896 and scheme representation of CoW arteries in zebrafish brain at 32 hour post fertilization (hpf) (A), 54 hpf (B), 3 day post fertilization (dpf) (C), and 4 dpf (D). Green channel represents flt4:yfp fluorescence, Red channel represents kdrl:hras-mcherry fluorescence, and Merge panel combines both channels. Arrows point to the CoW arteries with kdrl:hras-mcherry signal. Scale bar = 50 μm

(E) Average intensity of kdrl:hras-mcherry in caudal division of internal carotid arteries (CaDI), basal communicating artery (BCA), and posterior communicating segments (PCS) at 32 hpf (n=6, 2 independent experiments), 54 hpf (n=12, 4 independent experiments), 3 dpf (n=15, 4 independent experiments), and 4 dpf (n=9, 2 independent experiments), two-way analyses of variance followed by Tukey’s multiple comparisons, represented with mean ± SD, * p≤0.05, ** p≤0.01

Abbreviations: hpf: hour post fertilization, dpf: day post fertilization, EC: endothelial cell, l-CaDI: left caudal division of internal carotid artery, r-CaDI: right caudal division of internal carotid artery, BCA: basal communicating artery, l-PCS: left posterior communicating segment, r-PCS: right posterior communicating segment

Vascular smooth muscle cell (VSMC) differentiation on circle of Willis (CoW) arteries

(A-D) Confocal live images of Tg(acta2:mcherry, kdrl:cerulean)ca8/sd24 TgBAC(pdgfrb:egfp)ncv22 and scheme representation of vascular endothelium and mural cells on CoW arteries in zebrafish brain at 32 hour post fertilization (hpf) (A), 54 hpf (B), 3 day post fertilization (dpf) (C), and 4 dpf (D). White channel represents kdrl:cerulean fluorescence, Red channel represents acta2:mcherry, Green channel represents pdgfrb:egfp, Merge 1 panel combines all three channels, Merge 2 combines acta2:mcherry in red and pdgfrb:egfp in green. Arrows point to the CoW arteries with pdgfrb:egfp and acta2:mcherry signal. Scale bar = 50 μm

(E) Number of pdgfrb+ vascular mural cell progenitors per 100 μm vessel length on caudal division of internal carotid arteries (CaDI), basal communicating artery (BCA), and posterior communicating segments (PCS) at 32 hpf (n=3, 1 independent experiment), 54 hpf (n=6, 1 independent experiment), 3 dpf (n=6, 1 independent experiment), and 4 dpf (n=5, 1 independent experiment), two-way analyses of variance followed by Tukey’s multiple comparisons, represented with mean ± SD, ** p≤0.01, *** p≤0.001, **** p≤0.0001

(F) Number of acta2+ VSMCs per 100 μm vessel length on CaDI, BCA, and PCS at 32 hpf (n=5, 1 independent experiment), 54 hpf (n=6, 2 independent experiments), 3 dpf (n=24, 6 independent experiments), and 4 dpf (n=25, 6 independent experiments), two-way analyses of variance followed by Tukey’s multiple comparisons, represented with mean ± SD, **** p≤0.0001

Abbreviations: hpf: hour post fertilization, dpf: day post fertilization, EC: endothelial cell, VSMC: vascular smooth muscle cell, l-CaDI: left caudal division of internal carotid artery, r-CaDI: right caudal division of internal carotid artery, BCA: basal communicating artery, l-PCS: left posterior communicating segment, r-PCS: right posterior communicating segment

Computational fluid dynamic (CFD) simulation of circle of Willis (CoW) arteries

(A-C) Confocal live images of Tg(kdrl:gfp)zn1 injected with dextran and representation of vascular diameter, CFD simulated flow and wall shear stress (WSS) within the CoW arteries at 54 hour post fertilization (hpf) (A), 3 day post fertilization (dpf) (B), and 4 dpf

(C). Arrows point to the CoW arteries with high flow velocity and WSS. Scale bar = 50 μm

