Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Engineered 3D vessel-on-chip using hiPSC-derived endothelial- and vascular smooth muscle cells

Engineered 3D vessel-on-chip using hiPSC-derived endothelial- and vascular smooth muscle cells Stem Cell Reports Report Engineered 3D vessel-on-chip using hiPSC-derived endothelial- and vascular smooth muscle cells 1,2,4 1,4 1 3 Marc Vila Cuenca, Amy Cochrane, Francijna E. van den Hil, Antoine A.F. de Vries, 2 1 1, Saskia A.J. Lesnik Oberstein, Christine L. Mummery, and Valeria V. Orlova * Department of Anatomy and Embryology, Leiden University Medical Center, Einthovenweg 20, 2333ZA Leiden, the Netherlands Department of Clinical Genetics, Leiden University Medical Center, 2333ZA Leiden, the Netherlands Department of Cardiology, Leiden University Medical Center, 2333ZA Leiden, the Netherlands These authors contributed equally *Correspondence: [email protected] https://doi.org/10.1016/j.stemcr.2021.08.003 SUMMARY Crosstalk between endothelial cells (ECs) and pericytes or vascular smooth muscle cells (VSMCs) is essential for the proper functioning of blood vessels. This balance is disrupted in several vascular diseases but there are few experimental models which recapitulate this vascular cell dialogue in humans. Here, we developed a robust multi-cell type 3D vessel-on-chip (VoC) model based entirely on human induced pluripotent stem cells (hiPSCs). Within a fibrin hydrogel microenvironment, the hiPSC-derived vascular cells self-organized to form sta- ble microvascular networks reproducibly, in which the vessels were lumenized and functional, responding as expected to vasoactive stim- 2+ ulation. Vascular organization and intracellular Ca release kinetics in VSMCs could be quantified using automated image analysis based on open-source software CellProfiler and ImageJ on widefield or confocal images, setting the stage for use of the platform to study vascular (patho)physiology and therapy. multi-cell type 3D tissues and vessels-on-chip (VoC) (Tro- INTRODUCTION nolone and Jain, 2021). Typically, cells incorporated in Crosstalk between endothelial cells (ECs) and mural cells these microphysiological devices are derived from non-hu- (pericytes and vascular smooth muscle cells [VSMCs]) is man sources, human (tumor) cell lines, or directly from pri- pivotal for proper function of many blood vessels. mary human tissue. Primary human cells provide the Aberrant EC-mural cell crosstalk often leads to vascular closest mimic to human blood vessels but are of limited diseases that range from hypertension, atherosclerosis, availability and of variable genetic origin (Tronolone and vascular calcification, and coronary artery disease to Jain, 2021). While hiPSC derivatives are now regarded as stroke and other conditions (Owens et al., 2004). Animal an alternative, they have so far largely been used in combi- nation with primary cells in microfluidic chips. For models, including genetically modified mice, are widely example, human primary ECs, or hiPSC-ECs have been used to study vascular development and disease, but combined with human primary mural cells (Campisi these do not always capture patient phenotypes unless et al., 2018; van Dijk et al., 2020; van Duinen et al., 2019) the mutations are homozygous deletions, and differ- but not with mural cells derived from hiPSC, precluding ences associated with the genetic background and sus- opportunities to replicate (patient-specific) vascular dis- ceptibility are not evident in mice (Berry et al., 2019; eases originating in the mural cells. Van Norman, 2020). Human induced pluripotent stem Here, we overcome these limitations by incorporating cells (hiPSCs) generated from heathy individuals