Human and mouse brain-derived endothelial cells require high levels of growth factors medium for their isolation, in vitro maintenance and survival

The experimental protocol was approved by the ethics committee of the IRCCS Foundation Neurological Institute “C. Besta” (Milan, Italy) and IRCCS Foundation Ospedale Maggiore Policlinico, Mangiagalli and Regina Elena (Milan, Italy). Human fetal brains were obtained from five 10-12 week-old healthy foetuses, according to the ethical guidelines of the European Network for Transplantation (NECTAR). Tissue specimens were stored in medium DMEM/F12 (Gibco, Grand Island, NY) containing, penicillin 100 U/ml at 4°C for less than 48 h prior to processing.

Procedures involving animals and their care were conducted in conformity with all procedures following institutional guidelines which, in turn, are in compliance with national (D.L. No. 116, G.U. Suppl. 40, Feb. 18, 1992, Circolare No. 8, G.U., 14 Luglio 1994) and international laws and policies (EEC Council Directive 86/609, OJ L 358, 1 Dec.12, 1987; NIH Guide for the Care and use of Laboratory Animals, U.S. National Research Council, 1996). Brains from male CD1 mice (Charles River, Calco, LC, Italy) were removed and stored in DMEM/F12 plus antibiotics until used. After several washes with PBS/antibiotics, tissues were finely minced using surgical scissors and then incubated with Liberase Blendzyme 2 (Roche, Mannheim, Germany) at a concentration of 0.625 Wu/ml at 37°C on a rotator for 2 hours. After enzymatic digestion, the cell suspension was washed with D-PBS, centrifuged at 1200 rpm/10 min and then plated in 25 cm²-flask (one for each brain) coated with collagen type I (BD Bioscience, San Diego, CA, USA), in EndoPM, specifically developed by our laboratory (pending patent medium MI2011A 000201) to select and expand ECs. This medium contains very high levels of growth factors (Bovine Brain Extract (6 μg/mL), Fibroblast Growth Factor (5 ng/mL), and Epidermal Growth Factor (10 ng/mL) in particular) and a blend of hormones called Hormone Mix, composed by by apotransferrin (48.82 μg/mL), selenium (2.37 ng/mL), progesterone (2.88 ng/mL), putrescine (48.25 μg/mL), insulin (11.5 μg/mL). Cells were maintained at 37°C, 5% CO₂. After 24 h of culture, non adherent cells were removed from the flasks and reseeded in a new collagen type I -coated culture 25 cm²-flask in 6 mL of fresh EndoPM medium.

The medium was changed every 10 days, the cultured cells were passaged at a split ratio of 1:4 every 14 days and detached by TrypLE Select (Gibco). BMVECs at passages P10-P15 were used as indicated for all experiments. Experiments were performed in triplicates.

Human and Mouse Microvascular Endothelial Cells (HMVECs), cultured in complete EBM medium with growth factors supplement kit (Lonza Group Ltd, Basel Switzerland), were utilized as control samples.

All experiments were done on 12-well plates, in triplicate for each treatment. Cells were dissociated with TrypLE (Gibco) and diluted to 5 × 10⁵ cells/mL. 5 × 10⁵ cells were added into each well and placed into 5% CO₂, 37°C incubator. EndoPM or EBM were added to the specified well. At different time points, cells were dissociated with TrypLE and counted by cytometry. For proliferation experiments, media were changed weekly until the day of analysis.

Proliferation index represents the cell number at a specific time point divided by the number of input cells at time 0; survival index represents the number of harvested live cells 12 hours after plating in EndoPM or EBM (Table 1).

Human and mouse BMVECs were incubated in growth media at 37°C 5% CO₂ in a humidified atmosphere, in a 6-wells culture plate, 3 × 10⁵ cells per well. When 80% confluence was reached, cells were washed twice in PBS and incubated overnight in starvation medium composed of DMEM/F12 (Gibco), 1% L-Glutamine (Gibco), 0.5% FBS, to induce cells synchronization.

The day after, starved cells were treated with complete EndoPM or EBM medium for 5h or overnight. Cells were than detached with TripLE (Gibco) following manufacturer instruction and processed for cell cycle analysis.

