Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Acta Biomater. Author manuscript; available in PMC 2017 April 15.
Published in final edited form as:
PMCID: PMC4829480

Stable engineered vascular networks from human induced pluripotent stem cell-derived endothelial cells cultured in synthetic hydrogels


Here, we describe an in vitro strategy to model vascular morphogenesis where human induced pluripotent stem cell-derived endothelial cells (iPSC-ECs) are encapsulated in peptide-functionalized poly(ethylene glycol) (PEG) hydrogels, either on standard well plates or within a passive pumping polydimethylsiloxane (PDMS) tri-channel microfluidic device. PEG hydrogels permissive towards cellular remodeling were fabricated using thiol-ene photopolymerization to incorporate matrix metalloproteinase (MMP)-degradable crosslinks and CRGDS cell adhesion peptide. Time lapse microscopy, immunofluorescence imaging, and RNA sequencing (RNA-Seq) demonstrated that iPSC-ECs formed vascular networks through mechanisms that were consistent with in vivo vasculogenesis and angiogenesis when cultured in PEG hydrogels. Migrating iPSC-ECs condensed into clusters, elongated into tubules, and formed polygonal networks through sprouting. Genes upregulated for iPSC-ECs cultured in PEG hydrogels relative to control cells on tissue culture polystyrene (TCP) surfaces included adhesion, matrix remodeling, and Notch signaling pathway genes relevant to in vivo vascular development. Vascular networks with lumens were stable for at least 14 days when iPSC-ECs were encapsulated in PEG hydrogels that were polymerized within the central channel of the microfluidic device. Therefore, iPSC-ECs cultured in peptide-functionalized PEG hydrogels offer a defined platform for investigating vascular morphogenesis in vitro using both standard and microfluidic formats.

Keywords: Thiolene, Biomaterials, Blood Vessel, Tissue Engineering, Microfluidic, 3D culture


The lack of a functional vasculature and pathological disruption of circulation are unresolved challenges that have limited the success for many tissue engineering and wound healing approaches [13]. Furthermore, the incorporation of a vascular component is expected to improve human cellular models by recapitulating cell-cell interactions during tissue formation [4, 5] and by supporting the function of model organ systems [6, 7]. Finally, while target organs such as the heart, liver or central nervous system have been the focus for many in vitro toxicity screening strategies [8], vascular models have also been identified as a promising tool for predictive toxicology [9, 10]. Therefore, several emerging applications would benefit from in vitro assays that enable systematic investigation of factors that promote blood vessel formation and stabilization [13].

Endothelial cells cultured in vitro will spontaneously “self-assemble” into organized networks [1117], and several studies have demonstrated that capillary tubules can be perfused when subjected to flow [1823]. While extracellular matrix (ECM) components such as collagen or Matrigel are often used as culture substrates when modeling vascular morphogenesis in vitro [1214, 16, 17], these materials can be limiting for screening approaches due to batch variability, properties that are sensitive to reaction conditions, and poorly-defined compositions [2426]. To address these limitations, synthetic strategies have increasingly been applied to investigate factors that instruct endothelial phenotypes [2735]. Hydrogels formed via thiol-ene photopolymerization represent an emerging class of cell culture materials [36, 37] that are formed through a radical-initiated step-growth mechanism that couples thiols and alkenes with high specificity [38]. A growing body of literature has demonstrated the versatility of thiol-ene photochemistry for incorporating biomolecules such as peptides, growth factors, gelatin, and hyaluronic acid into synthetic hydrogels [4, 3537, 3947]. Hydrogels formed via thiol-ene photopolymerization enable spatial patterning of biochemical and mechanical properties [35, 3941], sequestering and controlled release of growth factors [45], rapid photopolymerization for 3D bioprinting of encapsulated cells [44], and protein-free backgrounds for identifying ECM components deposited in the matrix during cellular remodeling [47]. Thus, thiol-ene chemistry offers a potentially powerful tool for modeling vascular morphogenesis by providing control over a wide range of matrix properties relevant to blood vessel formation [4, 35].

While engineering platforms provide control over the 3D microenvironment when modeling vascular morphogenesis [13], the heterogeneity and donor-to-donor variability of primary human endothelial cells may be limiting for applications that require standardization or scale-up [1, 48]. Human umbilical vein endothelial cells (HUVECs) can be used for standardized screening of angiogenesis inhibitors in vitro, but require processing and pre-validation to identify donor sources with similar function [10]. Human pluripotent stem cells [4951] offer promise for predictive toxicology [8], and have been used to derive endothelial cells that form vascular networks in vitro and functional blood vessels in vivo [23, 30, 52, 53]. Importantly, human induced pluripotent stem cell-derived endothelial cells (iPSC-ECs) can be produced with high batch uniformity [23], which may be beneficial for vascular disease models or screening approaches that require standardization or scale-up [9, 54].

The strategy reported here combines a uniform endothelial cell source [23], a tunable synthetic ECM [36], and a tri-channel microfluidic device [55] to model vascular morphogenesis in vitro. Thiol-ene photopolymerization was used to fabricate peptide functionalized PEG hydrogels permissive towards cellular remodeling [36] and the iPSC-ECs were previously characterized by high lot-to-lot purity to at least 6 passages [23]. The passive flow concept that is the essence of the microfluidic device described here uses standard culture techniques for loading cells and exchanging media [55], and therefore provides an accessible format that takes advantage of microscale features such as decreased requirements for reagents and cells [56]. Our combined results provide evidence that iPSC-ECs self-assemble into vascular networks through physiologically-relevant mechanisms when cultured in PEG hydrogels, and capillary tubules with lumens were stable for at least two weeks when the hydrogels were polymerized within the microfluidic device.