and pa- hiPSC-ECs with hiPSC-VSMCs in entirely hiPSC-based tients are a useful source of vascular cells and they do VoCs. We have previously described robust protocols to reflect the genetic background of the individual from derive ECs and VSMCs from multiple healthy hiPSC lines whom they are derived (Samuel et al., 2015). Several with little batch-to-batch variability (Halaidych et al., methods have been described to generate ECs and 2018, 2019; Orlova et al., 2014a, 2014b). The function- VSMCs from hiPSCs and some have already been ality of hiPSC-VoC was compared with similar VoCs con- used to model vascular disease-specific abnormalities taining hiPSC-ECs and human primary mural cells of the (Cochrane et al., 2019). Nevertheless, and despite recent advances, many current same development origin, namely human brain vascular in vitro models of blood vessels fail to emulate the inte- smooth muscle cells (HBVSMCs) and primary human grated, complex and multicell-type composition of the brain vascular pericytes (HBVPs). In all cases, we showed human vasculature and do not include a mimic of blood the microenvironment in the microfluidic device sup- flow (Duval et al., 2017). To address this, microfluidic de- ported the formation of a 3D perfusable, self-assembled vices have been engineered that do incorporate these fea- microvascular network. We optimised the culture condi- tures and provide the environment for the formation of tions and developed an automated quantification Stem Cell Reports j Vol. 16 j 2159–2168 j September 14, 2021 j ª 2021 The Author(s). 2159 This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). A Agglutinin + hiPSC-VSMCs Agglutinin + HBVSMCs Agglutinin + HBVPs + hiPSC-VSMCs + HBVSMCs + HBVPs (i) (i) (i) SOX17 Agglutinin SOX17 Agglutinin SOX17 Agglutinin (iii) (iii) (iii) (ii) (ii) (ii) D E Dextran (70 kDa) Agglutinin (i) Dextran (70 kDa) Agglutinin (iii) + hiPSC-VSMCs + hiPSC-VSMCs + HBVSMCs + HBVPs (ii) F H J K G I L M + hiPSC-VSMCs + HBVSMCs + HBVPs (i) Collagen IV CD31 (i) Collagen IV CD31 (i) Collagen IV CD31 (iii) (iii) (iii) (ii) (i) (i) (legend on next page) 2160 Stem Cell Reports j Vol. 16 j 2159–2168 j September 14, 2021 framework of the vascular network, mural cell hiPSC-EC sprouting (Figure 1Aiv). Vacuoles started to morphology, EC-mural cell interaction and extracellular appear in hiPSC-ECs 12–24 h after seeding (Figure S2A, matrix (ECM) composition. Finally, we demonstrated a white arrowheads) followed by proliferation and remodel- functional application: the automated quantification of ing up to 72 h. An interconnected microvascular network vasoactive responses. formed as early as day 2 (Figure S2A) and this spanned the complete microfluidic channel by day 7 (Figures 1B, S1A, and S2A). All combinations of mural cells with hiPSC-ECs RESULTS resulted in vascular lumen formation (Figure 1C). Further- more, these lumenized networks were perfusable by fluo- Characterization of the VoC model rescent beads (10 mm) or FITC-Dextran (70 kDa) (Figures Using the commercially available AIM Biotech 3D cell cul- 1D and 1E; Video S1) under gravity-driven flow. Microvas- ture chips, hiPSC-ECs (Halaidych et al., 2018; Orlova cular networks formed in the presence of hiPSC-VSMCs or et al., 2014a, 2014b) were combined with hiPSC-VSMCs primary mural cells showed similar morphologies, with (Halaidych et al., 2019), HBVSMCs or HBVPs (Figure 1Ai) no significant difference in vessel density (%, Figure 1F), in a fibrin hydrogel (Figure 1Aii) and the cell/gel mix was average vessel length (mm, Figure 1G), mean vessel diam- injected into the middle channel of the microfluidic chip eter (mm, Figure 1H), branching point (BP) density (BP/ (Figure 1Aiii). As