Cell cycle was evaluated by flow cytomety analysis after synchronization in starvation medium. After the synchronization time, starved cells were treated with complete EndoPM or EBM medium for 5 h or overnight. 5 × 10⁵ human or mouse BMVECs cells were fixed in 1 ml cold ethanol (70% vol/vol in PBS) with gentle vortexing (about 5 s) to obtain a mono-dispersed cell suspension, and maintained in ethanol, at 4°C, for at least 2 h. After ethanol fixation, cells were centrifuge 1200 rpm/10 min and the ethanol thoroughly decanted. The pellet was then re-suspended in 500 μL of PBS and incubated with 1 mg/mL RNAse A, at 37°C for 15 minutes. Cells were centrifuge for 1200 rpm /10 min and the supernatant discarded. The pellet was then resuspended in 500 μL of PBS with the addition of 50 μg/mL of Propidium iodide (PI) staining solution. Cells were kept in the dark at 37°C for 20 min and then analysed using a FACScalibur flow cytometer and Cell Quest software (BD Bioscience).

Human BMVECs were cultured in EndoPM or EBM in standard conditions. Normal human dermal fibroblasts (NHDF, Promocell GmbH, Heidelberg, Germany) were cultured in MEM Eagle Medium (Lonza) supplemented with 10% FBS and 1% L-Glutamine. Co-cultures (n = 3) were produced by pooling BMVECS and NHDF with a defined ratio (BMVECS:NHDF = 3:1) in the following medium mix (EndoPM: EMEM medium = 4:1). Fibroblasts were previously labeled with CFDA-SE (Invitrogen, Carlsbad, CA, USA) to permit their identification in co-culture systems. The data were acquired by immunofluorescence.

Human and murine BMVECs were plated on collagen-coated permanox chamber slides (Nunc, Naperville, IL, USA). Cells were fixed and analysed for the presence of endothelial markers by means of immunostainings as previously described [17]. VE-Cadherin, Claudin-5 and β-catenin were detected, according to manufacturer instructions, 5 days after incubations of cells in EndoPM medium without Hormone Mix.

Three separate immunofluorescence analyses were performed on human and murine BMVECs; positive cells were counted in a blind manner.

Human and murine BMVECs were characterized for endothelial marker expression also by means of FACS. The controls were isotype-matched mouse IgG. The cytometric analyses were done with a FACScalibur flow cytometer and Cell Quest software (BD Bioscience). The antibodies used for immunofluorescence and flow cytometry are described in Table 2.

200 μL of Matrigel (12.5 mg/mL, BD Bioscience) at 4°C were transferred to pre-chilled 24-wells culture plates. After gentle agitation to ensure complete coating, plates were incubated for 30 minutes at 37°C to allow the solidification of Matrigel. Human and murine BMVECs were then seeded at a concentration of 6 × 10⁴/well in EndoPM [18]. Cord formation was detectable after 5-7 hours of incubation in ten fields randomly photographed.

To determine the uptake of Dil labeled acetylated low-density lipoprotein, cells were incubated with 10 mg/mL Dil-Ac-LDL (Molecular Probes, Invitrogen) at 37°C for 4 hours. The cells were washed with PBS and mounted with Fluorsave™ (Calbiochem-Millipore, Temecula, CA, USA). The slides were analyzed using a Nikon Eclipse TE300 inverted microscope equipped with a Zeiss Axiovision device camera.

2 × 10⁴ BMVECs were plated on collagen coated insert of Transwell (Corning Life Science, Union City, CA, USA) with membrane filter (0.4-μm pore size) in EndoPM medium in the upper and in the lower chamber. 5 days before assay the cells were grown in EndoPM medium without Hormone Mix, until they have reached confluence (day 3 = semiconfluence). The confluence was determined by hematoxylin-eosin staining of sentinel well. FITC-dextran (4 μL, 25 mg/mL initial concentration) (Sigma-Aldrich, St.Louis, MO, USA) was filled to the insert. Every 30 min, 50 μL of medium was collected from the lower chamber. The aliquots were diluted to 1 mL with 1× PBS. 100 μL of each diluted sample were transferred into 96-well black plates and the fluorescent content at 492/520 nm absorption/emission wavelengths for FITC-dextran was measured.

Transfer permeability of the EC monolayer correlates with the fluorescent intensity in the lower chamber. No fluorescent intensity is detected in the lower chamber once the EC monolayer reaches confluency [19].

After 60 DIV in EndoPM, endothelial genes expression of human and murine BMVECs was analyzed by means of qualitative PCR and quantitative Real Time PCR and compared with HMVECs.

