Skip to main content
NCBI home page
Springer logoLink to Springer
. 2026 Oct 14;30(12):91. doi: 10.1007/s10404-026-02939-3

Spatial distribution and density of fibroblasts determine angiogenic response of endothelial cells

Pu Zhang 1,2, Melur K Ramasubramanian 2,✉, Li-Heng Cai 1,3,4,5,✉
PMCID: PMC13668921  PMID: 42889306

Abstract

Fibroblasts play indispensable roles in orchestrating revascularization and angiogenesis of endothelial cells. However, the specific conditions under which the influence of fibroblasts becomes important in modulating angiogenic formation remain to be determined. Here, we extend a previously established microfluidic co-culture system to enable precise manipulation of the density of human lung fibroblasts (HLF) and their distance to human umbilical vein endothelial cells (HUVECs) in a wide range within fibrin hydrogels. Using live cell imaging and image analysis, we quantify the effects of fibroblast density, HLF- HUVEC cell–cell distance, and culturing time on cell migration, morphology, microvasculature, and endothelial sprouting. We find that, within the system and parameter range explored here, for fibroblasts to have a significant positive impact in the angiogenesis of endothelial cells, the fibroblast density should be no less than ~3×106 cells/ml and the cell–cell distance should be no greater than ~2 mm. The experimental findings can be captured by a minimal model accounting for the concentration profile of pro-angiogenic factors secreted by fibroblasts. Our results provide insights into the remodeling of vasculature and the guidance for engineering vascularized tissue mimics.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s10404-026-02939-3.

Keywords: Microfluidic device, Angiogenesis, Revascularization, Fibrin hydrogel, Angiogenic response, Endothelial

Introduction

Angiogenesis, the formation of new blood vessels from existing vasculature, is essential for various biological processes including tissue regeneration (Jahani et al. 2020; Tonnesen et al. 2000), wound healing (Tonnesen et al. 2000; Darland and D’Amore 2001), and cancer progression (Liu et al. 2023; Kalluri 2016). As the primary cellular component of blood vessels, endothelial cells line the inner surface of vessels to support tissue growth and repair. While endothelial cells are the primary drivers of vascular formation, their behavior is profoundly influenced by surrounding stromal cells, particularly fibroblasts, which modulate angiogenesis through biochemical signaling and extracellular matrix remodeling (Hartlapp et al. 2001; Stupack and Cheresh 2002; Soncin 1992; Dellaquila et al. 2023). Coordinated interactions among endothelial cells, mural cells, and stromal tissues govern sprouting, migration, lumen formation, and vessel stabilization during vascular development (Darland and D’Amore 2001). In parallel, ECM remodeling modulates angiogenesis by altering biochemical composition and mechanical properties, including growth factor sequestration, matrix stiffness, and bioactive matrix fragments that influence endothelial proliferation and sprouting (Sottile 2004). In vitro co-culture models further demonstrate the pro-angiogenic role of fibroblasts. For instance, by pre-mixing human umbilical vein endothelial cells (HUVECs) and human lung fibroblasts (HLF) and culturing them within a fibrin gel, an early co-culture system showed that fibroblasts could enhance endothelial sprouting and the formation of capillary networks via paracrine signaling mechanisms (Chen et al. 2009). Moreover, fibroblasts can secrete matrix proteins to stiffen fibrin gel that correlates with enhanced lumen formation (Newman et al. 2011). However, these co-culture systems have limited control over the proximity of fibroblasts to endothelial cells.

Unlike conventional co-culture systems, microfluidic platforms allow patterning fibroblasts and endothelial cells at controlled distances within hydrogels (Kim et al. 2013; Ahn et al. 2021; Chen et al. 2023; Offeddu et al. 2019; Whisler et al. 2014; Zheng et al. 2012). This lab-on-chip platform enables engineering perfusable three-dimensional (3D) microvascular networks and tumor vasculatures based on the spatially controlled co-culture of endothelial cells with other cell types in fibrin gels (Kim et al. 2013). However, in this classic microfluidic co-culture system, two fibroblast channels symmetrically distribute on the two sides of a HUVECs channel (Kim et al. 2013). This double-sided geometric configuration inevitably creates a non-monotonic distribution in fibroblasts secreted pro-angiogenic factors with concentration peaks at location with equal distance from the two fibroblast channels. This may complicate the understanding of effects of HLF-HUVEC cell–cell distance on the angiogenic response of HUVECs, which is known to be sensitive to the concentration and distribution of pro-angiogenic factors. Additionally, only a limited range of fibroblast density (5 to 8 ×106 cells/ml) and HLF-HUVEC cell–cell distance (500 to 1000 µm) have been explored. Consequently, quantitative data linking specific fibroblast densities and spatial configurations to angiogenic efficiency remains limited.

Here, we extend the classic double-sided microfluidic co-culture device to a single-sided system (Fig. 1a) to achieve a monotonic gradient profile of fibroblasts secreted pro-angiogenic factors (Fig. 1b). We systematically vary the fibroblast density within a wide range from 1 to 10 ×106 cells/ml and the HLF-HUVEC cell–cell distance from 1 to 10 mm. Using live cell imaging and image analysis, we quantify the dependence of cell migration, morphology, microvasculature, and endothelial sprouting on fibroblast density, cell–cell distance, and culturing time. We find that, within the system and parameter range explored here, for fibroblasts to have a significant positive impact in the angiogenesis of endothelial cells, the fibroblast density should be no less than ~3×106 cells/ml and the cell–cell distance should be no greater than ~2 mm. Accounting for the concentration profile of pro-angiogenic factors secreted by fibroblasts, we develop a minimal model that successfully captures the dependence of sprout length on fibroblast density and cell–cell distance. Our results provide insights into the remodeling of vasculature and the guidance for engineering vascularized tissue mimics.

Fig. 1.