2.1 Cell culture

Human induced pluripotent stem cell-derived endothelial cells (“iPSC-ECs”, Cellular Dynamics, iCell® Endothelial Cells) were cultured according to the manufacturer’s protocol. Briefly, iPSC-ECs were expanded to passage 3 and cryopreserved for additional use. Passage 3 iPSC-ECs were thawed and plated at 10,000–15,000 iPSC-ECs/cm2 onto tissue culture plates treated with 3 µg/cm2 fibronectin (Invitrogen) and passaged every 3 to 4 days with TrypLE (Invitrogen). The manufacturer’s recommenced growth medium was used for culturing iPSC-ECs, which consists of VascuLife VEGF Medium (Lifeline Cell Technologies) that was modified as follows: 10 mL glutamine supplement was added to 500 mL medium (rather than the 25 mL provided) and the FBS supplement was replaced with iCell® Endothelial Cells Medium Supplement (Cellular Dynamics). Cells were incubated at 37°C and 5% CO2 for all experiments. Microfluidic channels. Microfluidic experiments were conducted using iPSC-ECs cultured in the growth medium described above (“Control”), or in growth medium supplemented within 100–1000 ng/mL VEGF-165 (Catalog # 293-VE, Lot 114714051, 97% purity from R&D), as described in Results and Discussion. The iPSC-ECs were encapsulated in PEG hydrogels that were polymerized within the central channel of the tri-channel device (see details below). All fluid was removed from both outer channels daily and replaced with a total of 10 µL of fresh medium.

2.2 Microfluidic Device Fabrication

The tri-channel microfluidic device was fabricated as previously described without modification [55]. Briefly, polydimethylsiloxane (PDMS, Sylgard 184 Silicone Elastomer Kit, Dow Corning) elastomer base and curing agent were mixed at a 10:1 ratio and degassed under vacuum for 45 min (room temperature). The degassed PDMS was poured over SU-8 master molds, which were generated using standard soft lithography methods [57]. The PDMS was cured for 4 hr (80°C), allowed to cool to room temperature, and removed from the master mold. The PDMS device was autoclaved for 20 min at 120° C. Six hr before loading the cell/monomer solution, devices were oxygen-plasma-treated to bond the PDMS channels to the inside of a glass-bottom Petri dish (MatTek).

2.3 Poly(ethylene glycol) (PEG) hydrogel preparation

Poly(ethylene glycol) (PEG) hydrogels were formed using thiol-ene photopolymerization chemistry (Fig. 1A) [36]. For most experiments, 8-arm PEG-norbornene was purchased from a commercial source (JenKem USA: 20,000 MW, 8ARM (TP)-NB-20K). For some experiments, 8-arm PEG-NB monomer was synthesized as previously described [35]. Stock PEG solutions were prepared by adding 0.8 mL 1× PBS to 300 mg lyophilized 8-arm PEG-NB powder (final volume = 1 mL) and filtered through a 0.2 µm nylon syringe filter (Fisher) for a final concentration of 300 mg/mL sterile monomer. Monomer solutions for cell encapsulation were prepared in 1× PBS with 40 mg/mL 8-arm PEG-NB in which 40–60% of the available norbornene arms were cross-linked with a matrix metalloproteinase (MMP)-degradable peptide with cysteines flanking the active sequence (KCGGPQG*IWGQGCK, GenScript; active sequence in bold; * = cleavage site) [58, 59]. To promote cell adhesion [60], 2 mM CRGDS (GenScript, active sequence in bold) was incorporated as a pendant group through the terminal cysteine.

Figure 1
Synthetic extracellular matrix (ECM) formulations for culturing iPSC-ECs

2.4 Quantification of Shear Modulus

To measure the shear modulus of PEG hydrogels, 72 µL precursor solutions were pipetted into 8.0 mm diameter, 1.2 mm depth Teflon wells and cured for 8 sec using 365 nm UV light at a dose rate of 90 mW/cm2. The resulting hydrogels were swollen to equilibrium in 1× PBS for 24 hr. The samples were tested using an Ares-LS2 rheometer (TA Instruments). A 20 g force was applied to the samples via parallel plate crossheads and a strain sweep test at 1 Hz fixed frequency was performed from 0.1 to 20% strain. If the sample was not robust enough to withstand a 20 g force the gap between the parallel plates of the rheometer was set to 1.0 mm distance. Measurements for complex shear modulus for each hydrogel formulation were taken at 1 Hz, 2–20% strain.

2.5 Cell encapsulation and polymerization of PEG hydrogels

Cells pellets were re-suspended in a 2× photoinitiator solution (0.1 % wt./wt. Irgacure 2959) and mixed 1:1 with a 2× PEG monomer solution. For experiments using standard plates, 5 µL of the cell/monomer solution was pipetted onto tissue culture treated surfaces and polymerized for 2 min using ~5–10 mW/cm2 UV light centered at ~365 nm (multiwell plates were placed on the top shelf of the exposure stand for a UVP XX-15L series UV lamp, Fisher). Microfluidic device. For microfluidic experiments, 2.5 µL of the cell/monomer solution was polymerized in the middle channel of the tri-channel device.

2.6 Microscopy

Immunofluorescence images were collected using a Nikon A1R laser scanning confocal microscope with Plan Apo 10×, Plan Fluor 20× Ph1 DLL, or Plan Apo 20× DIC M objectives (0.95–3.35 µm z-steps) unless otherwise noted. Some confocal z-stacks were drift corrected before creating maximum projection images using the “Align Current ND Document” command in NIS Elements. Time-lapse and viability images were collected using a Nikon TI Eclipse fluorescence microscope (20 µm z-steps). For time-lapse imaging, cells were housed in a Nikon environmental chamber with external heater (in vivo Scientific) and CO2 regulator (in vivo Scientific) to control temperature and CO2 levels. Some time-lapse images were drift corrected using the StackReg plugin [61] for ImageJ [62, 63].

2.7 Quantification of cell viability

Cell viability was quantified for iPSC-ECs encapsulated in PEG hydrogels using the LIVE/DEAD® Viability/Cytotoxicity kit (Life Technologies). Cells were rinsed with 1× PBS and stained with calcein AM and ethidium homodimer-1 for 30 min using the manufacturer’s recommended dilutions. Samples were then washed with 1× PBS and fixed in 10% buffered formalin (4% formaldehyde, Fischer) for 30 min. Following fixation, cell nuclei were stained with 1:500 DAPI (Sigma-Aldrich) for 30 min. The fraction of dead cells was quantified by determining the ratio of ethidium homodimer+ nuclei to total nuclei.

