Mice

Bmp10 mutant mice bearing conditional knockout (Bmp102loxP(2f)) or lacZ reporter (Bmp10lacZ(3f)) alleles were generated at Gachon University. Generation of Bmp9-knockout mice (Bmp9-KO) on a C57BL/6J background was previously described [19]. ROSA26CreER/+ (R26CreER/+) was utilized for inducible global gene deletion ([31]; JAX stock #008463). Tagln-Cre [32], where Cre recombinase is transiently expressed in cardiomyocytes at a mid-gestational period ([33]; JAX stock #017491), was used to delete Bmp10 in embryonic cardiomyocytes, and Scl-CreER mice were used for endothelial cell-specific deletion of Eng [34]. To induce Cre activity for deletion of the Bmp10 gene, tamoxifen (Sigma, T5648) was given to control (Bmp102f/2f), Bmp9-KO (Bmp9−/−;Bmp102f/2f), Bmp10-iKO (R26CreER/+;Bmp102f/2f), and Bmp9/10-dKO (R26CreER/+;Bmp9−/−;Bmp102f/2f) mice with following dosages and schedules: for postnatal day (P) 6 analysis and survival study, 50 μg tamoxifen dissolved in corn oil (Sigma, C8267) was intragastrically injected at P1, P2, P3, and P4; for neonatal AVM analysis, tamoxifen (75 μg/g body weight for control and Bmp10-iKO littermates, 35 μg/g body weight for Bmp9-KO and Bmp9/10-dKO littermates) was intragastrically injected at P2 and P3; for wound-induced AVM study and survival study at adult mice, tamoxifen (100 μg/g body weight) was intraperitoneally injected for three consecutive days. All mice were maintained in a C57BL/6J × 129Sv hybrid background to compare their phenotypes. We utilized CreER-positive Bmp10-iKO and CreER-negative littermate control mice derived from intercross of R26CreER/+;Bmp102f/2f with Bmp102f/2f. CreER-positive Bmp9/10-dKO (R26CreER/+;Bmp9−/−;Bmp102f/2f) and CreER-negative littermate R26+/+;Bmp9−/−;Bmp102f/2f (Bmp9-KO) mice were derived from the intercross of R26CreER/+;Bmp9−/−;Bmp102f/2f and Bmp9−/−;Bmp102f/2f mice. Neonatal Scl-CreER;Eng2f/2f and control littermates were given an intragastric injection of 50 μg tamoxifen at P1, P2, and P3, and the retinas were harvested at P7. Tamoxifen-injected CreER-negative littermates were used as controls. For generation of Bmp101f allele, tamoxifen treated R26CreER/+;Bmp102f/+ male mice were mated with WT mice. Among the offspring, a R26+/+;Bmp101f/+ mouse was selected. Transmission of Bmp10-1f allele and absence of the 2f allele were confirmed for more than eight generations. All housing, breeding and procedures were performed according to the NIH Guide for the Care and Use of Experimental Animals and approved by the Institutional Animal Care and Use Committees of Gachon University (LCDI-2017–0069 and LCDI-2018–0086), University of Florida (UF 20,141,417), and Barrow Neurological Institute (530).

Construction of conditional knockout vector for targeting Bmp10

A BAC DNA [129S7/AB2.2 library, bMQ438f03 (87,389,727–87,451,271)] containing mouse Bmp10 gene was obtained from Source BioScience (UK). A vector targeted Bmp10 was constructed by using a homologous recombination method [35]. The mouse Bmp10 gene contains two exons (Online Figure IA). An 18.6-kb genomic DNA fragment of Bmp10 was retrieved from the BAC DNA and inserted into pLMJ235 possessing the diphtheria toxin A fragment (DT-A) gene as a negative selection marker. A frt-SD/SA-IRES-LacZ-Neo-frt-loxP (LacZ/Neo) cassette including a reporter (LacZ) and a positive selection marker (Neo, neomycin-resistant gene) was inserted at 477-bp upstream of exon 2. Downstream loxP sequence was inserted at 234-bp downstream of exon 2.