Fig. 1

Open in a new tab

A microfluidic device for studying the formation of microvascular networks regulated by the proximity and cell density of fibroblasts. a A tile-scan image of the microfluidic device. The cell–cell distance d and fibroblast density ρ can be independently controlled to study their roles in the formation of the microvascular network and sprouts. From left to right channels: 1M for culture medium, 2H for HUVECs, 3G for fibrin hydrogel only, 4M for culture medium, and 5F for fibroblasts. b A schematic that shows the difference in the concentration profile of growth factors between the classical double-sided and our single-sided cell seeding configuration. c A schematic of the microfluidic device in (a) to illustrate the formation of microvascular network and sprouts from the endothelial cell channel. d Tile scan fluorescence confocal images of pre-vascularized network and angiogenic sprouts in microfluidic channels. (i) Microvascular network within the endothelial channel, (ii) and angiogenic sprouting inside the Channel 3G after 48 h, using a 1 mm wide acellular channel and seeding density of 3×106 cells/ml for HUVEC and 5×106 cells /ml for HLF. e Fluorescent microbeads of 1 µm in diameter are introduced from Channel 1 M and transport within the microvascular network. The trajectories of these microbeads, represented by yellow lines, illustrate their transport paths through the interconnected luminal spaces of the vascular structure

Results and discussion

Design and fabrication of a microfluidic device for studying microvasculature

Fibroblasts secrete a variety of pro-angiogenic factors, such as vascular endothelial growth factor (VEGF), basic fibroblast growth factor (bFGF), platelet-derived growth factor (PDGF), and stromal-derived factor 1 (SDF-1), which collectively stimulate endothelial cell proliferation, migration, and tube formation (Carmeliet and Jain 2011; Margolis et al. 2021; Salcedo et al. 1999; Salvucci et al. 2002). Thus, the formation of vasculature is determined by both the concentration and the gradient of pro-angiogenic factors sensed by HUVECs (Johnson and Wilgus 2014; Wang et al. 2020). These two parameters are expected to be dependent on both (i) the distance between HUVECs and HLF and (ii) the density of HLF. To explore this, we design a microfluidic device that consists of five adjacent channels, which are consecutively loaded with cell culture medium (Channel 1 M), HUVECs (Channel 2H), fibrin gel (Channel 3G), cell culture medium (Channel 4 M), and HLF (Channel 5F), as illustrated from left to right in Fig. 1a. The dimension of the channels is fixed at 1 mm in width (except for fibrin gel Channel 3G) and 150 µm in height, large enough for monitoring the angiogenesis within a few days. Moreover, the medium reservoir has a diameter of 6 mm and a height of 1.5 mm (Fig. S1), sufficient for supplying nutrients for cell growth for at least 48 h, by which the culture medium is changed. Note that our design is inspired by but different from the classic microfluidic devices, which have two stromal fibroblast channels symmetrically distributed on the two sides of Channel 2H (Kim et al. 2013); this geometric configuration inevitably creates non-monotonic distribution in fibroblasts secreted pro-angiogenic factors with concentration peaks at location with equal distance from the two fibroblast channels, as illustrated by the dashed line in Fig. 1b. By contrast, our design uses a single fibroblast channel to achieve a gradient profile of fibroblasts secreted pro-angiogenic factors (red line in Fig. 1b). Moreover, the concentration of pro-angiogenic factors at the Channel 2H can be tuned by either the concentration of the HLF or the width of the acellular fibrin gel Channel 3G that separates the HUVECs and the HLF.

In the microfluidic device, instead of using relatively thick microscope slides, we use a thin cover slip as the substrate. This allows using lens with relatively high magnification and short working distance to obtain high-resolution images of cell morphology. Additionally, we pre-inject the microfluidic channels with carbon dioxide (CO₂), which can dissolve into water; this avoids trapped air bubbles when loading the fibrin gel and cells.

Besides the channels for cell culture medium, all channels are loaded with fibrin gel, which is formed by thrombin-catalyzed conversion of fibrinogen to fibrin, a process mimicking that in blood clotting. For the channel containing cells, the cells are pre-mixed with the fibrin gel at a prescribed cell density. The concentration of the fibrin gel is fixed at 2.5 mg/ml, corresponding to 7.4 µM, at which the diameter of the fibrin fiber is about Inline graphic (Xia et al. 2021). Since fibrin is a semiflexible biopolymer, the mesh size of the fibrin gel is determined by the fibrin volume fraction ϕ and the fiber diameter d: Inline graphic (MacKintosh et al. 1995). This pore size is large enough to ensure sufficient transport of nutrients and oxygen for cell growth and migration. Moreover, the concentration of thrombin within the fibrinogen is 1 U/ml (~ 200 nM), resulting in a crosslinked fibrin gel with a shear modulus 200 Pa, as shown by the small amplitude oscillatory shear measurements in Fig. S2. This stiffness is high enough to mechanically support cell infiltration and the formation of capillary-like structure (Morin and Tranquillo 2013; Moura et al. 2024) (Fig. 1c). Additionally, the degradation rate of the fibrin gel of this concentration is fast enough to accommodate matrix remodeling during tissue formation (Ahmann et al. 2010). Indeed, our microfluidic setup allows for vasculogenic development, as exemplified by a tile scan fluorescence image of the whole device, which shows a microvascular network (μVN) (panel i in Fig. 1d) formed by HUVECs expanding into the adjacent channel filled with fibrin gel only (panel ii in Fig. 1d). We note that fibrin gels exhibit batch-to-batch variability and possess inherent translational and mechanistic limitations. The specific gel composition and mechanical properties used here are selected to provide a reproducible and pro-angiogenic environment for controlled spatial interrogation. Extension of this framework to alternative natural or synthetic matrices will be important for future translational studies.

To further validate the functionality of the engineered microvascular networks, we assess their perfusability using 1 µm fluorescent microbeads under hydrostatic pressure. Specifically, we load a suspension of microbeads into the microfluidic device from one of the two medium reservoirs for HUVECs. Within a few seconds, the microbeads enter the vascular network through the medium Channel 1M and transport through the interconnected vessel structures, as shown by Fig. 1e and Movie S1. Moreover, tracking the bead trajectories confirms that the microbeads are confined within the vessel lumen, as shown by the yellow lines in Fig. 1e. These results demonstrate that our microfluidic co-culture system supports the formation of hollow, interconnected microvasculature in vitro. Future studies will evaluate the functional perfusability of these networks using established perfusion assays, such as fluorescent dextran or microsphere perfusion.

Effects of cell–cell distance on microvasculature

We define cell–cell distance as the separation distance between HUVEC and HLF cell populations. To explore the effects of cell–cell distance on the formation of microvasculature, we fix the concentrations of HUVECs and fibroblasts, respectively, at 3 and 5 ×106 cells per ml, comparable to those in existing studies (Kim et al. 2016). We vary the width of fibrin gel Channel 3G, or the distance between the HUVEC and fibroblast channels, in a wide range from 1 to 10 mm, which is greater than the typical distance 1 mm explored in literature (Kim et al. 2016). The cells are fluorescently labeled by cell tracker for live cell imaging without impairing their biological activity (Ladaigue et al. 2023). Since cell trackers typically allow for live cell imaging for up to 3 days, following initial staining, the cells are re-stained to extend the 5-day experiment window.