2.8 Immunofluorescence (IF) imaging

Standard well plates. Antibody dilutions were 1:200 mouse anti-CD31 (MAB2148, R&D Systems or M082301–2, DAKO), 1:100 goat anti-VE cadherin (AF938, R&D Systems), and 1:200 secondary antibodies (Alexa fluor 488 donkey anti-mouse, A-21202 or Alexa fluor 568 donkey anti-goat, A-11057, Life Technologies). Cells encapsulated in PEG hydrogels were fixed in 10% buffered formalin (4% formaldehyde, Fischer) for 30–60 min. Following fixing, samples were rinsed 2×15 minutes with wash buffer (0.05% Triton X-100 in 1× PBS). Cells were then permeabilized and blocked for at least 60 min using 0.25% Triton X-100 and 1% (wt/wt) bovine serum albumin (BSA, Fisher Scientific) in 1× PBS, followed by an additional two rinses with wash buffer. Samples were then incubated with primary antibody in incubation buffer (0.05% Triton X100 and 1% BSA in 1× PBS) for 4 hr at room temperature or overnight at 4°C. After rinsing 2 × 15 min and once for at least 30 min in wash buffer, samples were incubated with secondary antibody and 1:500 DAPI (Sigma-Aldrich) in incubation buffer for 4 hr at room temperature or overnight at 4°C. Samples were rinsed 2 × 15 min in wash buffer and stored in 1× PBS at 4°C until imaging (at least overnight). Microfluidics device. Cells encapsulated within microfluidic devices were fixed with 4% paraformaldehyde (PFA) for 30 min. Samples for immunofluorescence imaging of CD31 and VE-Cadherin were prepared as described above. For experiments that illustrate F-actin only, samples were rinsed with 1× PBS and then incubated with 1:1000 DAPI and 1:50 phalloidin in 1× PBS (added to the side channels) for 1 hr at room temperature, followed by at least 2×15 min incubation in 1× PBS. All samples were stored in 1× PBS until imaging.

2.9 RNA sequencing

Total RNA was isolated using the RNeasy Kit (Qiagen) and included lysing in 350 µL RLT lysis buffer and the optional DNase treatment. cDNA libraries were prepared from 50 ng of total RNA and indexed with Illumina’s TruSeq RNA Sample Preparation Kit v2 (RS-122-2001 and RS-122-2002). Final indexed cDNA libraries were pooled with 12 uniquely indexed TruSeq cDNA libraries for a total of 6 samples per lane. Multiplexed samples were sequenced on an Illumina HiSeq 2500 with a single 51 bp read and a 7 bp index read. Base-calling and demultiplexing were performed using Casava (v1.8.2). Sequences were filtered and trimmed to remove low quality reads, adapters, and other sequencing artifacts. The remaining reads were aligned to 19084 Refseq genes extracted from the Illumina iGenomes reference, selecting only those with ‘NM_’ annotations. Bowtie (v0.12.9) was used for alignment, allowing two mismatches in a 28 bp seed [64]. Reads with more than 200 alignments were excluded from further analysis. RSEM (v1.2.3) was used to estimate isoform and relative gene expression levels (transcripts per million or “TPM”) [65]. LOG-fold changes are calculated as LOG2[(iPSC-EC TPM+1)/(iPSC TPM +1)] to avoid calculation errors associated with samples that have zero expression. Differential gene expression. Differentially expressed genes between individual samples were calculated from RSEM expected read counts (EC) using EBSeq (v1.5.3) [66], with median quantile normalization of EC and a maximum False Discovery Rate of 0.005. Differentially expressed genes between individual samples were calculated in EBSeq using expected read counts and the False Discovery Rate cutoff was set at FDR ≤ 0.005. Gene Ontology analysis. The DAVID Bioinformatics Database Functional Annotation Tool (v6.7) was used to identify Gene Ontology terms [6770] from differentially expressed genes. DAVID analysis was done using the following settings: Gene Ontology category GOTERM_BP_FAT; Threshold options: Counts = 10, EASE = 0.001. GO lists presented in Supplemental Tables 2–3 were limited to the Top 25 GO terms and Benjamini corrected p-value ≤ 0.05 (thus, some lists included fewer than 25 terms). Code availability (EBSeq). Details about the EBSeq algorithm were previously described in detail [66]. EBSeq download and documentation is available at:

Additional files and instructions for downloading the user interface and instructions for installing the EBSeq toolshed for Galaxy are available at:

2.10 Statistical analysis

Statistical significance for data in Suppl. Fig. 5 was calculated using a one-way ANOVA with a post hoc Tukey-Kramer test for individual comparisons (α = 0.05).


Induced pluripotent stem cell-derived endothelial cells (iPSC-ECs) were encapsulated in peptidefunctionalized poly(ethylene glycol) (PEG) hydrogels to establish an in vitro vascular model using a uniform cell source [23] and a synthetic extracellular matrix (ECM) [36]. Thiol-ene photopolymerization was used to incorporate protease-degradable peptide crosslinks [58] and cell adhesion peptides [60] into PEG hydrogels to provide a synthetic ECM permissive towards cellular remodeling (Fig. 1A) [36]. The iPSC-ECs were previously characterized by uniform purity between lots and functional characteristics that included thrombin-dependent barrier function, TNF-α responsiveness, and shear stress-induced alignment [23]. Here, calcein/ethidium homodimer staining (Fig. 1B–C) and time-lapse microscopy (Suppl. Fig. 1, Suppl. Movie 1) demonstrated that iPSC-ECs were viable and self-assembled into interconnected vascular networks during the first three days of culture in peptide-functionalized PEG hydrogels. After encapsulation, iPSC-ECs condensed into clusters, elongated, and extended protrusions to establish connections (Suppl. Fig. 1A, Suppl. Movie 2), which resembled vasculogenic sprouting in vivo [71, 72]. Sprouting from existing tubules played a dynamic role during iPSC-EC assembly (Suppl. Fig. 1B, Suppl. Movie 1), as sprouts often retracted before establishing new connections or formed tubules that later disassembled (Suppl. Movie 3). Vacuoles also appeared to play a role in tubule formation by iPSC-ECs (Suppl. Movie 4), such as previously observed in vitro and in vivo [73]. By day 3, capillary networks with polygonal organization were evident throughout the hydrogel spot (Suppl. Fig. 1C), which resembled mechanisms described for endothelial cells in naturally-derived matrices in vitro [1517] and blood vessel development in vivo [71, 72].