a control, hiPSC-ECs in a fibrin hydrogel mm , Figure 1I), extravascular spaces (%, Figure 1J), or without mural cells were used (Figure S1). Endothelial number of hiPSC-ECs (Figure 1K). In contrast, microvas- growth medium-2 (EGM-2), supplemented with vascular cular networks formed by hiPSC-ECs alone were less orga- endothelial growth factor (VEGF) (50 ng/mL), was used nized than microvascular networks with hiPSC-VSMCs to support microvascular network formation. Microfluidic (Figure S1A). Although lumen formation was observed chips were perfused through gravity-driven flow by add- in microvascular networks formed using only hiPSC- ing 100 mL of medium to the inlet and 50 mL to the outlet ECs, these appeared irregular and broken (Figure S1B). of each medium channel. The gravity-driven flow was re- The instability of microvascular networks without established every 24 h allowing medium exchange in the hiPSC-VSMCs was also evidenced by increased leakage microfluidic chip (Table S1). On day 1, after hiPSC-ECs of FITC-Dextran from the vessel network (70 kDa; Fig- had begun to self-organize, g-secretase inhibitor DAPT ure S1I). There were, however, no significant differences (10 mM) was added to the medium for 24 h to promote in vessel density (%, Figure S1C), branching point (BP) Figure 1. Characterization of VoC (A) Schematic of the VoC protocol. hiPSC-ECs were cultured with hiPSC-VSMCs, HBVSMCs, or HBVPs (i). Cells were mixed in a fibrin hydrogel (ii) and injected into (AIM Biotech) microfluidic chips (iii). EGM-2 was supplemented with VEGF (50 ng/mL), chips were refreshed daily for 7 days. EGM-2 was also supplemented with DAPT (10 mM) on day 1 for 24 h (iv). (B) Representative immunofluorescence images of microvascular network showing hiPSC-EC (magenta; agglutinin) vessels spanning the complete length of the microfluidic channel. Images showing hiPSC-ECs cultured with hiPSC-VSMCs, HBVSMCs, or HBVPs, respectively (103). Scale bars, 200 mm. (C) Representative confocal images of microvascular network showing hiPSC-ECs (gray; agglutinin) and hiPSC-EC nuclei (cyan; SOX17). Images displaying xyz (i), xy (ii), and yz cross-sectional perspectives (iii). Images showing hiPSC-ECs cultured with hiPSC-VSMCs, HBVSMCs, or HBVPs, respectively (403). Scale bars, 100 mm. (D) Representative Immunofluorescence images showing hiPSC-ECs (magenta; agglutinin) and perfusion of 70 kDa FITC-Dextran (green). Images showing hiPSC-ECs cultured with hiPSC-VSMCs, HBVSMCs, or HBVPs, respectively (103). Scale bars, 50 mm. (E) Representative confocal images showing hiPSC-ECs (magenta; agglutinin) and perfusion of 70 kDa FITC-Dextran (green) in hiPSC-VoC. Images displaying xyz (i), xy (ii), and yz cross-sectional perspectives (iii) (403). Scale bars, 100 mm. (F–K) Quantification of vessel density (%) (F), average vessel length (mm) (G), mean diameter (mm) (H), branching point (BP) density (BPs/mm ) (I) extravascular spaces (%) (J), and number of hiPSC-ECs (K), from hiPSC-ECs cultured with hiPSC-VSMCs, HBVSMCs, or HBVPs, respectively, are shown. Data are shown as ±SD from N = 3, n = 6; three independent experiments with two microfluidic channels per experiment. (L) Representative confocal images of microvascular network showing hiPSC-ECs (magenta; CD31) and ECM (yellow; collagen IV). Images displaying xyz (i), xy (ii), and yz cross-sectional perspectives (iii). Images showing hiPSC-ECs cultured with hiPSC-VSMCs, HBVSMCs, or HBVPs, respectively (403). Scale bars, 100 mm. (M) Quantification of collagen IV density (%) from hiPSC-ECs cultured with hiPSC-VSMCs, HBVSMCs, or HBVPs, respectively, are shown. Data are shown as ±SD from N = 3, n = 8; three independent biological replicates with two to three