(D) Average vascular diameter in caudal division of internal carotid arteries (CaDI), basal communicating artery (BCA), and posterior communicating segments (PCS) at 54 hpf (n=6, 1 independent experiment), 3 dpf (n=6, 1 independent experiment), and 4 dpf (n=6, 1 independent experiment), two-way analyses of variance followed by Tukey’s multiple comparisons, *** p≤0.001

(E-G) Red blood cell (RBC) velocity in CaDI, BCA, and PCS at 54 hpf (n=4, 1 independent experiment) (E), 3 dpf (n=5, 1 independent experiment) (F), and 4 dpf (n=4, 1 independent experiment) (G), ordinary one-way ANOVA analyses with Tukey’s multiple comparisons, represented with mean ± SD, * p≤0.05, ** p≤0.01

(H) Average wall shear stress (WSS) throughout the CoW arteries at 54 hpf (n=3, 1 independent experiment), 3 dpf (n=3, 1 independent experiment), and 4 dpf (n=3, 1 independent experiment), ordinary one-way ANOVA analyses with Tukey’s multiple comparisons, **** p≤0.0001

(I-K) Average wall shear stress (WSS) in CaDI, BCA, and PCS at 54 hpf (n=3, 1 independent experiment) (I), 3 dpf (n=3, 1 independent experiment) (J), and 4 dpf (n=3, 1 independent experiment) (K), ordinary one-way ANOVA analyses with Tukey’s multiple comparisons, represented with mean ± SD, * p≤0.05, ** p≤0.01, *** p≤0.001

Abbreviations: hpf: hour post fertilization, dpf: day post fertilization, RBC: red blood cell, WSS: wall shear stress, l-CaDI: left caudal division of internal carotid artery, r-CaDI: right caudal division of internal carotid artery, BCA: basal communicating artery, l-PCS: left posterior communicating segment, r-PCS: right posterior communicating segment

Blood flow is required for vascular smooth muscle cell (VSMC) differentiation on circle of Willis (CoW) arteries

(A) Scheme representation of in vitro cell co-culture experiment

(B) Representative immunofluorescence images of brain pericytes after exposure of pulsative flow and laminar flow. Cells were stained for ACTA2 (cyan), and cytosolic GFP label (magenta). Scale bar = 10 μm

(C) Morphological measurement of brain pericyte length/width ratio before and after exposure of pulsative flow and laminar flow (n = 63 cells), two-tailed Mann–Whitney test, represented with mean ± SD, **** p≤0.0001

(D) Protein level of PDGFRB (n = 153 cells), ACTA2 (n = 222 cells), and TAGLN (n = 150 cells) in arbitrary unit (A.U.) after exposure of pulsative flow and laminar flow, two-tailed Mann–Whitney test, represented with mean ± SD, **** p≤0.0001

(E-G) Confocal live images of Tg(acta2:mcherry; kdrl:gfp)ca8/zn1 and scheme representation of vascular endothelium and VSMCs on CoW arteries at 3 day post fertilization (dpf) and 4 dpf in control embryos (E), embryos injected with 0.35 ng tnnt2a morpholino (MO) (F), and embryos treated with 25 μM nifedipine from 54 hpf (G). Red channel represents acta2:mcherry, Green channel represents kdrl:gfp, and Merge panel combines both channels. Arrows point to the CoW arteries with acta2:mcherry signal. Scale bar = 50 μm

(H) Number of acta2+ VSMCs per 100 μm vessel length on caudal division of internal carotid arteries (CaDI), basal communicating artery (BCA), and posterior communicating segments (PCS) at 3 dpf in uninjected control (n=6, 1 independent experiment) and embryos injected with 0.35 ng tnnt2a MO at one to two cell stage (n=6, 1 independent experiment), two-tailed Mann–Whitney test on each vessel’s comparison, represented with mean ± SD, *** p≤0.001, **** p≤0.0001