RNA-Seq was used to analyze global gene expression for iPSC-ECs during vascular network formation in PEG hydrogels (Suppl. Table 1). Immunofluorescence imaging demonstrated that iPSC-EC networks in PEG hydrogels were CD31+ and VE-Cadherin+ by immunofluorescence (Fig. 2A–C), and both genes were highly expressed by RNA-Seq for 2D and 3D culture (Fig. 2D–E). Normalized gene expression was ranked for iPSC-ECs within the Gene Ontology (GO) categories vasculature development (GO:0001944) and biological adhesion (GO:0022610) (Suppl. Fig. 2), which are functional terms from the Gene Ontology Consortium [69, 70] chosen for their relevance to mechanisms that guide vascular morphogenesis [17, 74, 75]. Genes that were highly expressed by iPSC-ECs included many regulators of vascular function, such as cellmatrix and cell-cell adhesion genes (e.g., KDR/VEGFR-2, CD99, MCAM/CD146, CLDN5/claudin-5, integrins, etc.) and genes relevant to matrix remodeling (e.g., laminins, collagens, TIMPs, and MMPs; Suppl. Fig. 2) [17, 74]. Thus, iPSC-ECs were characterized by gene expression profiles and cell-cell adhesions that are characteristic of an endothelial phenotype, which agrees with previous results using the same cell line [23].

Figure 2
Expression of characteristic adhesion components by iPSC-ECs cultured in PEG hydrogels

EBseq [66] was then used to identify differentially expressed genes for iPSC-ECs compared to undifferentiated iPS cells (Suppl. Table 2) or iPSC-ECs in 3D (PEG hydrogels) relative to 2D (TCP) culture (Suppl. Table 3). Differentially expressed genes were further analyzed for functional properties by identifying GO terms using the DAVID Functional Annotation Tool [6770]. Genes upregulated by iPSC-ECs relative to iPS cells were enriched within GO categories that included vasculature development (Fig. 3), biological adhesion, cell motion (GO:0006928), and others that were relevant to vascular morphogenesis (see Suppl. Table 2 for full lists). EBSeq identified 23 upregulated and 16 downregulated vasculature development genes for iPSC-ECs in 3D relative to 2D culture (Fig. 4A; at least one time point). Genes that were upregulated by iPSC-ECs in 3D culture also included 20 regulation of locomotion genes (GO:0040012) and 19 response to hypoxia genes (GO:0001666), whereas genes upregulated in 2D culture were predominantly enriched within GO categories related to proliferation (e.g., 127 cell cycle genes) (Suppl. Table 3). Upregulated genes in 3D culture also included Notch signaling genes, adhesion genes, and others that regulate mechanisms such as tip and stalk cell specification and branching during angiogenesis (Fig. 4B; NOTCH4, GJA4, GJA5, CLDN5, UNC5B, etc.) [74, 76], which is consistent with observations by time-lapse microscopy.

Figure 3
Differentially expressed vasculature development genes for iPSC-ECs relative to undifferentiated iPS cells
Figure 4
Differentially expressed vasculature development genes for iPSC-ECs in 2D and 3D culture

Vascular morphogenesis is dependent on coordinated interactions between endothelial cells and the ECM, which includes proteolytic matrix remodeling, a balance of cell-cell and cell-matrix adhesions, and deposition of matrix components such as collagens and laminins (e.g., for basement membrane assembly) [16, 17, 74]. Genes that were upregulated by iPSC-ECs in 3D relative to 2D culture included COL1A1, COL1A2, COL6A3, and ITGA1, which encode the α1 and α2 subunits of type I collagen, the α3 subunit of type VI collagen, and integrin α1, respectively (Fig. 4, Suppl. Table 3). Integrin α1 is a cell adhesion receptor for interstitial collagens, as well as ECM components of the basement membrane (collagen IV and laminin-1 [77, 78]) and vascular subendothelium (collagen VI [79, 80]) [81]. Further, endothelial cells have been characterized by increased expression of collagen I during tube formation in vitro [8285] and α1 integrin during angiogenesis in vivo [86, 87] and vascular morphogenesis in vitro [88]. Additional genes that have been implicated in ECM remodeling during vascular morphogenesis and were upregulated for iPSC-ECs in 3D culture included MMP9 (matrix metalloproteinase 9), TIMP3 (tissue inhibitor of metalloproteinase-3) and NID2 (nidogen-2) [17] (Suppl. Table 3), while other related matrix remodeling and cell adhesion genes were highly expressed in both 2D and 3D culture (Suppl. Fig. 2). Our combined results demonstrate that iPSC-ECs self-assemble in PEG hydrogels through mechanisms that are consistent with in vitro and in vivo vascular morphogenesis [16, 17, 71, 74].

Finally, vascular network formation was investigated for iPSC-ECs encapsulated in PEG hydrogels that were polymerized within a tri-channel microfluidic device (Figs. 5A) [55]. The passive flow microfluidic format uses standard pipetting to introduce cells and exchange media [55], and such “tubeless” designs can be arrayed and interfaced with automated liquid handlers for enhanced-throughput applications [89]. Vascular network formation was investigated within PEG hydrogels with shear moduli ranging from 183 ± 10 – 1612 ± 95 Pa, which were fabricated by maintaining a constant concentration for the PEG-NB backbone (2 mM 8-arm PEG-NB, 16 mM available norbornene groups) while varying the fraction of MMP-degradable peptide crosslinks (40–60% molar ratio thiol:norbornene) (Fig. 5B). Only limited tubule formation was evident after eight days when 1×107 iPSC-ECs per mL were encapsulated in 40–60% crosslinked hydrogels and cultured in basal medium (“Control”, includes 5 ng/mL VEGF, Fig. 5C,E,G). Vascular network formation was improved for each PEG formulation when iPSC-ECs were cultured in basal medium supplemented with 200 ng/mL VEGF (“200 VEGF”, Fig. 5D,F,H), and further optimization for iPSC-ECs encapsulated in 50% crosslinked PEG hydrogels demonstrated that vascular networks were most pronounced at intermediate VEGF concentrations (Suppl. Fig. 3). The optimal VEGF dose qualitatively increased when iPSC-ECs were encapsulated in PEG hydrogels at higher cell density (8.5×107 iPSC-ECs per mL), and vascular network formation was completely disrupted at the highest VEGF concentration (1000 ng/mL) (Suppl. Fig. 4). Importantly, capillary tubules with lumens were stable to at least 14 days in the microfluidic device when 1×107 iPSC-ECs per mL were encapsulated in 50% crosslinked PEG hydrogels and cultured in medium that was supplemented with 100 ng/mL VEGF (Fig. 6). Thus, vascular network formation by iPSC-ECs depended on an optimal VEGF concentration that was cell density-dependent, which is consistent with the role for VEGF during vascular morphogenesis in vitro [17, 90] and in vivo [74, 91].

