Endothelial Sprouting and Network Formation in Collagen- and Fibrin-based Modular Microbeads

2.1 Cell culture

Umbilical cords were obtained from the University of Michigan Mott Children’s Hospital via an IRB-exempt protocol and human umbilical vein endothelial cells (HUVEC) were isolated via methods similar to those previously described [30]. In brief, the umbilical cord was washed in phosphate buffer saline (PBS) and digested in collagenase type I solution (195 U/ml, Worthington Biochemical, Lakewood, NJ) at 37°C for 20 min. The digested tissue was washed in PBS and subsequently centrifuged (200×g for 5 min). HUVEC were plated in tissue culture flasks and supplied with endothelial growth media (EGM-2, Lonza). After 24 hours, HUVEC were rinsed with PBS thrice to remove non-adherent cells and supplied with fresh media that was changed every 48 hours. Cells from passages 3 and 4 were utilized for experiments. Normal human lung fibroblasts (NHLF, Lonza Inc., Walkersville, MD) were cultured in Media 199 (M199, Life Technologies, Grand Island, NY) with 10% fetal bovine serum (FBS), and 1% penicillin/streptomycin (Life Tech). Culture media was replaced every 48 hours and cells from passages 7–12 were used in experiments.

2.2 Microbead fabrication

The process used to create protein microbeads is shown schematically in Figure 1. Two matrix formulations were used: 100% fibrin (FIB) and 40/60 wt% collagen/fibrin (COL-FIB). Collagen stock solution (4.0 mg/ml) was made under sterile conditions by dissolving lyophilized type I bovine collagen (MP Biomedicals, Solon, OH) in 0.02 N acetic acid. Fibrinogen stock solution (4.0 mg/ml clottable protein) was made by dissolving bovine fibrinogen (Sigma Aldrich, St. Louis MO) in serum-free endothelial growth medium (SFEGM-2) at 37°C. The solution was then filter sterilized after completely dissolving the protein. Equal numbers of HUVEC and NHLF were suspended in FIB or COL-FIB composite hydrogel solutions at specific cell concentrations (Low: 5×10⁵ cells/ml, High: 2×10⁶ cells/ml). To make 1.0 mL of cellular FIB hydrogel solution, the following components were added to the cell pellet and mixed thoroughly: 255 μL SFEGM-2, 100 μL of FBS (10% final), 20 μL of 50 U/mL thrombin (1 U/mL final), and 625 μL fibrinogen stock solution (2.5 mg/mL final). To make 1.0 mL of cellular COL-FIB composite hydrogel solution, the following components were added to the cell pellet and mixed thoroughly: 93 μL SFEGM-2, 100 μL FBS (10% final), 100 μL of 5X Dulbecco’s Modified Eagle Medium (DMEM (10% Final), 12 μL of 87.5 mM glyoxal (1 mM Final, Sigma) for crosslinking collagen, 50 μL of 0.1 N NaOH to neutralize acidic collagen, 20 μL of 50 U/mL thrombin (1 U/mL final), 250 μL collagen stock solution (1.0 mg/mL final), and 375 μL fibrinogen stock solution (1.5 mg/ml final).

To create protein microbeads, 3.0 mL of matrix/cell suspension was quickly added to a stirred bath (600 rpm) containing 75 mL of 100 cSt polydimethylsiloxane (PDMS, Dow Corning) cooled in an ice bath. Emulsification proceeded for 5 min, and the temperature of the PDMS bath was then increased to 37 °C to induce gelling for an additional 25 min. The mixture of PDMS and formed microbeads was then transferred to 50 mL centrifuge tubes and 5 mL of PBS containing 0.1% L101 surfactant (PBS-L101; BASF, Florham Park, NJ) was added to each tube and mixed gently by inverting for 5 min. The mixture was then centrifuged at 200xg for 5 min. The PDMS layer was carefully removed and then pelleted microbeads were washed again using PBS-L101. Once the microbeads were thoroughly washed, they were supplied with an appropriate amount of culture medium (EGM-2) and used for further experiments.

