Vasculogenic Potency of Bone Marrow- and Adipose Tissue-Derived Mesenchymal Stem/Stromal Cells Results in Differing Vascular Network Phenotypes in a Microfluidic Chip

Abstract

The vasculature is an essential, physiological element in virtually all human tissues. Formation of perfusable vasculature is therefore crucial for reliable tissue modeling. Three-dimensional vascular networks can be formed through the co-culture of endothelial cells (ECs) with stromal cells embedded in hydrogel. Mesenchymal stem/stromal cells (MSCs) derived from bone marrow (BMSCs) and adipose tissue (ASCs) are an attractive choice as stromal cells due to their natural perivascular localization and ability to support formation of mature and stable microvessels in vitro. So far, BMSCs and ASCs have been compared as vasculature-supporting cells in static cultures. In this study, BMSCs and ASCs were co-cultured with endothelial cells in a fibrin hydrogel in a perfusable microfluidic chip. We demonstrated that using MSCs of different origin resulted in vascular networks with distinct phenotypes. Both types of MSCs supported formation of mature and interconnected microvascular networks-on-a-chip. However, BMSCs induced formation of fully perfusable microvasculature with larger vessel area and length whereas ASCs resulted in partially perfusable microvascular networks. Immunostainings revealed that BMSCs outperformed ASCs in pericytic characteristics. Moreover, co-culture with BMSCs resulted in significantly higher expression levels of endothelial and pericyte-specific genes, as well as genes involved in vasculature maturation. Overall, our study provides valuable knowledge on the properties of MSCs as vasculature-supporting cells and highlights the importance of choosing the application-specific stromal cell source for vascularized organotypic models.

Keywords: endothelial cells; in vitro vascularization; mesenchymal stem cells; microfluidic chip; organ-on-a-chip; pericytes.

FIGURE 1

Characterizing interstitial flow within the used microfluidic chip. (A) Schematic representation of the experimental set-up for flow measurements. Gravity-driven flow [(90 µl + 90 µl)+(50 µl + 50 µl)] across the hydrogel area was generated by applying 50 µl of 1xDPBS to the medium reservoirs (pink) and 90 µl of 70 kDa Rhodamine B isothiocyanate–Dextran in 1xDPBS for the opposing medium reservoirs (red). Spatial change of the fluorescent wave front was imaged sequentially. Cells are depicted as dots. (B) Example of original and binarized image used for estimating maximal flow rate. (C) Example of the tracked waveforms of two consecutive images. Solid and dashed lines demonstrate the averaged waveform locations presenting the difference (red arrow) between averaged waveform locations between these two image indexes. The average change is then used to estimate current flow rate. Donor cell lines BMSC 1 and ASC 3 were used for the flow characterization co-cultures.

FIGURE 2

Morphology of the microvascular networks formed by EC-BMSC and EC-ASC co-cultures and ECs cultured alone. Microvascular networks were formed via vasculogenesis process by GFP-HUVECs cultured alone or in combination with supporting stromal cells—BMSCs or ASCs, for 6 days. (A) Confocal micrographs representing the overall architecture of microvascular networks established by EC-BMSC and EC-ASC co-cultures (5 million ECs/ml and 1 million MSCs/ml) and ECs alone (5 million cells/ml). Scale bars, 200 µm. Both EC-BMSC and EC-ASC co-cultures formed interconnected microvascular networks spanning the entire hydrogel over the course of 6 days. ECs cultured alone formed multicellular aggregates but failed to develop interconnected networks. (B) Hollow lumens could be observed in the cross-section images in all three conditions. Scale bars, 50 µm. Dash line depicts microposts that separate hydrogel from media channels. Donor cell lines BMSC 1 and ASC 2 were used for generation of data presented in the figure.

FIGURE 3

Immunocytochemistry analysis of basement membrane deposition and endothelial cell-cell junction formation in MSCs supported microvascular networks. (A) The EC networks (green) are surrounded by collagen IV (red), one of the major components of basement membrane, in both EC-BMSC and EC-ASC co-cultures (5 million ECs/ml and 1 million MSCs/ml). Empty basement membrane sleaves found in EC-ASC co-culture are marked by arrowheads. Scale bars, 100 µm. (B) Continuous and intact intercellular connections (red) are present along the border of ECs (green) in both conditions, as shown by immunostaining for adherens junction protein VE-cadherin. Nuclei (blue) are stained with DAPI. Scale bars, 50 µm. Donor cell lines BMSC 2 and ASC 1 were used for generation of data presented in (A). Donor cell lines BMSC 3 and ASC 2 were used for generation of data presented in (B).