Within the Channel 2H, after being cultured for 48 h, the HUVECs start to form microvascular network regardless of the distance from the fibroblasts, as shown by the representative fluorescence images on the left column of Fig. 2a. We use total sprout area as the primary angiogenic metric because it provides a robust, non-directional measure of global endothelial invasion driven by fibroblast-secreted growth factor gradients. This metric integrates collective cell migration, proliferation, and matrix remodeling into a single quantitative descriptor, minimizing variability arising from local morphological heterogeneity. Quantitatively, the normalized area of microvasculature, defined as the area of microvascular network normalized by the total area of the HUVEC channel, is about 25% and exhibits no significant difference across various cell–cell distance (Fig. 2b). This behavior is consistent with existing understanding that fibrin gel provides a favorable environment for vasculogenesis of HUVECs (Moura et al. 2024). By contrast, the ability of HUVECs to invade the adjacent empty fibrin gel Channel 3G strongly depends on the cell–cell distance. For instance, at 48 h, only for the case with 1 mm cell–cell distance, HUVECs exhibit some extent of angiogenesis behavior by migrating into Channel 3G to form a network (second column to the left in Fig. 2a). The difference in angiogenesis becomes dramatic after 120 h, at which the normalized sprouting area, defined as the area of the sprouts divided by the total area of the sprouting region, decreases significantly from 10 to 2% as the cell–cell distance increases from 1 to 5 mm, as visualized by the right column in Fig. 2a and quantified in Fig. 2c.

Fig. 2.

Fig. 2

Open in a new tab

Vasculogenesis and angiogenesis at various cell–cell distances. The concentrations of HUVECs and fibroblasts are fixed at 3×106 cells/ml and 5×106 cells/ml, respectively. By contrast, the distance between the HUVEC and fibroblast channels varies from 1 to 5 mm. a Representative fluorescence confocal images of the vasculogenesis and angiogenesis channels over 120 h. b Normalized microvascular network area in the vasculogenic channel at 48 h. c Normalized area of sprouts, defined as area of the sprouts divided by the total area of the sprouting region, at 120 h. d Dependence of normalized area of sprouts on culturing time for various cell–cell distances. Data are presented as mean ± SD (n = 4 independent microfluidic devices per condition). Each data point represents the average of four ROIs from a single device. *p<0.05; **p<0.01; ***p<0.001; n.s., not significant

To further explore the angiogenic sprouting in response to the cell–cell distance, we quantify the dependence of the normalized sprouting area on culturing time for various cell–cell distances (Fig. 2d). Interestingly, the normalized sprouting area increases nearly linearly with culturing time. As the cell–cell distance increases by five times from 1 to 5 mm, the normalized sprouting rate, defined as the normalized sprouting area divided by culturing time, decreases by nearly 7 times from 8.0×10–4 to 1.2×10–4 per hour. Consistent with this observation, the sprout diameter increases with the decrease in cell–cell distance (Fig. S4a). However, further increasing the cell–cell distance to 10 mm does not decrease the sprouting area (Fig. S3). These phenomena are likely because of the concentration of angiogenic factors secreted by fibroblasts decreases with the increase of cell–cell distance, and that at 10 mm distance, the concentration becomes too low to induce cell response. These results justify that the largest separation condition serves as an effective no fibroblast control.

Effects of fibroblast density on microvasculature

To explore the effects of fibroblast density on vasculogenesis and angiogenesis, we fix the cell–cell distance at 1 mm and the concentration of HUVECs at 3×106 cells/ml, while changing the concentration of HLF from 0 to 10 ×106 cells/ml. Higher HLF concentrations (> 10×106 cells/ml) markedly increase the viscosity of cell suspension, posing challenges in loading the cells into the microfluidic channels. Nevertheless, the range of cell density explored here is wider than 5 to 8 ×106 cells/ml in previous studies (Kim et al. 2013, 2016; Du et al. 2017).

Interestingly, vasculogenesis of HUVECs within the fibrin gel is nearly independent of the HLF concentration. By contrast, the process of angiogenesis strongly depends on HLF concentration. For instance, at 48 h, the normalized area of microvasculature is nearly constant (24±6)% regardless of the HLF concentration (fluorescence confocal images on the left of Fig. 3a and data in Fig. 3c). By contrast, at 120 h, the normalized sprouting area increases by nearly eleven times from (2±1)% to (23±4)% as the HLF concentration increases from 1 to 10 ×106 cells/ml (second column to the left in Fig. 3a and data in Fig. 3d). Yet, there is no significant difference in sprouting between the 1 ×106 cells/ml and the negative control without HLF.

Fig. 3.

Fig. 3

Open in a new tab

Effects of fibroblast density on microvasculature. The distance between the HUVEC and fibroblast channels is fixed at 1 mm, and the seeding cell density of HUVECs is maintained at 3×10⁶ cells/ml. By contrast, the density of fibroblasts is varied from 0 (negative control, n.c.) to 10×10⁶ cells/ml. a Representative fluorescence images for various fibroblast densities over 120 h. b Immunofluorescence images of angiogenic sprouts (HUVECs, red) and the proliferation of fibroblasts (green) within the fibrin gel Channel 3G. c The normalized microvascular network area at 48 h. d The normalized sprouting area at 120 h. e Dependence of the normalized area of sprouts on culturing time. Dashed lines: The sprouting rate for various cell densities. Data are presented as mean ± SD (n = 4 independent microfluidic devices per condition). Each data point represents the average of four ROIs from a single device. *p<0.05; **p<0.01; ***p<0.001; n.s., not significant

The effects of HLF cell density on angiogenesis are further confirmed by the sprouting rate. Over the 5-day culturing period, for the HLF cell densities explored, the normalized sprouting area increases linearly with culturing time (Fig. 3e). As the concentration of HLF cells decreases by nearly two times from 10 to 5 ×106 cells/ml, the normalized sprouting rate decreases slightly from 9.7×10–4/hr to 8.6×10–4/hr (red and blue dashed lines in Fig. 3e). For HLF density of 5 ×106 cells/ml, the sprouting rate (8.6×10–4/hr) is consistent with independent measurements (8.0×10–4/hr), as shown by the red circles in Fig. 2d. Further reducing the cell density to 1×106 cells/ml leads to a sprouting rate of 0.2×10–4/hr, this sprouting rate has no significant difference compared to the control group without HLF. Similar trend is observed for the dependence of sprout diameter on cell density (Fig. S4b). These observations are consistent with the understanding that, below a threshold cell density, the concentration of signaling molecules secreted by HLF is insufficient for cell–cell communications that rely on concentration-sensitive signaling pathways (Zhou et al. 2022; Dudley and Griffioen 2023). Taken together, our results unambiguously show that the threshold density of HLF to promote angiogenesis of HUVECs is approximately 1×106 cells/ml.