2.3 Characterization of cell incorporation and microbead morphology

The number of cells incorporated within microbeads was determined by staining with 4′,6-diamidino-2-phenylindole (DAPI, Life Technologies). The FIB and COL-FIB samples were collected and fixed with zinc-buffered formalin solution (Z-fix, Anatech, Battle Creek, MI) for 5 min. Fixed samples were washed in PBS and stained with 10 nM DAPI solution for 10 min. Bright field and fluorescent images were taken using an optical microscope (Olympus IX81, Olympus, Center Valley, PA). The number of cells per microbead and microbead diameters were determined from these images using ImageJ (National Institutes of Health, Bethesda, MD). At least 100 randomly selected microbeads per condition and per experiment were used for image analysis to determine microbead diameter, cell distribution, and number of cells incorporated into the microbeads.

2.4 Cell viability and proliferation assays

The viability of cells encapsulated in microbeads and in bulk gels was determined using the Live/Dead® staining kit (Life Technologies). Bulk gel samples with the same protein and cell concentration as the microbeads were also prepared by allowing the hydrogel/cell mixture to gel in 24-well plates (500 μL per well) at room temperature for 5 minutes, followed by an additional 20 to 25 min at 37 °C in an incubator. Samples were collected, washed in 10 mM PBS and re-suspended in PBS containing 1 μM calcein-AM and 2 μM ethidium homodimer for 45 min at 37°C. Samples were then imaged using a fluorescence microscope.

Both cell seeded COL-FIB and FIB microbeads were embedded in fibrin gels after their fabrication process. To embed microbeads in gels, one tenth of a 3 mL microbead preparation was collected and pelleted to make three fibrin hydrogels with microbeads per time point (~500 μL per hydrogel). These hydrogels were made by adding the following components to the microbead pellet and mixing thoroughly: 382.5 μL SFEGM-2, 150 μL of FBS (10% final), 30 μL of 50 U/mL thrombin (1 U/mL final), and 937.5 μL fibrinogen stock solution (2.5 mg/mL final). Hydrogels were molded into standard 24-well culture plates, and were left at room temperature for 5 min before being placed in an incubator for 25 minutes at 37°C. One milliliter of EGM-2 was added to each hydrogel containing the cellular microbeads. Culture medium was replaced the next day and then on every second day for the duration of the culture period.

Cell proliferation was determined by quantifying total DNA in samples collected at specific time intervals. To isolate the total DNA from the samples, fibrin gels containing microbeads were digested in 10 mM Tris-HCL/0.4 M Guanidine hydrochloride solution for 3 hours. The digested samples were then centrifuged at 10,000 × g for 10 min and the supernatant containing the DNA was aspirated, and then diluted with water to reduce Guanidine-HCl interference with the DNA assay. DNA content was determined using the PicoGreen® DNA assay kit (Invitrogen). Calf thymus DNA standards were used to generate standard curves.

2.5 Imaging and analysis of endothelial sprouts

To distinguish the microbeads from the 3D fibrin environment in which they were embedded, FIB and COL-FIB microbeads fabricated as described above were subsequently coated with a FITC-fibrinogen layer by incubating them in FITC-fibrinogen (9 μg/mL, Life Technologies) in EGM-2 at 37 °C for 25 min prior to embedding within the larger hydrogel. The same embedding process utilized for the proliferation assay (Section 2.4) was employed to incorporate the cellular microbeads within acellular fibrin hydrogels. To visualize vascular sprouts, the endothelial cell-specific marker Ulex Europaeus Agglutinin I (UEA-I, Vector Laboratories, Burlingame, CA) was used. In brief, samples were collected at specified time intervals, washed twice in PBS, and fixed in zinc-buffered formalin (Z-fix, Anatech) at room temperature. Fixed samples were then washed and incubated in staining solution containing 1% BSA, 20 μg/ml rhodamine-labeled UEA-I, and 10 nM DAPI in PBS for 45 min at room temperature. Before imaging, samples were rinsed twice with PBS for 5 min to remove unbound residual stains.