Figure 5
Vascular network formation as a function of crosslinking density for iPSC-ECs cultured in PEG hydrogels that were polymerized within a microfluidic device
Figure 6
Capillary networks with lumens for iPSC-ECs after 14 days of culture in PEG hydrogels that were polymerized within a microfluidic device

Previous studies have demonstrated that endothelial cells will self-assemble into vascular networks within engineered matrices under static conditions, but these networks often regress within a few days in the absence of mural cells such as fibroblasts, mesenchymal stem cells, or pericytes [2934]. The absence of pericytes in PDGF-B and PDGFR-β knockout mouse models leads to hyperplasia and other blood vessel abnormalities, but microvessel density, length, and branching are similar to wild type mice and the mutant animals survive into adulthood [92]. Thus, in vivo capillaries retain some level of function even in the absence of pericytes [92], which suggests that a lack of mural cells does not entirely account for regression of vascular networks in vitro. HUVECs form in vitro networks that are stable for at least five days in collagen when soluble factors are optimized [15], and up to 10 days in starPEG-heparin hydrogels that sequester VEGF, bFGF, and SDF1α during encapsulation [34]. Here, iPSC-ECs in PEG hydrogels formed capillary tubules with lumens that were stable for at least 14 days in microfluidic channels when VEGF concentration was optimized (Fig. 6). Vascular network stability could also be improved for iPSC-ECs in PEG hydrogels in standard well plates by adding a second cell-free hydrogel layer to provide mechanical support after initial cellular selfassembly (Suppl. Fig. 5). These combined results demonstrate the value of engineering approaches for extending the stability of in vitro vascular networks using endothelial cell monocultures.


The aim of the present study was to develop a vascular model by encapsulating human induced pluripotent stem cell-derived endothelial cells (iPSC-ECs) [23] in a synthetic extracellular matrix (ECM) formed via thiol-ene photopolymerization [36]. Time-lapse microscopy, immunofluorescence imaging, and RNA-sequencing (RNA-Seq) demonstrated that iPSC-ECs encapsulated in poly(ethylene glycol) (PEG) hydrogels self-assembled into capillary networks through mechanisms that are consistent with in vitro and in vivo vascular morphogenesis. Capillary tubules with lumens were stable for at least 14 days when iPSC-ECs were encapsulated in PEG hydrogels that were polymerized within a passive flow microfluidic device. Thus, the in vitro model described here mimics important aspects of vascular morphogenesis by incorporating iPSC-ECs into chemically-defined PEG hydrogels using standard and microfluidic formats.

Significance Statement

Human induced pluripotent stem cell-derived endothelial cells (iPSC-ECs) cultured in synthetic hydrogels self-assemble into capillary networks through mechanisms consistent with in vivo vascular morphogenesis.

Supplementary Material




The authors would like to acknowledge funding from the National Institutes of Health (NIH R01HL093282-01A1, R21EB016381-01, 1UH2TR000506-01, T32HL007889, T32HL07936, R01EB10039, and the Biotechnology Training Program NIGMS5T32GM08349), and the UW-Madison Graduate Engineering Research Scholars program. Mechanical testing data was obtained using the Ares LS2 rheometer at the UW-Madison Soft Materials Laboratory.


Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Competing Financial Interests

J.A.T. was a founder and stockholder for Cellular Dynamics, Inc at the time of manuscript submission. D.J.B. holds equity in Bellbrook Labs, LLC, Tasso, Inc., Stacks for the Future, LLC and Salus Discovery, LLC. W.L.M. is a founder and stockholder for Stem Pharm, Inc., and Tissue Regeneration Systems, Inc.