Remarkably, for the HLF with the highest density of 10×106 cells/ml, at relatively long culture time of Day 5, we observe a nonuniform cell density within the fibrin gel Channel 3G. Compared to that in the middle of Channel 3G, the cell density is relatively high with more cells located at the left and right edges of the channel, which are, respectively, adjacent to the HUVECs Channel 2H and the culture medium Channel 4M. To further explore this behavior, we stain the F-actin and nuclei of all cells using phalloidin and Hoechst 33342, respectively. Moreover, we identify endothelial cells using a red fluorescent molecule tagged mouse monoclonal antibody against CD31, a transmembrane glycoprotein expressed on endothelial cells (Fig. 3b). This dual immunostaining reveals that the endothelial cells are mainly located at the side adjacent to the HUVECs Channel 2H. However, the cells adjacent to the culture medium Channel 4M are fibroblasts. These results suggest that the fibroblasts are not confined within their own channel at relatively long culture times. Instead, they migrate across the culture medium Channel 4M of 1 mm in width, which is nearly 50 times of the cell size, and enter the fibrin gel Channel 3G adjacent to Channel 2H. This behavior may arise from matrix deposition or surface conditioning that enhances adhesion along PDMS surfaces. The presence of fibroblasts within the gel is expected to increase the local concentration of fibroblast-derived pro-angiogenic factors near the endothelial sprouts (Otsu et al. 2009; Barkefors et al. 2008). In addition, because migrating fibroblasts are brought into closer proximity to endothelial cells, we cannot exclude the possibility that direct fibroblast–endothelial interactions also contribute to the enhanced sprouting observed under this extreme condition. Consistent with these effects, the highest fibroblast density exhibits a notably high angiogenic sprouting rate (9.7×10–4/hr) (Fig. 3e). By contrast, at fibroblast density below 5×106 cells/ml, we do not observe fibroblast invasion into the gel, yet sprouting remains strongly dependent on fibroblast density (Fig. 3e). This observation is consistent with paracrine signaling being the dominant mechanism over most of the parameter space investigated. These results show that a threshold cell density is necessary to sustain sufficient pro-angiogenic signaling, highlighting the synergistic effects of fibroblasts invasion, local biochemical signals, and effective angiogenesis.

Modeling sprout length induced by fibroblasts

We develop a phenomenological model to describe the effects of fibroblast density and cell–cell distance on sprout length. To construct the phenomenological model, we quantify sprout length rather than sprout area at later stages of angiogenesis (Day 5). At this time point, endothelial sprouts exhibit defined, elongated morphology, and sprout length measured perpendicular to the cell channels directly reflects directional extension toward fibroblast-derived signals. In contrast, sprout area is more appropriate for early-stage analysis (Day 1–3), where cell invasion is collective and individual sprouts are not yet well defined.

The sprout length is proportional to migration speed of endothelial cells, which is determined by the concentration of pro-angiogenic signaling factors secreted by the fibroblasts. For a fixed cell–cell distance, the higher the density of fibroblasts, the higher the concentration of pro-angiogenic factors. However, for living cells, the concentration of secreted molecules is not linearly proportional to the cell density. Instead, the dependence of the concentration of secreted molecules typically consists of three phases: slow initial increase, rapid exponential increase, and eventually saturation due to limited resources for cell growth and activity. This behavior is often described by sigmoid function (Neurohr and Amon 2020):

graphic file with name d33e696.gif 1

Here, ℓ is the sprout length, ρ is the density of fibroblasts, Inline graphic is the crossover cell density above which the HLF start to promote the angiogenesis, Inline graphic is the characteristic cell density of the exponential growth, and Inline graphic describes the characteristic sprout length at certain cell–cell distance d.

For a fixed fibroblast density, as the cell–cell distance d increases, the concentration of secreted pro-angiogenic factors decreases, and so does the sprout length. However, the sprout length remains finite even without the presence of fibroblasts, as HUVEC cells themselves can proliferate and migrate within fibrin gel, resulting in vasculogenesis and then angiogenesis (Morin and Tranquillo 2013; Peterson et al. 2014). Moreover, as the cell–cell distance is above a characteristic value, the impact of fibroblasts on the angiogenesis of endothelial cells is expected to diminish, which often follows an exponential decay (Vilanova et al. 2018; Francis and Palsson 1997). Thus, we use an offset exponential function to describe the impact of cell–cell distance on sprout length:

graphic file with name d33e746.gif 2

Here, Inline graphic is the sprout length intrinsic to the proliferation and migration of endothelial cells themselves, Inline graphic is the maximum sprout length when the two cell types are in contact, and Inline graphic is the characteristic cell–cell distance below which fibroblasts strongly impact the behavior of endothelial cells. Note that we ignore the contribution of the medium channel (1 mm in width) to the cell–cell distance d, as the diffusion coefficient of pro-angiogenic factors such as VEGF in culture medium (~ 10–9 m2/s) is about two orders of magnitude higher than that in fibrin gel (~ 10–11 m2/s) (Bang et al. 2022; Li and Mooney 2016). Thus, it is reasonable to assume that the concentration of fibroblasts secreted pro-angiogenic factors is uniform within the fibroblast and medium channels, and that the cell–cell distance is largely determined by the width of the fibrin gel Channel 3G.

We determine the coefficients of the models by fitting them to the two sets of experimental data.

Our experiments show that for a fixed fibroblast density Inline graphic cells/ml, the sprout length is nearly constant 0.2 mm at cell–cell distance d≥ 5 mm. This gives Inline graphic mm and reduces Eq. (2) to a two-parameter model. The model describes the experimental data remarkably well (Fig. 4a) with fitting parameters of Inline graphic mm and Inline graphic mm. These results suggest that the cell–cell distance should be no more than 2 mm for fibroblasts to have a significant positive impact on angiogenesis of endothelial cells.

Fig. 4.