microfluidic channels per experiment. One- way ANOVA with Tukey’s multiple comparison. See also Figures S1 and S2A–S2D and Video S1. Stem Cell Reports j Vol. 16 j 2159–2168 j September 14, 2021 2161 A SM22 Agglutinin + hiPSC-VSMCs SM22 Agglutinin + HBVSMCs SM22 Agglutinin + HBVPs + HBVSMCs + HBVPs + hiPSC-VSMCs (i) (i) SM22 Agglutinin SM22 Agglutinin (i) SM22 Agglutinin (iii) (iii) (iii) (ii) (ii) (ii) D E G H + hiPSC-VSMCs + HBVSMCs + HBVPs SM22 mCherry SM22 mCherry SM22 mCherry SM22 Intensity 0 40 (legend on next page) 2162 Stem Cell Reports j Vol. 16 j 2159–2168 j September 14, 2021 2 density (BP/mm , Figure S1F), extravascular space (%, Fig- significant difference in the cell length (mm, Figure 2D) or ure S1G), or number of hiPSC-ECs (Figure S1H) in micro- cell circularity (Figure 2E). Analysis of the contractile vascular networks formed either with or without hiPSC- marker SM22 in mural cells in microvascular networks VSMCs. By contrast, average vessel length (mm, Fig- showed a significantly lower normalized mean cell inten- ure S1D) and mean vessel diameter (mm, Figure S1E) sity in hiPSC-VSMCs compared with HBVSMCs (Figures were significantly reduced or increased respectively in 2F and 2G) while the total number of SM22 + cells was significantly lower in microvascular networks formed microvascular networks formed from hiPSC-ECs alone compared with microvascular networks formed in the with HBVSMCs (Figure 2H). Notably, we also observed presence of hiPSC-VSMCs. No significant differences that all mural cells displayed significantly higher SM22 were found in ECM deposition, evidenced by changes in staining intensity when in contact with hiPSC-ECs (Figures the relative density of collagen IV, between any cell com- 2F and 2I), which indicated that heterotypic cell-cell con- binations (Figures 1L, 1M, S1J, and S1K). Furthermore, the tact in VoC culture could further promote mural cell matu- presence of fibronectin was confirmed by immunostain- ration. Furthermore, long-term hiPSC-VoC culture resulted ing in microvascular networks formed with mural cells in an increase over time in the percentage of mural cells (Figure S2D). Finally, long-term culture of hiPSC-ECs in located close to the hiPSC-EC vessel wall (Figures S2E VoC with mural cells demonstrated that the microvas- and S2F). cular network architecture was stable and underwent 2+ continuous increase in vessel density over 21 days (Figures Assessment of hiPSC-VSMC Ca dynamics in the VoC S2A–S2C), although we observed an increase in density of model vessel networks formed with primary HBVSMCs on day To further assess the functionality of hiPSC-VSMCs in the 2+ 21 (Figures S2A–S2C). hiPSC-VoC, we measured intracellular Ca release in hiPSC-VSMCs engineered to express an ultra-sensitive 2+ Characterization of hiPSC-VSMCs and primary mural Ca sensor (GCaMP6f) (Chen et al., 2013). First, hiPSC- cells in the VoC model derived neural crest (NC) intermediates were transduced hiPSC-VSMCs, HBVSMCs, and HBVPs self-organized and with a lentiviral vector (LV) expressing GCaMP6f self-oriented toward the developing hiPSC-EC microvas- (Figure S3A). Transduction of hiPSC-NC cells with LVs en- cular networks, as early as day 1 (Figure S2E). On day 7, coding either enhanced green fluorescent protein or all mural cells were located at extravascular positions along GCaMP6f did not change expression of the surface the entire length of the microvascular network (Figure 2A), marker CD271 (Figure S3B). hiPSC-VSMCs engineered to 2+ surrounding hiPSC-EC lumens (Figure 2B). No significant express GCaMP6f showed no intracellular Ca release difference was observed in the percentage of hiPSC-VSMCs, upon perfusion with medium only (pre-stimulated, Figures 