Generation for Bmp10 conditional knockout mice

The linearized targeting vector (45 μg) was electroporated into 2 × 107 J1 embryonic stem (ES) cells. Approximately 300 G418-resistant colonies were randomly picked. Homologous recombination was screened by genomic Southern blot analyses using external 5′-probes (Online Figure IB) and tested the first loxP site by PCR analysis (Online Figure IC) using specific primers (Online Table I). Targeted ES cells were injected into blastocysts of the C57BL/6 (B6) strain. ES cell culture and blastocyst injections were performed by standard methods. Male chimeras were bred to B6 females to establish the Bmp10lacZ(3f) strain on a 129/B6 hybrid background. Germ-line transmission of the targeted allele was confirmed by genotyping (Online Figure IC). Bmp10lacZ/+ mice were crossed with FLP–deleter mice (ROSA26FLPe) to remove the LacZ/Neo cassette flanked by the frt sequences for generating the conditional Bmp102f/+ mice. Genotypes of wild-type, Bmp102f/+ and Bmp101f/+, alleles were analyzed by PCR (Online Figure ID and 1E) using the primer sets listed in Online Table I.

Whole-mount X-gal staining

Whole-mount X-gal staining was performed using littermate control and Bmp103f/+ mice carrying a lacZ reporter. Fresh tissues were fixed with fixation solution [1% formaldehyde, 0.2% glutaraldehyde, 2 mM MgCl2, 5 mM EGTA, and 0.02% NP-40 in PBS (pH7.4)] for 10 min with gentle agitation. After three times washing with PBS, the samples were incubated in X-gal staining solution [5 mM K3[Fe(CN)6], 5 mM K4[Fe(CN)6], 2 mM MgCl2, 0.01% Na-deoxycholate, 0.02% NP-40, and 0.75 mg/mL X-gal in PBS (pH7.4)] at 37 °C overnight with gentle agitation. After three times washing with PBS, samples were sequentially dehydrated using increasing concentration of methanol, cleared with a BABB organic solvent (benzyl alcohol/benzyl benzoate = 1:1; Fisher, AC148390010, AC105860010) and photographed under a stereomicroscope with a CCD camera (Leica). For the histological analysis, X-gal-stained samples were embedded in paraffin following hydration with ethanol and clearing with xylene. The sections with 7 μm thickness were counterstained with nuclear fast red (Vector laboratories).

Histology

Collected tissues were fixed in 4% paraformaldehyde, dehydrated with the serial dilutions of ethanol, cleared with xylene, and embedded in paraffin. Seven-μm-thick sections were rehydrated with ethanol and stained with hematoxylin and eosin.

Antibodies

For immunostaining: anti-CD31 (Biocare medical, CM303B; 1:200), anti-Ki67 (Thermo, RM9106S1; 1:200), anti-VE-Cadherin (BioLegend, 348,501; 1:100), anti-Erg1/2/3-Alexa Fluor 647 (Santa Cruz, sc-376293; 1:100), and anti-SMA-Cy3 (Sigma, C6198; 1:200) antibodies were used. For western blotting: anti-BMP10-GFD (R&D systems, MAB6038; 1:1000), anti-BMP10-Pro (R&D systems, AF3956; 1:1500), and anti-TNNI3 (abcam, ab56357; 1:3000) antibodies were used.

Immunofluorescence

The eyeballs of neonatal pups were enucleated and fixed in 4% paraformaldehyde for 50 min at room temperature. Retinas were dissected, blocked for 1 h in blocking buffer [0.25% Triton X-100, 0.01% Na-deoxycholate, 3% BSA in PBS (pH 7.4)], and then incubated in blocking buffer at 4 °C overnight with Isolectin B4 (IB4; Sigma, L2140) or primary antibodies as described in antibodies. The retinas were washed five times with PBST (0.25% Triton X-100 in PBS) and incubated with appropriate fluorescence-conjugated secondary antibodies (Invitrogen) in blocking buffer at RT for 2 h. The retinas were washed five times with PBST and mounted on microscope glass slides with ProLong™ Gold Antifade Mountant (Invitrogen, P36930). The images were acquired by fluorescence microscopy (Keyence, BZ-X710).

Morphometric analysis

Morphometric measurements of retinal vasculature were done using ImageJ software. The radial length of the retina vessel was measured as the distance from the optic disk to the outermost vessel in the vascular front. Two measurements were obtained from each retina and averaged. Vascular density was measured as CD31-positive vessel area in 500 μm × 500 μm fields located between an artery and a vein in each retina and presented as a percentage. The diameter of arteries and veins was measured at 500 μm from the optic disk. The number of branch points was measured manually in four 500 µm × 500 µm fields located between an artery and a vein in each retina and averaged. The number of tip cells was measured in 300 µm × 500 µm fields located at the peripheral front between an artery and a vein. Endothelial cell size in veins was quantified as total cell area with CD31 and VE-Cadherin-stained boundaries divided by the number of Erg1/2/3-positive endothelial cell nuclei. Endothelial cell number was calculated as the total number of Erg1/2/3-positive endothelial cell nuclei per 100 μm length of vessels.