Fig. 4

Open in a new tab

Modeling of sprout length as a function of cell–cell distance and fibroblasts seeding density. a Sprout length, ℓ, as a function of cell–cell distance, d. Symbols: data points represent mean values with error bars indicating standard deviations. Solid line: Weighted fit using sigmoid function [Eq. (2)]: Inline graphic mm, Inline graphic ×106 cells/ml, and Inline graphic ×106 cells/ml. b Sprout length as a function of cell density, ρ . Solid line: Weighted fit using Eq. (1) with fitting parameters Inline graphic mm and Inline graphic mm. Inline graphic mm is the sprout length intrinsic to the proliferation and migration of endothelial cells themselves. c Diagram-of-state for the dependence of the sprout length on cell density and cell–cell distance. Filled symbols: our data; empty symbols: data from literature with cell–cell distance of 0.7 mm

Similarly, our minimal model [Eq. (1)] describes well the experimental data on the dependence of sprout length ℓ on fibroblast density ρ for a fixed cell–cell distance (Fig.4b). The best fit gives Inline graphic mm, Inline graphic ×106 cells/ml, and Inline graphic ×106 cells/ml. These results indicate that for a fixed cell–cell distance d=1 mm, the maximum sprout length that can be achieved by increasing fibroblast density is Inline graphic mm. Moreover, if the fibroblasts concentration is above Inline graphic ×106 cells/ml, the effects of fibroblast density saturate and thus the sprout length will be maximized. Consistent with this understanding, Eq. (2) predicts that for Inline graphic cells/ml and d=1 mm, the sprout length will be Inline graphic mm, comparable to the maximum sprout length predicted by Eq. (1).

Note that our phenomenological models are based on limited conditions explored in this study, and extension to additional endothelial–stromal combinations would further test the model. The sigmoidal dependence on fibroblast density and the exponential decay with separation distance are motivated by the well-established dose–response behavior of paracrine signaling and diffusion-limited attenuation of signaling over distance, respectively. The concentrations of fibroblast-secreted pro-angiogenic factors are not directly measured. Among the model parameters, only the baseline sprout length, Inline graphic mm [Eq. (2)], is directly determined from the experimental data as the plateau value at large separation distances. Nevertheless, the prediction from our minimal model successfully captures the behavior reported in literature that measured the sprout length for cell–cell distance of 0.7 mm (empty squares, Fig. 4c). Taking together, these results suggest that, within the system and parameter range explored here, for fibroblasts to have significant positive impact in the angiogenesis of endothelial cells, the fibroblast density should be no less than ~3×106 cells/ml and the HLF-HUVEC cell–cell distance should be no greater than ~2 mm, [Inline graphic cells/ml, Inline graphic 2 mm], as shown by the plateau region in Fig. 4c.

Conclusion and outlook

In summary, we have developed a microfluidic co-culture system and exploited it to systematically investigate the effects fibroblast densities and spatial configurations on angiogenic efficiency of endothelial cells. We vary the fibroblast density from 1 to 10 ×106 cells/ml and the HLF-HUVEC cell–cell distance from 1 to 10 mm; these ranges are much wider than those explored in literature. Using live cell imaging and image analysis, we quantify the dependence of endothelial morphology and sprouting on fibroblast density, cell–cell distance, and culture time. Total sprout area and sprout length are used as complementary morphological metrics to characterize the temporal evolution of endothelial growth, rather than to distinguish distinct biological processes such as vasculogenesis-like and angiogenesis-like behavior. Within the system and parameter range explored here, for fibroblasts to have a significant positive impact in the angiogenesis of endothelial cells, the fibroblast density should be no less than ~3×106 cells/ml and the cell–cell distance should be no greater than ~2 mm.

Like the classic microfluidic co-culture system (Kim et al. 2013), our system enables engineering perfusable vasculature in vitro. The difference is that our co-culture system allows for generating a monotonic gradient profile of pro-angiogenic factors secreted by fibroblasts. This simplifies the understanding of the effects of fibroblast density and spatial proximity in angiogenic response of endothelial cells. Accounting for the concentration profile of pro-angiogenic factors secreted by fibroblasts, we have developed a phenomenological model that successfully captures the dependence of sprout length on fibroblast density and cell–cell distance. Note that our model does not distinguish the roles of specific pro-angiogenic factors secreted by fibroblasts as well as the physical properties of fibrin gels, which may impact the angiogenic outcomes (Vilanova et al. 2018; Mehdizadeh et al. 2013; Chaplain 2000; Chaplain et al. 2006), and we did not systematically evaluate donor-to-donor variability in HUVECs or HLFs. Although different cell lots may alter absolute sprouting magnitude, the spatial scaling trends and threshold behaviors identified here arise from defined geometric constraints and are therefore expected to remain robust across sources. By establishing quantitative relationships between fibroblast distribution, density, and angiogenic outcomes, our results provide actionable guidelines for engineering functional vasculature. This becomes particularly useful with the advent of 3D bioprinting technologies, which offer exquisite control over the cell location in 3D space (Kolesky et al. 2016; Zhu et al. 2024, 2022).

Although the HUVEC/HLF co-culture model is widely used, it has inherent limitations (Nowak-Sliwinska et al. 2018). Endothelial cells exhibit significant tissue-specific heterogeneity, and stromal populations such as fibroblasts or mesenchymal stem cells (MSCs) differ in their paracrine and matrix remodeling profiles. The relationships identified here are therefore not intended to represent universal biological constants. Rather, they define a spatial–density framework for interpreting angiogenic responses. Because our microfluidic platform is cell-agnostic, the same experimental parameters can be directly applied to more physiologically relevant endothelial sources, such as human induced pluripotent stem cell (iPSC)–derived endothelial cells (ECs) (O’Connor et al. 2022; Williams and Wu 2019; Rufaihah et al. 2011), and alternative stromal populations. This approach provides a foundation for systematically evaluating how different EC–stromal combinations interpret spatial and density cues, advancing the development of more translatable in vitro angiogenesis models.

Materials and methods

Fabrication of microfluidic devices

We use soft lithography (McDonald et al. 2000; Mazutis et al. 2013) to fabricate microfluidic devices in poly(dimethyl siloxane) (PDMS). The fabrication process involves five steps: (i) mask design using computer-aided design (CAD) software, (ii) mask printing (Artnet Pro Inc), (iii) fabrication of the SU8-on-Si wafer masters (SU8-2150, Microchem, USA), (iv) curing of the PDMS (Sylgard 184, Dow Corning) using the master as a mold, and (v) bonding of the PDMS slabs to a glass substrate.

Cell culture

Primary human umbilical vein endothelial cells (HUVECs PCS-100-010, ATCC) are cultured in endothelial growth medium (Kit-VEGF PCS-100-041, ATCC). Normal primary human lung fibroblasts (HLF PCS-201-013, ATCC) are cultured in fibroblasts growth medium (Kit-Low serum PCS-201-041, ATCC). HUVECs and HLF are ready for the experiments after passages 3 to 5 and 5 to 7, respectively. All cells are maintained in a humidified incubator at 37 °C and 5% CO2.