FIGURE 4

MSCs of different origin induce microvascular network formation in ECs with significantly different vasculature area, vessel diameter, and vasculature length. The presence of MSCs greatly improves microvasculature parameters compared to ECs cultured alone. (A) Representative images of microvascular networks derived from EC-BMSC and EC-ASC co-cultures and EC alone at day 6 and corresponding binary images. Scale bars, 200 µm. (B) Morphological parameters (vasculature area, average vessel diameter, and vasculature length) of MSCs supported microvascular networks are shown individually for BMSCs and ASCs derived from different donors. Data are presented as means of 9 ROIs (3 devices, 3 ROIs per device). Comparison between BMSCs and ASCs as well as comparison of BMSCs and ASCs to EC alone was performed on mean values from EC-BMSC and EC-ASC co-cultures (3 donors, 3 devices per donor, 3 ROIs per device). ***—p < 0.001 with Unpaired student’s t-test. p values were corrected to control the false discovery rate. Donor cell lines BMSC 2 and ASC 1 were used for generation of data presented in (A).

FIGURE 5

Mesenchymal stem cell transition towards perivascular cells. (A,B) BMSCs as well as ASCs have pericytic characteristics marked by positive expression of PDGFR-β (A) and α-SMA (B). The immunostaining revealed that majority of MSCs are stained for PDGFR-β in both co-cultures. BMSCs are more organized and localize closely to the microvessels whereas ASCs have random appearance scattered throughout the hydrogel. In contrast to PDGFR-β staining, considerably less MSCs were stained for a-SMA in EC-ASC co-culture compared to EC-BMSC co-culture. Scale bars, 50 µm. (C) α-SMA positive MSCs (red) localize in close proximity and wrap around the vascular structures (green) in both EC-BMSC and EC-ASC co-cultures. White arrowheads indicate interaction of pericytes (red) with ECs (green). Scale bars, 50 µm. Nuclei (blue) are stained with DAPI. Donor cell lines BMSC 3 and ASC 1 were used for generation of data presented in (A). Donor cell lines BMSC 1 and ASC 2 were used for generation of data presented in (B, C).

FIGURE 6

BMSCs have greater pericytic capacity when co-cultured with ECs in a microfluidic chip compared to ASCs based on significantly higher pericyte area and pericytes coverage. (A) Whole chip scans showing alignment of perivascular MSCs (red, marked by α-SMA staining) with vasculature (green) in EC-BMSC and EC-ASC co-cultures at day 6. Images are maximum intensity projections of confocal stacks 200 µm height. Scale bars, 1 mm. (B) Quantifications of pericyte area and pericyte coverage (vasculature covered by pericytes) are shown individually for BMSCs and ASCs derived from different donors. Data are presented as means of 15 ROIs (3 devices, five ROIs per device); error bars, SD; *** denotes p < 0.001 with Unpaired student’s t-test. Comparison between BMSCs and ASCs was performed on mean values from EC-BMSC and EC-ASC co-cultures (3 donors, 3 devices per donor, 5 ROIs per device). Donor cell lines BMSC 1 and ASC 2 were used for generation of data presented in (A).

FIGURE 7

Measured average flow rates (Q) [(90 µl + 90 µl)+(50 µl + 50 µl)] on two independent biological replicates (N = 2) on (A) day 0, (B) day 2, and (C) day 4 for both EC-ASC (*) and EC-BMSC (□) co-cultures. (C) At day 4, there is minimal flow front movement between two consecutive image indexes due to decreased flow rate. This results in reduced accuracy of the used analysis method that is seen as oscillation of the measured flow rate values. Donor cell lines BMSC 1 and ASC 3 were used for the flow characterization co-cultures.

FIGURE 8

Simulated flow profiles for EC-ASC (blue) and EC-BMSC (red) co-cultures at days 0, 2 and 4 show a general decrease in flow rates and increase in total flow duration from day 0–4. Donor cell lines BMSC 1 and ASC 3 were used for the flow characterization co-cultures.

FIGURE 9

Quantitative RT-PCR analysis of gene expression in ASC- and BMSC-supported microvascular networks. Heatmap shows relative mRNA expression of pericyte-specific (PC) and endothelial-specific (EC) genes, ECM protein (ECM), angiogenic growth factors (GF), and other genes involved in vasculature morphogenesis and stability in EC-ASC and EC-BMSC co-cultures after 6 days of culture. Gene expression fold change of EC-BMSC and EC-ASC cocultures was computed relative to the average expression in EC-ASC co-cultures. The reference genes were 18S and GAPDH (for PECAM1, VEGFA, PDGFRB, CSPG2, ACTA2, ANGPT1, FGF2) and PECAM1 (for VWF, CDH5, KDR, FLT1, ANGPT2).