Mice were obtained from Charles River Italy, Calco (LC), Italy. CD1 mice, 8-10 weeks of age, were anesthetized with a solution of Tribromoethanol (Avertin®), at a dose of 0.5 mg/g body weight. 2 × 10⁵ human BMVECs in 500 μL of HA hydrogel (Extracel™-X Hydrogel Kit, Glycosan BioSystems, Salt Lake City, UT, USA) were injected with 18-gauge needle in the abdominal quadrant keeping mice in the Trendelenburg position (head down and legs up) to decrease the risk of bowel perforation. Control samples were taken from sham operated mice treated only with HA hydrogel (Glycosan). Two weeks after transplantation, the abdominal quadrants were removed and the isolated patches were fixed in 4% paraformaldehyde. Serial sections of 20 μm, cut using a vibratome, were processed by immunofluorescence. To detect human cells, tissue sections were stained with human nuclear antigen (hNAg). Cell engraftment in the abdominal quadrant was evaluated by immunohistological analysis with UEA-1 and Draq5 (Table 2). Ten fields from 20 tissues sections were randomly selected and the images were acquired with Leica TCS SP2 AOBS (Leica Microsystems, Heidelberg, Germany) confocal laser scanning microscope.

Each experiment was performed in triplicate, values are expressed as mean ± SD. Comparisons of parameters were performed using the Student's t-test. Values of P < 0.05 (*), P < 0.01 (**) and P < 0.001 (***) were considered statistically significant.

Human and mouse BMVECs cultured in EndoPM or complete EBM were examined for their proliferation rate by growth curve, proliferation index and survival index (Figure 1). Results indicated that human and mouse cells cultured in EndoPM, compared to EBM culture condition, showed a significant increase of growth rate (Figure 1A), proliferation (Figure 1B) and survival (Figure 1C). The mean ± SD of repeated experiments, performed with different cells lines, are reported in Table 1.

Cell cycle was evaluated using a flow cytometer after overnight synchronization in starvation medium. The data showed that EndoPM facilitated the mitosis in ECs by speeding the time of duplication. Following appropriate synchronization time, starved cells were treated with complete EndoPM or EBM medium for 5h or overnight. In all the samples examined, the population of cells in mitosis was greater after restart in EndoPM than EBM.

The results are reported in Figure 1D-G and indicate that after overnight incubation in starvation medium the mean % ± SD of G0/G1 of human cells was 66.4% ± 0.9; S phase was 2.9% ± 1.2; M phase was 17.8% ± 3.3; the mean ±SD of G0/G1 of murine cells was 53.8% ± 0.7; S phase was 2.7% ± 1.6; M phase was 19.1% ± 2.6.

For human cells, after a further overnight restart with complete EBM medium, the mean % ± SD of G0/G1 was 49.7% ± 9.1; S phase was 43.2% ± 6.5; M phase was 2.6% ± 0.1 (Figure 1D); while restarting with complete EndoPM medium induced a mean % ± SD of G0/G1 of 32.9% ± 0.4; S phase of 66.3% ± 9.0; M phase of 4.5% ± 0.7 (Figure 1E). For mouse cells, after further overnight restart with complete EBM medium, the mean % ± SD of G0/G1 was 42.5% ± 10.1; S phase was 38.2% ± 7.0; M phase was 2.2% ± 0.2 (Figure 1F); while restarting with complete EndoPM medium induced a mean ±SD of G0/G1 of 33.2% ± 0.5; S phase of 67.2% ± 7.9; M phase of 4.4% ± 0.5 (Figure 1G).

These data suggest that EndoPM induced a G1 phase decrease compared with complete EBM, and was accompanied by a significant increase of cells in the S phase and a slight increase of M phase cells (P < 0.05 for both cells type). Figure 1 (D-G) is representative of repeated experiments performed with different cells lines.

To test the capability of EndoPM on the inhibition of a contaminant cell population, (e.g. fibroblasts), co-cultures of BMVECs and normal human dermal fibroblasts (NHDF) were produced by pooling both cell types at the ratio BMVECS:NHDF = 3:1. 24 h after seeding the mean densities of human BMVECs and NHDF were approximately equal in co-culture (Figure 1H). After 24 h of co-culture, the number of NHDF cells was stable. NHDF cells decreased after 7 and 14 days of culture. During co-culture condition, a significant reduction of NHDF density under the influence of EndoPM was detected compared with complete EBM culture. The reduction of NHDF cells after culture in EndoPM was approximately 95% after 14 days of culture (Figure 1H).

BMVECs cultured in EndoPM showed phenotypical and morphological stability during 12 months of culture and no evidence of overgrowth by contaminating cells was found. As seen by phase-contrast microscopy, EndoPM cultured BMVECs formed a monolayer with a cobblestone-like morphology composed of density-inhibited cells, which were not senescent, but still proliferative when split (Figure 2A).

In order to confirm the endothelial phenotype of BMVECs, several EC markers were examined by immunocytochemistry and FACS analysis.