Angiogenesis assay

We prepare a solution of 10 mg/ml fibrinogen (F8630, Sigma-Aldrich) and 0.45 U/ml aprotinin (A1153, Sigma-Aldrich) in Dulbecco’s phosphate-buffered saline (DPBS, Hyclone). In parallel, HUVECs are collected and suspended in cell culture medium at a concentration of 6.67×106 cells/ml, and fibroblasts are prepared at concentrations ranging from 1.33×106 to 13.3×106 cells/ml. We mix the fibrinogen solution with the cell suspension at 1:3 volume ratio, resulting in final concentrations of 2.5 mg/ml fibrinogen, 5×106 cells/ml HUVECs, and 1×106 to 10×106 cells/ml fibroblasts.

Before loading hydrogels, we pre-fill the channels with CO₂, which can dissolve into water to avoid the formation of air bubbles trapped between adjacent channels and within the micro post structures. To load fibroblasts into the microfluidic device, we add thrombin (T4648, 50 U/ml, Sigma-Aldrich) to the fibroblast–fibrinogen mixture at 50:1 volume ratio and immediately inject the mixture into Channel 5F. We wait for 3 min to allow for fibrin cross-linking and then load a fibrinogen solution (2.5 mg/ml in PBS) mixed with thrombin into the fibrin gel Channel 3G, followed by a 3-min clotting period to form the acellular fibrin gel. We then fill the Channel 2H with the HUVEC–fibrin gel mixture using the same method as that for Channel 5F. After waiting for 3 min, we fill the upper reservoirs of each device with endothelial cell growth kit-VEGF culture medium (PCS-100-041, ATCC). The lower reservoirs are gently aspirated to ensure proper filling of Channel 1 M and 4 M, and all reservoirs are then filled up and leveled with Kit-VEGF medium to establish static pressure between the two medium channels. We change the cell culture medium 24 h after cell seeding and refresh it every 48 h thereafter.

Live cell imaging and immunostaining

For live cell imaging, we use cell tracker to stain the cells. Specifically, CellTracker™ Green CMFDA dye (Invitrogen, MA) is diluted in pre-warmed, serum-free basal medium to a final concentration of 10 μM. After removing the culture medium from Channels 1 M and 4 M and reservoirs, the channels are rinsed with phosphate buffered saline (PBS, Hyclone) to eliminate residual medium. The CellTracker™ solution is then added to the channels, and the device is incubated for 40 min at 37 °C. Following incubation, the staining solution is removed, and the channels are washed three times with PBS to remove residual dyes. Fresh, pre-warmed culture medium containing serum is added, and the cells are incubated for an additional 30 min to allow internalization of fluorescent dye. Fluorescence imaging is performed using a fluorescence confocal microscope (Leica, SP8) with filter settings of excitation/emission: 492/517 nm. To allow for live-cell imaging over 5 days, the sample is re-stained on Day 3 using the same procedure.

For immunofluorescence staining, we wash the cultured samples once with PBS and fix the cells in 4% (w/v) paraformaldehyde in PBS for 15 min, then permeabilize the samples using a 0.15% Triton X-100 (Sigma) in PBS solution for 15 min. After blocking with 3% bovine serum albumin (BSA, Sigma) in PBS for 1 h, we incubate the samples overnight at 4 °C with antibody against CD31 (Alexa Fluor® 647, clone WM59). To stain F-actin and DNA, we add Alexa Fluor® 488-Phalloidin (66 nM) and Hoechst 33342 (1:1000) and incubate for 1 h at room temperature. The samples are washed three times and stored in PBS before imaging.

Perfusability of microvascular network

To assess the perfusability of microvascular networks, after 4 days of vasculogenic and angiogenic vessel formation, we stain the sample using Alexa Fluor® 488-Phalloidin to visualize endothelial structures. Then, we suspend 1 µm fluorescent polystyrene beads (F13080, Invitrogen) in PBS and introduce the suspension into the microfluidic device from one of the two medium reservoirs for HUVECs. Before doing so, the residual fluid from two medium reservoirs for HUVECs is aspirated. Note that the bead suspension is added to one reservoir only; this generates a concentration gradient that results in unidirectional flow of beads through the microvascular network. Microbead trajectories are analyzed using the TrackMate plugin in ImageJ. Individual bead paths are tracked and rendered as yellow lines overlaid on confocal images. Successful perfusion is defined by the presence of continuous bead trajectories confined within the luminal structures of the microvascular network, confirming the formation of potent, interconnected endothelial channels capable of supporting fluid transport.

Image and statistical analysis

Angiogenic sprout areas and lengths are quantified using ImageJ. Fluorescent images of HUVECs are acquired every 24 h after seeding, and each image is processed for background subtraction and thresholding to clearly define sprout boundaries. To ensure consistency, identical thresholding and measurement parameters are applied across all images in each experimental condition. Sprouts from four randomly selected regions of interest (ROIs) are analyzed for each condition. Measurements are averaged for each experimental group to ensure accurate quantification of angiogenesis.

For each experimental condition (e.g., specified gel separation distance or fibroblast density), we fabricate and prepare four microfluidic devices in parallel under identical configurations (n = 4). We seed and culture all replicates for a given condition on the same day to minimize variability associated with timing and handling. Across the full study, thirty-two devices successfully support angiogenic growth and are included in the analysis. An additional five to six devices fail during initial cell seeding, most commonly due to unintended leakage of the cell–gel suspension into adjacent channel regions, and we exclude these devices prior to data collection.

Statistical analysis is performed using one-way analysis of variance (ANOVA), followed by Tukey’s Honest Significant Difference (HSD) test to assess the significance of differences across all sprouting conditions for each group of study. Data are collected from multiple independent experiments using parallel microfluidic devices and are reported as mean ± standard deviation. All statistical calculations are conducted using SPSS.

Supplementary Information

Supplementary file 1 (PDF 489 KB) (489KB, pdf)
Download video file (970.5KB, avi)

Supplementary file 2. Configuration and dimension of microfluidic device, mechanical properties of fibrin gel, data on angiogenesis, and supplementary videos

Acknowledgements

We thank Dr. Jin Li at the Department of Orthopaedic Surgery for generously providing the laboratory space and equipment necessary for the fabrication of microfluidic devices used in this study.

Author contributions

Conceptualization: LHC, PZ, MKR Methodology: PZ Investigation: PZ, LHC Data Curation: PZ Visualization: LHC, PZ Funding acquisition: LHC, MKR Project administration: LHC, MKR Supervision: LHC, MKR Writing – original draft: LHC, PZ Writing – review & editing: LHC, PZ, MKR.

Funding

This work is supported by the UVA Office of the Vice President for Research and National Institutes of Health (1R35GM154912).

Data availability

All data are available in the manuscript or supplementary information.

Declarations

Conflict of interest

The authors declare no conflict of interest.

Footnotes

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Contributor Information

Melur K. Ramasubramanian, Email: mkr5a@virginia.edu

Li-Heng Cai, Email: liheng.cai@virginia.edu.