1. Novosel EC, Kleinhans C, Kluger PJ. Vascularization is the key challenge in tissue engineering. Adv Drug Deliv Rev. 2011;63:300–311. [PubMed]
2. Phelps EA, García AJ. Engineering more than a cell: vascularization strategies in tissue engineering. Curr Opin Biotechnol. 2010;21:704–709. [PMC free article] [PubMed]
3. Hasan A, Paul A, Vrana NE, Zhao X, Memic A, Hwang Y-S, et al. Microfluidic techniques for development of 3D vascularized tissue. Biomaterials. 2014;35:7308–7325. [PMC free article] [PubMed]
4. Schwartz MP, Hou Z, Propson NE, Zhang J, Engstrom CJ, Costa VS, et al. Human pluripotent stem cell-derived neural constructs for predicting neural toxicity. Proc Natl Acad Sci U S A. 2015;112:12516–12521. [PubMed]
5. Martins-Green M, Li Q-J, Yao M. A new generation organ culture arising from cross-talk between multiple primary human cell types. The FASEB Journal. 2005;19:222–224. [PubMed]
6. Takebe T, Sekine K, Enomura M, Koike H, Kimura M, Ogaeri T, et al. Vascularized and functional human liver from an iPSC-derived organ bud transplant. Nature. 2013;499:481-+. [PubMed]
7. Sekine H, Shimizu T, Sakaguchi K, Dobashi I, Wada M, Yamato M, et al. In vitro fabrication of functional three-dimensional tissues with perfusable blood vessels. Nat Commun. 2013;4 [PMC free article] [PubMed]
8. Scott CW, Peters MF, Dragan YP. Human induced pluripotent stem cells and their use in drug discovery for toxicity testing. Toxicol Lett. 2013;219:49–58. [PubMed]
9. Kleinstreuer NC, Yang J, Berg EL, Knudsen TB, Richard AM, Martin MT, et al. Phenotypic screening of the ToxCast chemical library to classify toxic and therapeutic mechanisms. Nat Biotechnol. 2014;32:583-+. [PubMed]
10. Sarkanen J-R, Mannerstrom M, Vuorenpaa H, Uotila J, Ylikomi T, Heinonen T. Intra-laboratory validation of a human cell based in vitro angiogenesis assay for testing angiogenesis modulators. Frontiers in Pharmacology. 2011;1 [PMC free article] [PubMed]
11. Folkman J, Haudenschild C. Angiogenesis in vitro. Nature. 1980;288:551–556. [PubMed]
12. Maciag T, Kadish J, Wilkins L, Stemerman MB, Weinstein R. Organizational behavior of human umbilical vein endothelial cells. The Journal of Cell Biology. 1982;94:511–520. [PMC free article] [PubMed]
13. Montesano R, Orci L, Vassalli P. In vitro rapid organization of endothelial cells into capillary-like networks is promoted by collagen matrices. J Cell Biol. 1983;97:1648–1652. [PMC free article] [PubMed]
14. Kubota Y, Kleinman HK, Martin GR, Lawley TJ. Role of laminin and basement membrane in the morphological differentiation of human endothelial cells into capillary-like structures. J Cell Biol. 1988;107:1589–1598. [PMC free article] [PubMed]
15. Davis GE, Camarillo CW. An alpha 2 beta 1 integrin-dependent pinocytic mechanism involving intracellular vacuole formation and coalescence regulates capillary lumen and tube formation in three-dimensional collagen matrix. Exp Cell Res. 1996;224:39–51. [PubMed]
16. Vernon RB, Sage EH. Between molecules and morphology - Extracellular matrix and creation of vascular form. Am J Pathol. 1995;147:873–883. [PubMed]
17. Davis GE, Senger DR. Endothelial extracellular matrix - Biosynthesis, remodeling, and functions during vascular morphogenesis and neovessel stabilization. CircRes. 2005;97:1093–1107. [PubMed]
18. Vickerman V, Blundo J, Chung S, Kamm R. Design, fabrication and implementation of a novel multi-parameter control microfluidic platform for three-dimensional cell culture and real-time imaging. Lab Chip. 2008;8:1468–1477. [PMC free article] [PubMed]
19. Whisler JA, Chen MB, Kamm RD. Control of perfusable microvascular network morphology using a multiculture microfluidic system. Tissue Eng Part C-Methods. 2014;20:543–552. [PMC free article] [PubMed]
20. Tourovskaia A, Fauver M, Kramer G, Simonson S, Neumann T. Tissue-engineered microenvironment systems for modeling human vasculature. Exp Biol Med. 2014;239:1264–1271. [PMC free article] [PubMed]
21. Moya ML, Hsu Y, Lee AP, Hughes CCW, George SC. In vitro perfused human capillary networks. Tiss Eng C: Methods. 2013 [PMC free article] [PubMed]
22. Kim S, Lee H, Chung M, Jeon NL. Engineering of functional, perfusable 3D microvascular networks on a chip. Lab Chip. 2013;13:1489–1500. [PubMed]
23. Belair D, Whisler J, Valdez J, Velazquez J, Molenda J, Vickerman V, et al. Human vascular tissue models formed from human induced pluripotent stem cell derived endothelial cells. Stem Cell Rev Rep. 2014:1–15. [PMC free article] [PubMed]
24. Yang YL, Leone LM, Kaufman LJ. Elastic moduli of collagen gels can be predicted from two-dimensional confocal microscopy. Biophys J. 2009;97:2051–2060. [PubMed]
25. Hughes CS, Postovit LM, Lajoie GA. Matrigel: A complex protein mixture required for optimal growth of cell culture. Proteomics. 2010;10:1886–1890. [PubMed]
26. Vukicevic S, Kleinman HK, Luyten FP, Roberts AB, Roche NS, Reddi AH. Identification of multiple active growth factors in basement membrane Matrigel suggests caution in interpretation of cellular activity related to extracellular matrix components. Exp Cell Res. 1992;202:1–8. [PubMed]
27. Zisch AH, Lutolf MP, Ehrbar M, Raeber GP, Rizzi SC, Davies N, et al. Cell-demanded release of VEGF from synthetic, biointeractive cell-ingrowth matrices for vascularized tissue growth. The FASEB Journal. 2003 [PubMed]
28. Leslie-Barbick JE, Moon JJ, West JL. Covalently immobilized vascular endothelial growth factor promotes endothelial cell tubulogenesis in poly(ethylene glycol) diacrylate hydrogels. J Biomater Sci-Polym Ed. 2009;20:1763–1779. [PubMed]
29. Moon JJ, Saik JE, Poche RA, Leslie-Barbick JE, Lee SH, Smith AA, et al. Biomimetic hydrogels with pro-angiogenic properties. Biomaterials. 2010;31:3840–3847. [PMC free article] [PubMed]
30. Kusuma S, Shen Y-I, Hanjaya-Putra D, Mali P, Cheng L, Gerecht S. Self-organized vascular networks from human pluripotent stem cells in a synthetic matrix. Proceedings of the National Academy of Sciences. 2013;110:12601–12606. [PubMed]