(I) Number of acta2+ VSMCs per 100 μm vessel length on CaDI, BCA, and PCS at 4 dpf in uninjected control (n=6, 1 independent experiment) and embryos injected with 0.35 ng tnnt2a MO at one to two cell stage (n=6, 1 independent experiment), two-tailed Mann–Whitney test on each vessel’s comparison, represented with mean ± SD, **** p≤0.0001

(I) Number of acta2+ VSMCs per 100 μm vessel length on CaDI, BCA, and PCS at 3 dpf in DMSO control (n=8, 2 independent experiments) and embryos treated with 25 μM nifedipine from 54 hpf (n=8, 2 independent experiments), two-tailed Mann–Whitney test on each vessel’s comparison, represented with mean ± SD, ** p≤0.01, **** p≤0.0001

(J) Number of acta2+ VSMCs per 100 μm vessel length on CaDI, BCA, and PCS at 4 dpf in DMSO control (n=9, 2 independent experiments) and embryos treated with 25 μM nifedipine from 54 hpf (n=8, 2 independent experiments), two-tailed Mann–Whitney test on each vessel’s comparison, represented with mean ± SD, ** p≤0.01, **** p≤0.0001

Abbreviations: hpf: hour post fertilization, dpf: day post fertilization, EC: endothelial cell, VSMC: vascular smooth muscle cell, MO: morpholino, l-CaDI: left caudal division of internal carotid artery, r-CaDI: right caudal division of internal carotid artery, BCA: basal communicating artery, l-PCS: left posterior communicating segment, r-PCS: right posterior communicating segment

Blood flow regulated transcription factor klf2a is expressed in circle of Willis (CoW) arteries

(A) Confocal live images of Tg(klf2a:h2b-gfp, kdrl:ras-mcherry)ig11/s896 and scheme representation of endothelial cells (ECs) in CoW arteries in zebrafish brain at 32 hour post fertilization (hpf), 54 hpf, 3 day post fertilization (dpf), and 4 dpf. Green channel represents klf2a:h2b-gfp, Red channel represents kdrl:ras-mcherry, and Merge panel combines both channels. Arrows point to the CoW arteries with klf2a:h2b-gfp signal. Scale bar = 50 μm

(B-E) Number of klf2a+ ECs per 100 μm vessel length on caudal division of internal carotid arteries (CaDI), basal communicating artery (BCA), and posterior communicating segments (PCS) at 32 hpf (n=6, 3 independent experiments) (B), 54 hpf (n=6, 2 independent experiments) (C), 3 dpf (n=11, 3 independent experiments) (D), and 4 dpf (n=12, 3 independent experiments) (E), ordinary one-way ANOVA analyses with Tukey’s multiple comparisons, represented with mean ± SD, ** p≤0.01, **** p≤0.0001

Abbreviations: hpf: hour post fertilization, dpf: day post fertilization, EC: endothelial cell, l-CaDI: left caudal division of internal carotid artery, r-CaDI: right caudal division of internal carotid artery, BCA: basal communicating artery, l-PCS: left posterior communicating segment, r-PCS: right posterior communicating segment

klf2a promotes vascular smooth muscle cell (VSMC) differentiation on anterior circle of Willis (CoW) arteries

(A-B) Confocal live images of Tg(acta2:mcherry; kdrl:gfp)ca8/zn1 and scheme representation of vascular endothelium and VSMCs on CoW arteries at 3 day post fertilization (dpf) (A) and 4 dpf (B) in uninjected control embryos, embryos injected with 11 ng klf2a morpholino (MO). Red channel represents acta2:mcherry, Green channel represents kdrl:gfp, and Merge panel combines both channels. Arrows point to the CoW arteries with acta2:mcherry signal. Scale bar = 50 μm

(C) Number of acta2+ VSMCs per 100 μm vessel length on caudal division of internal carotid arteries (CaDI), basal communicating artery (BCA), and posterior communicating segments (PCS) at 3 dpf in uninjected control (n=8, 2 independent experiments) and embryos injected with 11 ng klf2a MO at one to two cell stage (n=8, 2 independent experiments), two-tailed Mann–Whitney test on each vessel’s comparison, represented with mean ± SD, ** p≤0.01