2+ HBVSMCs, or HBVPs associated with the hiPSC-EC lumen S3C and S3D) and similar intracellular Ca release upon (% mural cells localized at the vascular network, Figure 2C). stimulation with the vasoconstrictor endothelin-I (ET-I) Quantification of mural cell morphology also showed no (post-stimulated, Figures S3C and S3D) to Fluo-4-labeled Figure 2. Quantitative assessment of the structural proprieties of hiPSC-VSMCs and primary mural cells in VoC (A) Representative immunofluorescence images of microvascular network showing the hiPSC-ECs (magenta; agglutinin) and mural cells (green; SM22) derived vasculature unit spanning the complete length of microfluidic channel. Images showing hiPSC-ECs cultured with hiPSC-VSMCs, HBVSMCs, or HBVPs, respectively (103). Scale bars, 200 mm. (B) Representative confocal images of microvascular network showing hiPSC-ECs (gray; agglutinin) and mural cells (green; SM22). Images displaying xyz (i), xy (ii), and yz cross-sectional perspectives (iii). Images showing hiPSC-ECs cultured with hiPSC-VSMCs, HBVSMCs, or HBVPs, respectively (403). Scale bars, 100 mm. (C–E) Quantification of the percentage of mural cells associated with the hiPSC-EC lumen (% mural cells localized at the vascular network) (C), mean mural cell length (mm) (D), and mural cell circularity factor (the circle is 1) (E) in hiPSC-VSMCs, HBVPs, and HBVSMCs. Data are shown as ±SD from N = 3, n = 6; three independent experiments with two microfluidic channels per experiment. (F) (Top) Representative confocal images of microvascular network showing hiPSC-ECs (gray; mCherry) and mural cells (red; SM22). (Bottom) Representative surface-rendered objects of confocal images showing microvascular network (gray; mCherry) and mural cells (colour-coded scale representing SM22 intensity). Images showing hiPSC-ECs cultured with hiPSC-VSMCs, HBVSMCs, or HBVPs, respectively (403). Scale bars, 100 mm. (G–I) Quantification of normalized mean cell SM22 intensity (G), number of SM22 + cells (H), and normalized mean cell SM22 intensity of mural cells in contact with hiPSC-ECs (I) in hiPSC-VSMCs, HBVPs and HBVSMCs. Intensity was normalized to hiPSC-VSMCs. Data are shown as ±SD from N = 3, n = 6; three independent experiments with two microfluidic channels per experiment. One-way ANOVA (C–E and G–H) and two-way ANOVA (I) with Tukey’s multiple comparison. *p < 0.05, **p < 0.001, ***p < 0.0001. See also Figures S2E and S2F. Stem Cell Reports j Vol. 16 j 2159–2168 j September 14, 2021 2163 hiPSC-VSMCs (Halaidych et al., 2019). Next, intracellular assumed positions surrounding the vascular wall, support- 2+ Ca release in hiPSC-VSMCs in the microvascular network ing the vessel and maintaining its functionality in 3D. was examined prior to- (basal state) and after medium Although we did not find any morphological differences refreshment on day 7. Significantly higher GCaMP6f fluo- in microvascular network organization and mural cell rescence was observed across the entire microvascular morphology between hiPSC-VSMCs and primary mural network after medium refreshment (Figures 3A and 3B; cells, expression of the contractile marker SM22 did differ Video S2). We then combined medium refreshment with between hiPSC-VSMCs and HBVSMCs. hiPSC-VSMCs and ET-1 stimulation (1 mM, Figure 3C; Video S2). To compare HBVPs showed lower SM22 staining intensities and higher 2+ the divergence in Ca responses, the average fluorescence total cell numbers in VoC culture. This could indicate that intensity of regions of interest over time (F/F ) was exam- hiPSC-VSMCs are less differentiated in VoC culture. ined by adapting previous methods (Halaidych et al., Notably, by incorporating and monitoring fluorescently 