Endothelial sprouts were characterized by measuring average sprout lengths, numbers of segments (branching), branch points, and lengths of each segment both inside and outside the microbeads through image analysis using a protocol similar to previous studies [30, 37, 44]. Inosculation of adjacent sprouts protruding from adjacent microbeads was assessed qualitatively. Fluorescence images were acquired using an optical microscope and randomly selected microbeads were used for quantification using ImageJ. All image analysis was performed using a blinded image evaluation method in which random images were provided to the evaluator with no information about the experimental conditions.

2.6 Statistical Analysis

Statistical analyses were performed using STATISTICA (StatSoft. Inc, Tulsa, OK) software. At least 3 independent experiments were performed for each assay and data are reported as mean ± standard error of the mean (SEM). One-way ANOVA followed by unequal N post-test was used to compare multiple conditions. Values of p ≤ 0.05 were considered statistically significant.

3.1 Cell encapsulation and retention in microbeads

HUVEC and NHLF were encapsulated at a 1:1 ratio in FIB and COL-FIB microbeads, at either low cell concentration (LC: 5.0×10⁵ cells/ml) or high concentration (HC: (2.0×10⁶ cells/ml). At the same starting concentration, larger microbeads contained greater total numbers of cells, as expected (Fig. 2A). The average microbead diameter was approximately 140 μm for both FIB formulations and the low cell concentration COL-FIB formulation; however the high cell concentration COL-FIB microbeads were significantly larger (approximately 200 μm in diameter, Fig. 2B, p<0.05). The distribution in the number of cells per microbead of the treatment populations was broader for the high cell concentration conditions, in contrast to microbeads containing the low cell concentration, which exhibited a sharper peak and a more homogeneous distribution (Fig. 2C). Overall cell incorporation increased as higher cell concentrations were used with the same material formulation (Fig. 2D, p<0.05). Microbeads with high cell concentration contained an average of 30–35 cells per microbead, while those with low cell concentration contained 5–10 cells per microbead. There was no statistical difference in the total number of cells incorporated between COL-FIB and pure FIB microbeads, when comparing at either the low concentration or high cell concentration conditions.

3.2 Viability and proliferation of cells encapsulated in microbeads

The viability of cells encapsulated in FIB and COL-FIB microbeads in the high cell concentration (2.0×10⁶ cells/ml) was quantified and compared to values from bulk macroscopic hydrogels (Fig. 3). Representative live (green)/dead (red) images of COL-FIB microbeads (Fig. 3A), FIB microbeads (Fig. 3B), COL-FIB bulk hydrogels (Fig. 3C), and FIB bulk hydrogels (Fig. 3D) show uniform cell distributions in both the microbeads and the hydrogels, with no spatial variations in viability evident. Cell viabilities in both COL-FIB and FIB microbeads were high (75% and 79% viable cells, respectively), with no statistically significant differences between the formulations. However, cell viability was significantly lower in the microbeads compared to their bulk hydrogel counterparts, which exhibited 87% and 85% viable cells, respectively (Fig. 3E).

COL-FIB and FIB microbeads made at the high cell concentration (2.0×10⁶ cells/ml) were embedded in surrounding 2.5 mg/ml fibrin hydrogels to examine cell proliferation and endothelial sprout formation. Figure 4A shows a schematic of the protocol used for cell proliferation studies. Microbeads were visualized inside the larger fibrin matrix using brightfield and fluorescent microscopy (Fig. 4B). In these images, FIB microbeads were embedded in a FIB hydrogel immediately after fabrication and were cultured for 7 days before imaging. In the fluorescence image, microbeads were coated with a FITC-fibrinogen layer to distinguish the microbead matrix from the surrounding fibrin. At day 0, both microbead formulations showed equal DNA content, which then increased over time in culture as cells proliferated. At day 7, there was a significant difference (p<0.05) in the DNA content between FIB and COL-FIB microbeads. By day 14, the DNA content was 2.5-fold higher (p<0.05) in COL-FIB microbeads than in corresponding FIB microbeads, indicating a significantly higher number of cells within these beads (Fig. 4C).