(D) Number of acta2+ VSMCs per 100 μm vessel length on CaDI, BCA, and PCS at 4 dpf in uninjected control (n=8, 3 independent experiments) and embryos injected with 11 ng klf2a MO at one to two cell stage (n=8, 3 independent experiments), two-tailed Mann–Whitney test on each vessel’s comparison, represented with mean ± SD

Abbreviations: hpf: hour post fertilization, dpf: day post fertilization, EC: endothelial cell, l-CaDI: left caudal division of internal carotid artery, r-CaDI: right caudal division of internal carotid artery, BCA: basal communicating artery, l-PCS: left posterior communicating segment, r-PCS: right posterior communicating segment

Schematic model of the developmental muscularization of the CoW arteries

The model shows how blood flow generate higher hemodynamics in anterior CoW arteries like the caudal division of internal carotid artery (CaDI) via the activation of endothelial klf2a signaling. Other posterior CoW arteries with straight shape like the posterior communicating segment (PCS) experience less hemodynamic force and showed moderate klf2a activation and VSMC differentiation.

Abbreviations: hpf: hour post fertilization, dpf: day post fertilization, EC: endothelial cell, VSMC: vascular smooth muscle cell, CaDI: caudal division of internal carotid artery, PCS: posterior communicating segment

List of zebrafish fluorescent transgenic lines used in the study

List of morpholino antisense oligonucleotides used in the study

Vascular smooth muscle cell (VSMC) differentiation on circle of Willis (CoW) arteries

(A-B) Confocal live images of Tg(acta2:mcherry, kdrl:gfp)ca8/zn1 in CoW arteries of zebrafish brain at 32 hour post fertilization (hpf), 54 hpf, 3 day post fertilization (dpf), and 4 dpf. Red channel represents acta2:mcherry, Green channel represents kdrl:gfp, and Merge panel combines both channels. Scale bar = 50 μm

(C) Linear regression with normalized # acta2+ cells in CaDI (left y-axis) and average wall shear stress (WSS) (right y-axis) with the developmental stages (54 hpf, 3 and 4 dpf, x-axis), and Pearson correlation coefficient (r) is provided, indicating the degree of linear relationship between two variables.

Abbreviations: hpf: hour post fertilization, dpf: day post fertilization, VSMC: vascular smooth muscle cell, l-CaDI: left caudal division of internal carotid artery, r-CaDI: right caudal division of internal carotid artery, BCA: basal communicating artery, l-PCS: left posterior communicating segment, r-PCS: right posterior communicating segment

Blood flow is not required for vascular smooth muscle cell (VSMC) maintenance or pdgfrb+ mural cell progenitor recruitment on circle of Willis (CoW) arteries

(A-B) Bright-field images of Tg(acta2:mcherry; kdrl:gfp)ca8/zn1 in control embryos treated with DMSO and embryos treated with 25 μM nifedipine from 54 hour post fertilization (hpf) at 3 day post fertilization (dpf) (A) and 4 dpf (B). Scale bar = 500 μm

(C) Quantitative heartbeat rate at 3 dpf and 4 dpf in DMSO control embryos and embryos treated with 25 μM nifedipine from 54 hpf (n=5 each group, 1 independent experiment), two-tailed Mann–Whitney test on each vessel’s comparison, represented with mean ± SD, **** p≤0.0001

(D-E) Confocal live images of Tg(acta2:mcherry; kdrl:gfp)ca8/zn1 in CoW arteries of zebrafish brain at 5 dpf in DMSO control embryos and embryos treated with 20 μM nifedipine from 4 dpf. Red channel represents acta2:mcherry, Green channel represents kdrl:gfp, and Merge panel combines both channels. Scale bar = 50 μm

(F) Number of acta2+ VSMCs per 100 μm vessel length on caudal division of internal carotid arteries (CaDI), basal communicating artery (BCA), and posterior communicating segments (PCS) at 5 dpf in DMSO control (n=5, 1 independent experiment) and embryos treated with 20 μM nifedipine from 4 dpf (n=4, 1 independent experiment), two-tailed Mann–Whitney test on each vessel’s comparison, represented with mean ± SD, * p≤0.05