References

  1. Ahmann KA, Weinbaum JS, Johnson SL, Tranquillo RT (2010) Fibrin degradation enhances vascular smooth muscle cell proliferation and matrix deposition in fibrin-based tissue constructs fabricated in vitro. Tissue Eng Part A 16:3261 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Ahn J, Yoon MJ, Hong SH, Cha H, Lee D, Koo HS, Ko JE, Lee J, Oh S, Jeon NL, Kang YJ (2021) Three-dimensional microengineered vascularised endometrium-on-a-chip. Hum Reprod 36:2720–2731 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bang S, Tahk D, Choi YH, Lee S, Lim J, Lee SR, Kim BS, Kim HN, Hwang NS, Jeon NL (2022) Engineering densely packed adipose tissue via environmentally controlled in‐bath 3D bioprinting. Adv Funct Mater 32:1–13 [Google Scholar]
  4. Barkefors I, Le Jan S, Jakobsson L, Hejll E, Carlson G, Johansson H, Jarvius J, Jeong WP, Noo LJ, Kreuger J (2008) Endothelial cell migration in stable gradients of vascular endothelial growth factor A and fibroblast growth factor 2: effects on chemotaxis and chemokinesis. J Biol Chem 283:13905–13912 [DOI] [PubMed] [Google Scholar]
  5. Carmeliet P, Jain RK (2011) Molecular mechanisms and clinical applications of angiogenesis. Nature 473:298–307 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Chaplain MAJ, McDougall SR, Anderson ARA (2006) Mathematical modeling of tumor-induced angiogenesis. Annu Rev Biomed Eng 8:233–257 [DOI] [PubMed] [Google Scholar]
  7. Chen X, Aledia AS, Ghajar CM, Griffith CK, Putnam AJ, Hughes CCW, George SC (2009) Prevascularization of a fibrin-based tissue construct accelerates the formation of functional anastomosis with host vasculature. Tissue Eng - Part A 15:1363–1371 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Chen SW, Blazeski A, Zhang S, Shelton SE, Offeddu GS, Kamm RD (2023) Development of a perfusable, hierarchical microvasculature-on-a-chip model. Lab Chip 23:4552–4564 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Darland DC, D'Amore PA (2001) Cell-cell interactions in vascular development. Curr Top Dev Biol 52:107–149 [DOI] [PubMed]
  10. Dellaquila A, Dujardin C, Le Bao C, Chaumeton C, Carré A, Le Guilcher C, Lam F, Simon-Yarza T (2023) Fibroblasts mediate endothelium response to angiogenic cues in a newly developed 3D stroma engineered model. Biomater Adv 154:213636 [DOI] [PubMed]
  11. Du P, Suhaeri M, Ha SS, Oh SJ, Kim SH, Park K (2017) Human lung fibroblast-derived matrix facilitates vascular morphogenesis in 3D environment and enhances skin wound healing. Acta Biomater 54:333–344 [DOI] [PubMed] [Google Scholar]
  12. Dudley AC, Griffioen AW (2023) The modes of angiogenesis: an updated perspective. Angiogenesis 26:477–480 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Francis K, Palsson BO (1997) Effective intercellular communication distances are determined by the relative time constants for cyto/chemokine secretion and diffusion. Proc Natl Acad Sci U S A 94:12258–12262 [DOI] [PMC free article] [PubMed]
  14. Hartlapp I, Abe R, Saeed RW, Peng T, Voelter W, Bucala R, Metz CN (2001) Fibrocytes induce an angiogenic phenotype in cultured endothelial cells and promote angiogenesis in vivo. FASEB J 15:2215–2224 [DOI] [PubMed] [Google Scholar]
  15. Jahani M, Rezazadeh D, Mohammadi P, Abdolmaleki A, Norooznezhad A, Mansouri K (2020) Regenerative medicine and angiogenesis; challenges and opportunities. Adv Pharm Bull 10:490–501 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Johnson KE, Wilgus TA (2014) Vascular endothelial growth factor and angiogenesis in the regulation of cutaneous wound repair. Adv Wound Care 3:647 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Kalluri R (2016) The biology and function of fibroblasts in cancer. Nat Rev Cancer 16:582–598 [DOI] [PubMed] [Google Scholar]
  18. Kim S, Lee H, Chung M, Jeon NL (2013) Engineering of functional, perfusable 3D microvascular networks on a chip. Lab Chip 13:1489–1500 [DOI] [PubMed] [Google Scholar]
  19. Kim S, Chung M, Ahn J, Lee S, Jeon NL (2016) Interstitial flow regulates the angiogenic response and phenotype of endothelial cells in a 3D culture model. Lab Chip 16:4189–4199 [DOI] [PubMed] [Google Scholar]
  20. Kolesky DB, Homan KA, Skylar-Scott MA, Lewis JA (2016) Three-dimensional bioprinting of thick vascularized tissues. Proc Natl Acad Sci U S A 113:3179–3184 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Ladaigue S, Paget V, Lefranc AC, Quitoco M, Bacquer E, Milliat F, Guipaud O (2023) Protocol for in vitro assessment of human monocyte transendothelial migration using a high-throughput live cell imaging system. STAR Protoc 4:102388 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Li J, Mooney DJ (2016) Designing hydrogels for controlled drug delivery. Nat Rev Mater 1:16071 [DOI] [PMC free article] [PubMed]
  23. Liu ZL, Chen HH, Zheng LL, Sun LP, Shi L (2023) Angiogenic signaling pathways and anti-angiogenic therapy for cancer. Signal Transduct Target Ther 8:1–39 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. MacKintosh FC, Käs J, Janmey PA (1995) Elasticity of semiflexible biopolymer networks. Phys Rev Lett 75:4425–4428 [DOI] [PubMed] [Google Scholar]
  25. Chaplain MA (2000) Mathematical modelling of angiogenesis. J Neurooncol 50:37–51 [DOI] [PubMed]
  26. Margolis EA, Cleveland DS, Kong YP, Beamish JA, Wang WY, Baker BM, Putnam AJ (2021) Stromal cell identity modulates vascular morphogenesis in a microvasculature-on-a-chip platform. Lab Chip 21:1150–1163 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Mazutis L, Gilbert J, Ung WL, Weitz DA, Griffiths AD, Heyman JA (2013) Single-cell analysis and sorting using droplet-based microfluidics. Nat Protoc 8:870–891 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. McDonald JC, Duffy DC, Anderson JR, Chiu DT, Wu H, Schueller OJA, Whitesides GM (2000) Fabrication of microfluidic systems in poly(dimethylsiloxane). Electrophoresis 21:27–40 [DOI] [PubMed] [Google Scholar]
  29. Mehdizadeh H, Sumo S, Bayrak ES, Brey EM, Cinar A (2013) Three-dimensional modeling of angiogenesis in porous biomaterial scaffolds. Biomaterials 34:2875–2887 [DOI] [PubMed] [Google Scholar]