2+ 2019)(Figure 3D). Comparison of Ca kinetic parameters, tagged hiPSC-derived vascular cells (Roberts et al., 2017), measured at the half-maximum level (F/F0) , showed we showed that the most important steps of vascular max significantly higher amplitudes (F/F , Figure 3E), intensity network formation and remodeling occurred in the first over time (AUC [F*s/F0], Figure 3F) and duration, and 7 days of VoC culture. We also showed that these microvas- 2+ slower decay of the Ca transient without changes in the cular networks are stable for up to 3 weeks. Vessel density time to peak (s, Figures 3G–3I) upon ET-I stimulation. increased over time, although not much remodeling occur- ring beyond day 7 of culture. Specifically, microvascular networks formed with primary HBVSMCs showed the Modeling hiPSC-derived EC-VSMC crosstalk in the highest increase in vessel density, indicating that an in- VoC model crease in hiPSC-VSMC number observed on day 7 might To demonstrate the potential utility of hiPSC-VoC for dis- be advantageous for supporting long-term VoC culture. ease modeling, we examined the loss of EC-VSMC crosstalk Moreover, we demonstrated that hiPSC-VSMCs in the upon blocking NOTCH signaling using the small-molecule VoC were responsive to vasoactive stimulation by quanti- g-secretase inhibitor DAPT (10 mM). The addition of DAPT 2+ fying changes in Ca kinetic parameters. We noted that (10 mM, day 5–7) affected overall microvascular network ar- hiPSC-VSMCs were activated simply after medium refresh- chitecture (Figure S4A) with significant changes in the ment, an important consideration for proper experimental mean vessel diameter although vessel density was similar control. It seemed likely this was mediated by factors in the between the groups (Figures S4B and S4C). DAPT supple- fresh media since shear-stress during gravity-mediated flow mentation had no significant effect on the percentage of in the VoC was too slow to be stimulatory. In addition, we hiPSC-VSMCs associated with the hiPSC-EC lumen (% confirmed more pronounced and coordinated hiPSC- mural cells localized at the vascular network, Figures 4A 2+ VSMC intracellular Ca release following addition of the and 4B). However, hiPSC-VSMCs showed a significant vasoconstrictor ET-I. decrease in the cell length (mm, Figure 4C) with a signifi- Conventional preclinical models have shown low suc- cant increase in cell circularity (Figure 4D). Analysis of cess in accurately predicting drug efficacy and toxicity in the contractile marker SM22 in hiPSC-VSMCs showed a human trials (Van Norman, 2020). We considered it essen- significantly lower normalized mean SM22 intensity (Fig- tial therefore to provide some evidence that hiPSC-VoC re- ures 4E and 4F) while no change in the total number of sponded to drugs as expected and that this property was SM22 + cells was observed, following addition of DAPT to conserved across different healthy hiPSC lines and batches. the cultures (Figure 4G). This indeed was the case: under conditions of vessel matu- ration where the NOTCH signaling pathway coordinates heterologous cell-cell crosstalk and maintains vessel integ- DISCUSSION rity, we showed that these features were disrupted in the This report describes the generation of hiPSC-derived hiPSC-VoC after inhibition by DAPT on days 5–7. In partic- microvascular networks composed of hiPSC-ECs and ular, DAPT addition appeared to reduce acquisition of a hiPSC-VSMCs. We showed that hiPSC-ECs form an inter- contractile-like identity by hiPSC-VSMCs. This notion connected microvascular network with perfusable lumens was supported by reduced expression the contractile and ECM deposits as in previous microfluidic studies based marker SM22 and