Western blot analysis

Mouse tissues were homogenized and prepared in ice-cold lysis buffer [50 mM Tris–HCl (pH 7.4), 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, 0.5% SDS, 0.5% Na-deoxycholate, phosphatase inhibitor mixture (Sigma, P5726 and P0044), and complete protease inhibitor mixture (Thermo, 87786)]. The supernatants after centrifugation at 15,000 × g, 4 °C for 30 min were used for measuring the protein concentration by DC Protein Assay kit (Bio-rad, 5000116). Samples were separated using SDS–polyacrylamide gels and transferred onto nitrocellulose membranes. The membranes were blocked with 5% non-fat milk in TBST (Tris-buffered saline with 0.2% Tween-20) for 1 h, probed using primary antibodies as described in antibodies, and further incubated with appropriate HRP-conjugated secondary antibodies. The bands were visualized by ECL substrate (Pierce, PI34096), and band intensities were quantified with Image Lab (Bio-rad).

Reverse transcription-quantitative real-time polymerase chain reaction (RT-qPCR)

RNAs were purified from various tissues of mice using a NucleoSpin RNA kit (Macherey–Nagel) and then reverse-transcribed using SuperScript III (Invitrogen). Comparative ΔCT method was applied to identify relative changes in mRNA expression using SYBR Premix (Applied biosystems, 4,367,659) and ABI StepOne real-time system (Applied biosystems) with the corresponding primer sets: mBmp10 (#1: 5′-ACA TCA TCC GGA GCT TCA AGA ACG-3′, 5′-AAC CGC AGT TCA GCC ATG ACG-3′; #2: 5′-GCA GAT GAG GTC GAA CAT GA-3′, 5′-GGC CTG AAT AAT TGC GTG TT-3′; #3: 5′-GGA TCC ACC AGA GTA CAT GCT-3′, 5′-GTC CAC GCC ATC ATA CAT CA-3′), Actin (5′-CCT GAA CCC TAA GGC CAA CCG-3′, 5′-GCT CAT AGC TCT TCT CCA GGG-3′). The expression levels were normalized to Actin as indicated.

Enzyme-linked immunosorbent assay (ELISA)

BMP9 and BMP10 protein levels in plasma or serum were measured using BMP9/10 ELISA kits (R&D systems, DY5566, DY2926) according to the manufacturer’s instructions. Briefly, a 96-well microtiter plate was coated with capture antibody and incubated with blocking solution [1% bovine serum albumin in PBS (pH 7.4)]. 100 μl of samples was added to each well and incubated at 4 °C overnight. After washing the plate five times with PBST (0.05% Tween-20 in PBS), the biotinylated detection antibody was incubated for 2 h at RT. The color change was detected via HRP-conjugated streptoavidin antibody. Neonatal and adult samples were measured with 1/4 or 1/8 dilution and with 1 or 1/2 dilution, respectively.

Diffusible iodine-based contrast-enhanced computed tomography (diceCT)

P6 pups were fixed in 10% formalin for 2 days, washed with PBS, and incubated in Lugol’s iodine solution (1.25% iodine and 2.5% potassium iodine in distilled water; Carolina Biological Supply) for 6 days. The samples were scanned in a Phoenix v|tome|x M scanner (GE Measurement & control). The resulting 2D X-ray data were processed using datos|x software v2.3 and analyzed using VG StudioMax 3.0 (Volume Graphics).

Skin wound generation and blue latex dye perfusion assay

Six to 16-week-old male and female mice were deeply anesthetized by placing the animal within an induction chamber and introducing 4% isoflurane gas, and anesthesia was maintained by 2% isoflurane. One wound was inflicted on the mid-dorsum and the other on the untagged ear using a sterile 4 mm biopsy punch (Integra, 33–34). Betadine was applied to the wound. To delete conditional alleles of Bmp10, tamoxifen at 100 μg/g body weight was intraperitoneally injected for three consecutive days from the day of wounding (day 0). To examine arterial vasculature and AVMs, the perfusion of latex blue dye was performed at day 6 for Bmp9/10-dKO mice and at day 7 for control, Bmp9-KO and Bmp10-iKO mice. Latex dye (Connecticut valley biological supply company) was injected into the left ventricle with a 26-gauge syringe, after sequential perfusion with 10 ml of dilator (10 unit/ml heparin, 0.04 mg/ml papaverine, and 100 μM sodium nitroprusside) and 10 ml of 10% formalin. Injected mice were briefly washed in PBS, and hair on the back was removed. After overnight post-fixation in 10% formalin, the dorsal skin and ear were collected, dehydrated, and cleared with a BABB organic solvent. For neonates, blue latex dye was injected with a 27.5-gauge syringe after anesthesia with ice and perfusion with 4 ml of the dilator and 4 ml of 10% formalin. The images were acquired by a stereomicroscope with a CCD camera.