31. Chen YC, Lin RZ, Qi H, Yang YZ, Bae HJ, Melero-Martin JM, et al. Functional human vascular network generated in photocrosslinkable gelatin methacrylate hydrogels. Adv Funct Mater. 2012;22:2027–2039. [PMC free article] [PubMed]
32. Culver JC, Hoffmann JC, Poché RA, Slater JH, West JL, Dickinson ME. Three-dimensional biomimetic patterning in hydrogels to guide cellular organization. Adv Mater. 2012;24:2344–2348. [PMC free article] [PubMed]
33. Turturro MV, Christenson MC, Larson JC, Young DA, Brey EM, Papavasiliou G. MMP-sensitive PEG diacrylate hydrogels with spatial variations in matrix properties stimulate directional vascular sprout formation. PLoS One. 2013;8:e58897. [PMC free article] [PubMed]
34. Chwalek K, Tsurkan MV, Freudenberg U, Werner C. Glycosaminoglycan-based hydrogels to modulate heterocellular communication in in vitro angiogenesis models. Sci Rep. 2014;4 [PMC free article] [PubMed]
35. Nguyen EH, Zanotelli MR, Schwartz MP, Murphy WL. Differential effects of cell adhesion, modulus and VEGFR-2 inhibition on capillary network formation in synthetic hydrogel arrays. Biomaterials. 2014;35:2149–2161. [PMC free article] [PubMed]
36. Fairbanks BD, Schwartz MP, Halevi AE, Nuttelman CR, Bowman CN, Anseth KS. A versatile synthetic extracellular matrix mimic via thiol-norbornene photopolymerization. Adv Mater. 2009;21:5005–5010. [PMC free article] [PubMed]
37. Lin C-C, Ki CS, Shih H. Thiol–norbornene photoclick hydrogels for tissue engineering applications. J Appl Polym Sci. 2015;132:41563. [PMC free article] [PubMed]
38. Hoyle CE, Bowman CN. Thiol-ene click chemistry. Angew Chem-Int Edit. 2010;49:1540–1573. [PubMed]
39. Singh SP, Schwartz MP, Lee JY, Fairbanks BD, Anseth KS. A peptide functionalized poly(ethylene glycol) (PEG) hydrogel for investigating the influence of biochemical and biophysical matrix properties on tumor cell migration. Biomaterials Science. 2014;2:1024–1034. [PMC free article] [PubMed]
40. Gramlich WM, Kim IL, Burdick JA. Synthesis and orthogonal photopatterning of hyaluronic acid hydrogels with thiol-norbornene chemistry. Biomaterials. 2013;34:9803–9811. [PMC free article] [PubMed]
41. Munoz Z, Shih H, Lin CC. Gelatin hydrogels formed by orthogonal thiol-norbornene photochemistry for cell encapsulation. Biomaterials Science. 2014;2:1063–1072.
42. Van Hove AH, Beltejar M-JG, Benoit DSW. Development and in vitro assessment of enzymatically-responsive poly(ethylene glycol) hydrogels for the delivery of therapeutic peptides. Biomaterials. 2014;35:9719–9730. [PMC free article] [PubMed]
43. Wang C, Tong XM, Yang F. Bioengineered 3D brain tumor model to elucidate the effects of matrix stiffness on glioblastoma cell behavior using PEG-based hydrogels. Mol Pharm. 2014;11:2115–2125. [PubMed]
44. Hynes WF, Doty NJ, Zarembinski TI, Schwartz MP, Toepke MW, Murphy WL, et al. Micropatterning of 3D microenvironments for living biosensor applications. Biosensors. 2014;4:28–44. [PMC free article] [PubMed]
45. Belair DG, Murphy WL. Specific VEGF sequestering to biomaterials: Influence of serum stability. Acta Biomater. 2013;9:8823–8831. [PMC free article] [PubMed]
46. Lin C-C, Raza A, Shih H. PEG hydrogels formed by thiol-ene photo-click chemistry and their effect on the formation and recovery of insulin-secreting cell spheroids. Biomaterials. 2011;32:9685–9695. [PMC free article] [PubMed]
47. Anderson SB, Lin CC, Kuntzler DV, Anseth KS. The performance of human mesenchymal stem cells encapsulated in cell-degradable polymer-peptide hydrogels. Biomaterials. 2011;32:3564–3574. [PMC free article] [PubMed]
48. Chi J-T, Chang HY, Haraldsen G, Jahnsen FL, Troyanskaya OG, Chang DS, et al. Endothelial cell diversity revealed by global expression profiling. Proceedings of the National Academy of Sciences. 2003;100:10623–10628. [PubMed]
49. Yu JY, Vodyanik MA, Smuga-Otto K, Antosiewicz-Bourget J, Frane JL, Tian S, et al. Induced pluripotent stem cell lines derived from human somatic cells. Science. 2007;318:1917–1920. [PubMed]
50. Takahashi K, Tanabe K, Ohnuki M, Narita M, Ichisaka T, Tomoda K, et al. Induction of pluripotent stem cells from adult human fibroblasts by defined factors. Cell. 2007;131:861–872. [PubMed]
51. Thomson JA, Itskovitz-Eldor J, Shapiro SS, Waknitz MA, Swiergiel JJ, Marshall VS, et al. Embryonic stem cell lines derived from human blastocysts. Science. 1998;282:1145–1147. [PubMed]
52. Samuel R, Daheron L, Liao S, Vardam T, Kamoun WS, Batista A, et al. Generation of functionally competent and durable engineered blood vessels from human induced pluripotent stem cells. Proc Natl Acad Sci U S A. 2013;110:12774–12779. [PubMed]
53. Levenberg S, Golub JS, Amit M, Itskovitz-Eldor J, Langer R. Endothelial cells derived from human embryonic stem cells. Proc Natl Acad Sci U S A. 2002;99:4391–4396. [PubMed]
54. Pitrez PR, Rosa SC, Praça C, Ferreira L. Vascular disease modeling using induced pluripotent stem cells: Focus in Hutchinson-Gilford Progeria Syndrome. Biochem Biophys Res Commun. 2015 [PubMed]
55. Bischel LL, Young EWK, Mader BR, Beebe DJ. Tubeless microfluidic angiogenesis assay with three-dimensional endothelial-lined microvessels. Biomaterials. 2013;34:1471–1477. [PMC free article] [PubMed]
56. Beebe DJ, Mensing GA, Walker GM. Physics and applications of microfluidics in biology. Annu Rev Biomed Eng. 2002;4:261–286. [PubMed]
57. Jo BH, Van Lerberghe LM, Motsegood KM, Beebe DJ. Three-dimensional microchannel fabrication in polydimethylsiloxane (PDMS) elastomer. J Microelectromech Syst. 2000;9:76–81.
58. Nagase H, Fields GB. Human matrix metalloproteinase specificity studies using collagen sequence-based synthetic peptides. Biopolymers. 1996;40:399–416. [PubMed]
59. Netzel-Arnett S, Sang QX, Moore WGI, Navre M, Birkedal-Hansen H, Van Wart HE. Comparative sequence specificities of human 72- and 92-kDa gelatinases (type IV collagenases) and PUMP (matrilysin) Biochemistry. 1993;32:6427–6432. [PubMed]
60. Pierschbacher MD, Ruoslahti E. Cell attachment activity of fibronectin can be duplicated by small synthetic fragments of the molecule. Nature. 1984;309:30–33. [PubMed]
61. Thevenaz P, Ruttimann UE, Unser M. A pyramid approach to subpixel registration based on intensity. IEEE Trans Image Process. 1998;7:27–41. [PubMed]