  30. Morin KT, Tranquillo RT (2013) In vitro models of angiogenesis and vasculogenesis in fibrin gel. Exp Cell Res 319:2409–2417 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Moura JA, Barlow HJ, Doak SH, Hawkins K, Muller I, Clift MJD (2024) Exploring the role of fibrin gels in enhancing cell migration for vasculature formation. J Funct Biomater 15:265 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Neurohr GE, Amon A (2020) Relevance and regulation of cell density. Trends Cell Biol 30:213–225 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Newman AC, Nakatsu MN, Chou W, Gershon PD, Hughes CCW (2011) The requirement for fibroblasts in angiogenesis: fibroblast-derived matrix proteins are essential for endothelial cell lumen formation. Mol Biol Cell 22:3791–3800 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Nowak-Sliwinska, P., Alitalo, K., Allen, E., Anisimov, A., Aplin, A. C., Auerbach, R., ... & Griffioen, A. W. (2018). Consensus guidelines for the use and interpretation of angiogenesis assays. Angiogenesis 21:425-532 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. O’Connor C, Brady E, Zheng Y, Moore E, Stevens KR (2022) Engineering the multiscale complexity of vascular networks. Nat Rev Mater 7:702–716 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Offeddu GS, Haase K, Gillrie MR, Li R, Morozova O, Hickman D, Knutson CG, Kamm RD (2019) An on-chip model of protein paracellular and transcellular permeability in the microcirculation. Biomaterials 212:115–125 [DOI] [PubMed] [Google Scholar]
  37. Otsu K, Das S, Houser SD, Quadri SK, Bhattacharya S, Bhattacharya J (2009) Concentration-dependent inhibition of angiogenesis by mesenchymal stem cells. Blood 113:4197–4205 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Peterson AW, Caldwell DJ, Rioja AY, Rao RR, Putnam AJ, Stegemann JP (2014) Vasculogenesis and angiogenesis in modular collagen–fibrin microtissues. Biomater Sci 2:1497–1508 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Rufaihah AJ, Huang NF, Jamé S, Lee JC, Nguyen HN, Byers B, De A, Okogbaa J, Rollins M, Reijo-Pera R, Gambhir SS, Cooke JP (2011) Endothelial cells derived from human iPSCS increase capillary density and improve perfusion in a mouse model of peripheral arterial disease. Arterioscler Thromb Vasc Biol 31:72–79 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Salcedo R, Wasserman K, Young HA, Grimm MC, Howard OMZ, Anver MR, Kleinman HK, Murphy WJ, Oppenheim JJ (1999) Vascular endothelial growth factor and basic fibroblast growth factor induce expression of CXCR4 on human endothelial cells: in vivo neovascularization induced by stromal-derived factor-1α. Am J Pathol 154:1125–1135 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Salvucci O, Yao L, Villalba S, Sajewicz A, Pittaluga S, Tosato G (2002) Regulation of endothelial cell branching morphogenesis by endogenous chemokine stromal-derived factor-1. Blood 99:2703–2711 [DOI] [PubMed] [Google Scholar]
  42. Soncin F (1992) Angiogenin supports endothelial and fibroblast cell adhesion. Proc Natl Acad Sci U S A 89:2232–2236 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Sottile J (2004) Regulation of angiogenesis by extracellular matrix. Biochim Biophys Acta 1654:13–22 [DOI] [PubMed] [Google Scholar]
  44. Stupack DG, Cheresh DA (2002) ECM remodeling regulates angiogenesis: endothelial integrins look for new ligands. Sci STKE 2002:pe7 [DOI] [PubMed]
  45. Tonnesen MG, Feng X, Clark RAF (2000) Angiogenesis in wound healing. J Investig Dermatol Symp Proc 5:40–46 [DOI] [PubMed]
  46. Vilanova G, Burés M, Colominas I, Gomez H (2018) Computational modelling suggests complex interactions between interstitial flow and tumour angiogenesis. J R Soc Interface 15:20180415 [DOI] [PMC free article] [PubMed]
  47. Wang WY, Lin D, Jarman EH, Polacheck WJ, Baker BM (2020) Functional angiogenesis requires microenvironmental cues balancing endothelial cell migration and proliferation. Lab Chip 20:1153–1166 [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Whisler JA, Chen MB, Kamm RD (2014) Control of perfusable microvascular network morphology using a multiculture microfluidic system. Tissue Eng - Part C Methods 20:543–552 [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Williams IM, Wu JC (2019) Generation of endothelial cells from human pluripotent stem cells: Methods, considerations, and applications. Arterioscler Thromb Vasc Biol 39:1317–1329 [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Xia J, Cai LH, Wu H, MacKintosh FC, Weitz DA (2021) Anomalous mechanics of Zn2+-modified fibrin networks. Proc Natl Acad Sci USA 118:e2020541118 [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Zheng Y, Chen J, Craven M, Choi NW, Totorica S, Diaz-Santana A, Kermani P, Hempstead B, Fischbach-Teschl C, López JA, Stroock AD (2012) In vitro microvessels for the study of angiogenesis and thrombosis. Proc Natl Acad Sci U S A 109:9342–9347 [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Zhou X, Franklin RA, Adler M, Carter TS, Condiff E, Adams TS, Pope SD, Philip NH, Meizlish ML, Kaminski N, Medzhitov R (2022) Microenvironmental sensing by fibroblasts controls macrophage population size. Proc Natl Acad Sci USA 119:e2205360119 [DOI] [PMC free article] [PubMed]
  53. Zhu J, He Y, Wang Y, Cai LH (2024) Voxelated bioprinting of modular double-network bio-ink droplets. Nat Commun 15:5902 [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Zhu J, He Y, Kong L, He Z, Kang KY, Grady SP, Nguyen LQ, Chen D, Wang Y, Oberholzer J, Cai LH (2022) Digital assembly of spherical viscoelastic bio-ink particles. Adv Funct Mater 32:2109004

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary file 1 (PDF 489 KB) (489KB, pdf)
Download video file (970.5KB, avi)

Supplementary file 2. Configuration and dimension of microfluidic device, mechanical properties of fibrin gel, data on angiogenesis, and supplementary videos

Data Availability Statement

All data are available in the manuscript or supplementary information.

Articles from Microfluidics and Nanofluidics are provided here courtesy of Springer

RESOURCES