Immunofluorescence analysis revealed strong positivity for CD31 (Figure 2B) and von Willebrand Factor (vWF) (Figure 2C). In addition, FACS analysis by multiple surface epitopes showed high expression of CD31 (human 99.42% ± 0.96; mouse 84.88% ± 16.48), CD146 (human 97.42% ± 4.3; mouse 84.19% ± 11.93), human CD105 (97.96% ± 3), mouse CD34 (82.89% ± 17.64), Ulex europaeus agglutinin-1 (UEA-1) (99.9% ± 0.07) and mouse Tie-2 (84.37% ± 3.48) (Figure 2D).

To determine whether EndoPM cultured BMVECs were able to show angiogenic function, 6 × 10⁴ cells were plated on Matrigel and the cultures were examined for capillary tube-like structure formation. BMVECs exhibited an angiogenic response within 5-7 h and migrated from the evenly distributed monolayer of cells to form net-like, capillary tube-like structures with open areas with no cells around (Figure 2E). This was observed for every sample of human and murine BMVECs, with no major differences observed.

Using immunofluorescence staining, in vitro endothelial functionality was demonstrated on BMVECs by evaluating: uptake of Dil labeled acetylated low-density lipoprotein (Dil-Ac-LDL) (Figure 2F), expression of Glucose transporter 1 (GLUT-1) (Figure 2G) and positivity for endothelial nitric oxide synthase (eNOS) (Figure 2H). A similar uptake of Dil-Ac-LDL (intense punctate staining in the perinuclear region), expression of GLUT-1 and eNOS were detected in human and murine population cultured in complete EBM condition. Overall, these observations further indicate that human and murine BMVECs cultured in EndoPM retain key phenotypic features of ECs, despite serial time-passaging and culture in vitro.

In addition, cells resulted positive for functional markers of cellular junctions as adherent junctions VE-Cadherin and β-catenin and tight junction Claudin-5, after 5 days culture in EndoPM medium without Hormone mix (Figure 3A, B, C).

Macromolecules (FITC-dextran) are used to examine the permeability of the endothelium monolayer formed after isolation of BMVECs by means of the described protocol. Although the readout of this assay is not sufficient to draw conclusions, it is clear that cells cultured in EndoPM medium and then plated in appropriate transwells strongly reduce macromolecules permeability throughout the barrier over 12 hours, as functional test of vascular permeability (Figure 3D).

Using RT-PCR assay, molecular analysis of EC biology was performed on BMVECs to compare gene expression between our isolated EndoPM human and mouse BMVECs and EC line cultured in complete EBM. The identified genes were classified into pro-angiogenic and adhesion molecules genes. BMVECs and HMVECs showed similar expression levels of the most representative angiogenic genes: vWF, KDR, FLT-1, VEGF-A, PECAM-1. In addition, an overexpression of pro-angiogenic genes such as angiopoietin (ANGPT1 +4.14 fold increase), AGTR1 (+3.28 fold increase), SELPLG (+4.02 fold increase) and ADAM17 (+2.68 fold increase) was found. Adhesion molecules genes up-regulated included ITGAV (+41.64 fold increase), ICAM1 (+8.05 fold increase), VCAM1 (+2.60 fold increase) and TNFAIP3 (+5.77 fold increase) (Figure 4A). Furthermore, CD31, vWF, Tie-2 and P-glycoprotein (P-gp) transcripts, analyzed by qualitative PCR, were detected in all cultures tested (Figure 4B).

In order to evaluate whether EndoPM cultured human BMVECs could integrate into mouse tissue and if they had the capability to construct new blood vessels, we opted for a xenograft model in the CD1 mouse strain. 2 × 10⁵ cells were engrafted in the abdominal quadrant of CD1 mouse. We analyzed 15 mice divided into two groups: group 1 (n = 10) underwent transplantation with EndoPM cultured human BMVECs embedded in hydrogel, group 2 (n = 5) underwent injection of hydrogel without BMVECs. Two weeks after transplantation, macroscopic differences were noticed in mice that received BMVECs in hydrogel (Figure 4C) compared with the mice that received only hydrogel (Figure 4D). Histological examination with human specific antibodies showed that implanted cells were immunoreactive for UEA-1 and for double staining with Draq5 and anti-human nuclei (hNAg), thus confirming the presence of human cells expressing the endothelial marker in many sections (Figure 4E). The specificity of the staining was established by the absence of human UEA-1 and anti human nuclei positive cells in the patch of mice that did not receive endothelial cells (Figure 4F).