62. Rasband WS. Image J. Bethesda, Maryland, USA: U.S. National Institutes of Health; 1997–2012.
63. Schneider CA, Rasband WS, Eliceiri KW. NIH Image to ImageJ: 25 years of image analysis. Nat Meth. 2012;9:671–675. [PubMed]
64. Langmead B, Trapnell C, Pop M, Salzberg S. Ultrafast and memory-efficient alignment of short DNA sequences to the human genome. Genome Biol. 2009;10:R25. [PMC free article] [PubMed]
65. Li B, Dewey CN. RSEM: accurate transcript quantification from RNA-Seq data with or without a reference genome. BMC Bioinformatics. 2011;12:323. [PMC free article] [PubMed]
66. Leng N, Dawson JA, Thomson JA, Ruotti V, Rissman AI, Smits BMG, et al. EBSeq: an empirical Bayes hierarchical model for inference in RNA-seq experiments. Bioinformatics. 2013;29:1035–1043. [PMC free article] [PubMed]
67. Huang DW, Sherman BT, Lempicki RA. Bioinformatics enrichment tools: paths toward the comprehensive functional analysis of large gene lists. Nucleic Acids Research. 2009;37:1–13. [PMC free article] [PubMed]
68. Huang DW, Sherman BT, Lempicki RA. Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat Protocols. 2008;4:44–57. [PubMed]
69. Consortium TGO. Gene Ontology Consortium: going forward. Nucleic Acids Research. 2015;43:D1049–D1056. [PMC free article] [PubMed]
70. Ashburner M, Ball CA, Blake JA, Botstein D, Butler H, Cherry JM, et al. Gene Ontology: tool for the unification of biology. Nature Genet. 2000;25:25–29. [PMC free article] [PubMed]
71. Czirok A, Zamir EA, Szabo A, Little CD. Multicellular sprouting during vasculogenesis. In: Santiago Schnell PKMSAN, Timothy JN, editors. Current Topics in Developmental Biology. Academic Press; 2008. pp. 269–289. [PMC free article] [PubMed]
72. Rupp PA, Czirók A, Little CD. αvβ3 integrin-dependent endothelial cell dynamics in vivo. Development. 2004;131:2887–2897. [PubMed]
73. Kamei M, Saunders WB, Bayless KJ, Dye L, Davis GE, Weinstein BM. Endothelial tubes assemble from intracellular vacuoles in vivo. Nature. 2006;442:453–456. [PubMed]
74. Herbert SP, Stainier DYR. Molecular control of endothelial cell behaviour during blood vessel morphogenesis. Nat Rev Mol Cell Biol. 2011;12:551–564. [PMC free article] [PubMed]
75. Ramjaun AR, Hodivala-Dilke K. The role of cell adhesion pathways in angiogenesis. The International Journal of Biochemistry & Cell Biology. 2009;41:521–530. [PubMed]
76. Roca C, Adams RH. Regulation of vascular morphogenesis by Notch signaling. Genes Dev. 2007;21:2511–2524. [PubMed]
77. Kleinman HK, McGarvey ML, Liotta LA, Robey PG, Tryggvason K, Martin GR. Isolation and characterization of Type IV procollagen, laminin, and heparin sulfate proteoglycan from the EHS sarcoma. Biochemistry. 1982;21:6188–6193. [PubMed]
78. Timpl R, Rohde H, Robey PG, Rennard SI, Foidart JM, Martin GR. Laminin - Glycoprotein from basement membranes. J Biol Chem. 1979;254:9933–9937. [PubMed]
79. Rand JH, Wu XX, Potter BJ, Uson RR, Gordon RE. Colocalization of von Willebrand factor and Type VI collagen in human vascular subendothelium. Am J Pathol. 1993;142:843–850. [PubMed]
80. Rand JH, Patel ND, Schwartz E, Zhou SL, Potter BJ. 150 kD von Willebrand factor binding protein extracted from human vascular subendothelium is Type VI collagen. J Clin Invest. 1991;88:253–259. [PMC free article] [PubMed]
81. Tulla M, Pentikäinen OT, Viitasalo T, Käpylä J, Impola U, Nykvist P, et al. Selective Binding of Collagen Subtypes by Integrin α1I, α2I, and α10I Domains. J Biol Chem. 2001;276:48206–48212. [PubMed]
82. Iruela-Arispe ML, Hasselaar P, Sage H. Differential expression of extracellular proteins is correlated with angiogenesis in vitro. Lab Invest. 1991;64:174–186. [PubMed]
83. Iruela-Arispe ML, Diglio CA, Sage EH. Modulation of extracellular matrix proteins by endothelial cells undergoing angiogenesis in vitro. Arteriosclerosis and Thrombosis. 1991;11:805–815. [PubMed]
84. Fouser L, Iruelaarispe L, Bornstein P, Sage EH. Transcriptional activity of the alpha1(I) collagen promoter is correlated with the formation of capillary-like structures by endothelial cells in vitro. J Biol Chem. 1991;266:18345–18351. [PubMed]
85. Myers JC. Differential expression of type I collagen and cellular fibronectin isoforms in endothelial cell variants. Kidney Int. 1993;43:45–52. [PubMed]
86. Senger DR, Perruzzi CA, Streit M, Koteliansky VE, de Fougerolles AR, Detmar M. The alpha(1)beta(1) and alpha(2)beta(1) Integrins provide critical support for vascular endothelial growth factor signaling, endothelial cell migration, and tumor angiogenesis. Am J Pathol. 2002;160:195–204. [PubMed]
87. Senger DR, Claffey KP, Benes JE, Perruzzi CA, Sergiou AP, Detmar M. Angiogenesis promoted by vascular endothelial growth factor: Regulation through alpha(1)beta(1) and alpha(2)beta(1) integrins. Proc Natl Acad Sci U S A. 1997;94:13612–13617. [PubMed]
88. Bell SE, Mavila A, Salazar R, Bayless KJ, Kanagala S, Maxwell SA, et al. Differential gene expression during capillary morphogenesis in 3D collagen matrices: regulated expression of genes involved in basement membrane matrix assembly, cell cycle progression, cellular differentiation and G-protein signaling. J Cell Sci. 2001;114:2755–2773. [PubMed]
89. Meyvantsson I, Warrick JW, Hayes S, Skoien A, Beebe DJ. Automated cell culture in high density tubeless microfluidic device arrays. Lab Chip. 2008;8:717–724. [PubMed]
90. Pepper MS, Ferrara N, Orci L, Montesano R. Potent synergism between vascular endothelial growth factor and basic fibroblast growth factor in the induction of angiogenesis in vitro. Biochem Biophys Res Commun. 1992;189:824–831. [PubMed]
91. Connolly DT, Heuvelman DM, Nelson R, Olander JV, Eppley BL, Delfino JJ, et al. Tumor vascular permeability factor stimulates endothelial cell growth and angiogenesis. J Clin Invest. 1989;84:1470–1478. [PMC free article] [PubMed]
92. Hellström M, Gerhardt H, Kalén M, Li X, Eriksson U, Wolburg H, et al. Lack of pericytes leads to endothelial hyperplasia and abnormal vascular morphogenesis. The Journal of Cell Biology. 2001;153:543–554. [PMC free article] [PubMed]