(G) Confocal live images of Tg(pdgfrb:egfp, kdrl:ras-mcherry)ncv22/s896 labeling vascular endothelium and pdgfrb+ vascular mural cell progenitors on the CoW arteries in zebrafish brain at 3 dpf. Red channel represents kdrl:ras-mcherry, Green channel represents pdgfrb:egfp. Scale bar = 50 μm

(H) Number of pdgfrb+ vascular mural cell progenitors per 100 μm vessel length on CaDI, BCA, and PCS in control embryos and embryos injected with 0.35 ng tnnt2a morpholino (MO) at 3 dpf (n=6, 1 independent experiment), two-way analyses of variance followed by Tukey’s multiple comparisons, represented with mean ± SD.

Abbreviations: hpf: hour post fertilization, dpf: day post fertilization, VSMC: vascular smooth muscle cell, l-CaDI: left caudal division of internal carotid artery, r-CaDI: right caudal division of internal carotid artery, BCA: basal communicating artery, l-PCS: left posterior communicating segment, r-PCS: right posterior communicating segment

Number of endothelial cells (ECs) in circle of Willis (CoW) arteries does not increase during klf2a activation

(A-B) Confocal live images of Tg(fli1:nls-gfp, kdrl:ras-mcherry)y7/s896 in CoW arteries of zebrafish brain at 32 hour post fertilization (hpf), 54 hpf, 3 day post fertilization (dpf), and 4 dpf. Red channel represents kdrl:ras-mcherry and Green channel represents fli1:nls-gfp. Scale bar = 50 μm

(C-F) Number of ECs per 100 μm vessel length on caudal division of internal carotid arteries (CaDI), basal communicating artery (BCA), and posterior communicating segments (PCS) at 32 hpf (n=9, 1 independent experiment) (C), 54 hpf (n=8, 1 independent experiment) (D), 3 dpf (n=8, 1 independent experiment) (E), and 4 dpf (n=9, 1 independent experiment) (F), ordinary one-way ANOVA analyses with Tukey’s multiple comparisons, represented with mean ± SD, * p≤0.05, *** p≤0.001, **** p≤0.0001

(G) Confocal live images of Tg(klf2a:h2b-gfp, kdrl:ras-mcherry)ig11/s896 in CoW arteries of zebrafish brain at 3 and 4 dpf in uninjected control and embryos injected with 11 ng of klf2a morpholino (MO). Red channel represents kdrl:ras-mcherry and Green channel represents klf2a:h2b-gfp. Scale bar = 50 μm

(H) Average intensity of klf2a:h2b-gfp in CaDI, BCA, and PCS at 3 dpf in uninjected control (n=3, 1 independent experiment) and embryos injected with 11 ng of klf2a MO (n=3, 1 independent experiment), two-tailed Mann–Whitney test on each vessel’s comparison, represented with mean ± SD, * p≤0.05, *** p≤0.001, **** p≤0.0001

(I) Average intensity of klf2a:h2b-gfp in CaDI, BCA, and PCS at 4 dpf in uninjected control (n=3, 1 independent experiment) and embryos injected with 11 ng of klf2a MO (n=3, 1 independent experiment), two-tailed Mann–Whitney test on each vessel’s comparison, represented with mean ± SD, * p≤0.05, **** p≤0.0001

(J) Linear regression with normalized # klf2a+ cells at 3 dpf (left y-axis) and normalized # acta2+ cells (right y-axis) with the different CoW arteries (CaDI, BCA and PCS, x-axis), and Pearson correlation coefficient (r) is provided, indicating the degree of linear relationship between two variables.

Abbreviations: hpf: hour post fertilization, dpf: day post fertilization, l-CaDI: left caudal division of internal carotid artery, r-CaDI: right caudal division of internal carotid artery, BCA: basal communicating artery, l-PCS: left posterior communicating segment, r-PCS: right posterior communicating segment