Embryo isolation and whole-mount staining

Pregnant mice were euthanized by cervical dislocation following anesthesia with isoflurane. Embryos were fixed in 4% paraformaldehyde at 4 °C overnight, washed with PBS, blocked for 1 h at room temperature in blocking buffer [0.3% BSA and 0.1% Tween-20 in PBS (pH 7.4)], and then incubated with primary antibodies in blocking buffer at 4 °C overnight. Embryos were washed five times with PBST (0.1% Tween-20 in PBS) and further incubated with appropriate fluorescence-conjugated secondary antibodies (Invitrogen) in blocking buffer at 4 °C overnight. Embryos were washed with PBST, dehydrated with serial dilution of methanol (25%, 50%, 75%, 90%, and 100% methanol in PBS), cleared with a BABB organic solvent, and photographed by fluorescence microscopy (Keyence, BZ-X710).

Complete blood count and hemoglobin level measurement

Blood was obtained from the tail vein of adults or superficial temporal vein of pups and collected into tubes containing EDTA (Sarstedt Inc, NC9414041; RAM Scientific, 07–6011; Greiner Bio-One, 450470). A complete blood count was carried out using Element HT5 (Heska). Hemoglobin level was examined using a hemoglobin photometer (Stanbio Laboratory, Hemopoint H2).

Tracer injection experiment

For Alexa Fluor 555 cadaverine tracer injection and detection in postnatal mice at P8, 40 μl of Alexa Fluor 555 cadaverine tracer (1 mg/ml, Invitrogen) was injected intraperitoneally. After 2 h, mice were anesthetized using ice and then perfused intracardially with 6 ml of the dilator and 4% paraformaldehyde. After dissection of eyes and brains, eyes were fixed in 4% paraformaldehyde for an additional 1 h at room temperature. Retinas were isolated and mounted on microscope glass slides with ProLong™ Gold Antifade Mountant. Brains were post-fixed at 4 °C overnight and skulls were removed. For tracer leakage detection, the images of whole brains and retinas were acquired using a fluorescence microscope (Keyence, BZ-X710). Leakage was quantified by measuring the fluorescence intensity of the brain and retina images with ImageJ.

Supplementation with BMP9 and BMP10

Bmp9/10-dKO mice were injected intragastrically with 50 μg tamoxifen dissolved in corn oil (Sigma, C8267) at P1, P2, P3, and P4. Control PBS, 100 ng of mouse BMP9 protein (R&D systems, 5566-BP), or 100 ng of mouse BMP10 protein (R&D systems, 6038-BP) was injected intraperitoneally and daily on the opposite side of the milk spot until the sample collection. Proteins were reconstituted in 4 mM HCl containing 0.1% BSA (R&D systems, Reconstitution buffer 04) according to the manufacturer’s instructions. 100 ng of protein was diluted with 4 μl PBS right before the injection and injected using a Hamilton syringe (Model 701 RN SYR). The analysis was performed at P8. Neonatal Scl-CreER;Eng2f/2f and control littermates were given intragastric injections of 50 μg tamoxifen at P1, P2, and P3, and the retinas were harvested at P7.

Statistical analysis

Data are expressed as mean ± standard error of the mean (SEM) or mean ± standard deviation of the mean (SD) with at least three independent experiments. Statistical analyses were performed using GraphPad Prism 8. The data distribution was assessed with D’Agostino & Pearson normality test or Shapiro–Wilk normality test. For normally distributed data, statistical comparisons between two groups and among multiple groups were performed by unpaired two-tailed student’s t test and either one-way or two-way analysis of variance (ANOVA) followed by Tukey’s post hoc test, respectively. For non-normally distributed data, statistical comparisons were performed by Kruskal–Wallis test followed by Dunn’s post hoc test. Survival curves were created by the method of Kaplan–Meier and were compared using the log-rank test. Assessments with P < 0.05 were considered statistically significant.