3.3 Endothelial sprouting from microbeads embedded in fibrin matrix

COL-FIB microbeads made with the lower cell concentration did not support endothelial sprouting when embedded in fibrin hydrogels (data not shown), while those made with the higher cell concentration retained the capacity for sprout formation when embedded in fibrin hydrogels. Therefore, the vasculogenic capacity of COL-FIB and FIB microbeads were compared only at the higher initial cell concentration. COL-FIB and FIB microbeads containing HUVEC and NHLF were embedded in fibrin hydrogels and cultured for 7 and 14 days (representative images shown in Fig. 5). The microbead matrix was coated with a FITC-fibrinogen layer (green) and the endothelial cells were identified by UEA-1 staining (red). In both microbead formulations, endothelial sprouts initially formed inside the microbead matrix and then invaded into the surrounding fibrin hydrogel.

The extent of sprouting both within the microbeads and in the surrounding matrix were characterized in terms of the number of segments (Fig. 6A), the average segment length (Fig. 6B), the number of branch points (Fig. 6C), and the total sprout length (Fig. 6D) (see also Tables 1 and 2, Supplementary data). Microbeads contained an average of 4 to 5 vessel segments by the end of 7 days in culture, with no further change in the number of segments within the microbeads by day 14 (Fig. 6A, “inside microbead”). The number of segments that extended from the FIB microbeads was significantly higher compared to COL-FIB microbeads at both time points. A higher number of segments protruding the COL-FIB microbeads was found with an additional week of culture (Fig. 6A, “outside microbead”). By day 14, the total number of vessel segments per microbead was 10–14, and this measure was significantly higher in FIB microbeads than in COL-FIB microbeads at both day 7 and day 14 of culture (Fig. 6A, “total per microbead”).

The average vessel segment length inside microbeads was approximately 90 μm, with no significant differences across matrix formulations or time points (Fig. 6B, “inside microbead”). Outside the microbeads, longer segments were produced by FIB microbeads than COL-FIB microbeads (p<0.05), and segment length increased significantly outside FIB microbeads between 7 and 14 days (Fig. 6B, p<0.05). Both COL-FIB and FIB microbeads on average contained at least one branch point, and the number of branch points was significantly higher in FIB microbeads at day 14 (Fig. 6C, p<0.05), compared to the other formulation and time point. The number of branch points outside the microbeads and the total number of branch points were also significantly higher in FIB microbeads compared to COL-FIB at both time points (Fig. 6C, p<0.05).

The combination of vessel network characteristics described also resulted in differences in the total sprout network length (Fig. 6D). The average total sprout length inside microbeads was 315–460 μm, and there were no statistically significant differences between matrix formulations or time points. However, the vessel networks that penetrated into the surrounding fibrin hydrogel were significantly longer when emerging from FIB microbeads compared to COL-FIB microbeads at both day 7 and 14 (Fig. 6D, “outside microbeads”, p<0.05). Total sprout length increased significantly in COL-FIB microbeads over time, but was lower than in FIB microbeads at all time points (Fig. 6D, “total per microbead”, p<0.05). By day 14, the microvascular network produced by HUVEC and NHLF encapsulated within FIB microbeads approached 2500 μm in total length.

3.4 – Inosculation between endothelial sprouts from adjacent microbeads

COL-FIB and FIB microbeads embedded in fibrin hydrogels were cultured for 1 or 2 weeks. The microbead matrix was labeled green prior to the embedding process to distinguish them from the surrounding fibrin matrix, while the HUVEC were stained in red (Fig. 7). Microbeads were generally separated from each other in the surrounding gel, though some clumping was observed. As described quantitatively in Figure 6, endothelial cell networks extended from the embedded microbeads by day 7 (Fig. 7A–D), and were more prominent in the pure FIB samples. The arrows in Figure 7 show anastomoses between endothelial sprouts from adjacent microbeads. This phenomenon is evident in both microbead formulations at both day 7 and day 14 and at 4× and 10 magnification. The nuclei of both EC and NHLF were stained in blue (Fig. 7B, D, F, and H). It can be seen that NHLF migrated out of the microbeads and into the surrounding matrix, and that endothelial sprouts extended and inosculated between microbeads. By day 14, relatively long distances (1000–2000 μm) were traversed by joined endothelial sprouts (Fig. 7E–H).