a less elongated morphology after EC- on other cell sources (Belair et al., 2015; Campisi et al., VSMC crosstalk inhibition by DAPT. We also observed 2018). We demonstrate that although hiPSC-ECs alone gradual reversal of vascular stability and signs of vessel can form a microvascular network, inclusion of mural cells regression, reflecting what is known about dysfunctional EC-VSMC crosstalk in developed vessels in vivo (Kerr facilitates hiPSC-EC self-organization and supports vessel et al., 2016). stability. Much like primary mural cells, hiPSC-VSMCs 2164 Stem Cell Reports j Vol. 16 j 2159–2168 j September 14, 2021 A Basal state mCherry GCaMP6f Media refreshment mCherry GCaMP6f C D Pre-stimulated Post-stimulated (ET-I 1μM) mCherry GCaMP6f mCherry GCaMP6f E F G H I 2+ Figure 3. Analysis of hiPSC-VSMCs Ca dynamics in VoC 2+ (A) Representative immunofluorescent images of intracellular Ca fluorescence showing hiPSC-ECs (gray; mCherry) and hiPSC-VSMCs (green; GCaMP6f) without- (basal state) and after EGM-2 refreshment on day 7 (103). Scale bar, 100 mm. (B) Normalized GCaMP6f intensity at day 7. GCaMP6f intensity was normalized to the condition prior to EGM-2 refreshment (basal state). Data are shown as ±SD of N = 3, n = 21; three independent experiments with seven microfluidic channels per experiment. 2+ (C) Representative confocal images of intracellular Ca fluorescence showing with hiPSC-ECs (gray; mCherry) and hiPSC-VSMCs (green; GCaMP6f) in pre- and post-stimulated (ET-I, 1 mM) states (203). Scale bar, 100 mm. (D) Normalized average fluorescence intensity F/F in hiPSC-VSMCs expressing GCaMP6f. Medium channels were gravity-flow perfused with EGM-2 alone or containing ET-I (1 mM). Stimulation time point is set as t = 5(s). Data are shown as ±SD of N = 3, n = 9; three independent experiments with three microfluidic channels per experiment. 2+ (E–I) Ca transient parameters: amplitude F/F (E), duration (s) (F), area under the curve (AUC, F*s/F ) (G), time to peak (s) (H), and 0 0 decay (s) (I) of channels gravity-flow perfused with EGM-2 alone or containing ET-I (1 mM). Data are shown as ±SD of N = 3, n = 9; three independent experiments with three microfluidic channels per experiment. Paired (B) Student’s t test. Wilcoxon-Mann-Whitney test (E–I). *p < 0.05, **p < 0.001, ***p < 0.0001; ns, not significant. See also Figure S3 and Video S2. Stem Cell Reports j Vol. 16 j 2159–2168 j September 14, 2021 2165 A B C FG Figure 4. Modeling loss of EC-VSMC crosstalk in VoC (A) Representative confocal images of microvascular network showing in hiPSC-ECs (gray; GFP) and hiPSC-VSMCs (green; RFP) in control and DAPT (10 mM) supplemented conditions. Images displaying xyz (i), xy, (ii), yz cross-sectional perspectives (iii), and enlargements of white framed areas (iv) (403). Scale bars, 100 mm in (i–iii) and 50 mm in (iv). (B–D) Quantification of the percentage of hiPSC-VSMCs associated with the hiPSC-EC lumen (% mural cells localized at the vascular network) (B), mean hiPSC-VSMCs length (mm) (C), and hiPSC-VSMC circularity factor (the circle is 1) (D) in control and DAPT (10 mM) supplemented conditions at day 7. Data are shown as ±SD from N = 3, n = 27; three independent experiments with nine microfluidic channels per experiment. (E) (Left) Representative confocal images of microvascular network showing hiPSC-ECs (gray; mCherry) and hiPSC-VSMCs cells (red; SM22). (Right) Representative surface-rendered objects of confocal images showing microvascular network (gray; mCherry) and hiPSC-VSMCs (colour-coded scale representing SM22 intensity) in control and DAPT (10 mM) supplemented conditions (403). Scale bars, 100 mm. (legend continued on next page) 2166 Stem Cell Reports j Vol. 16 j 2159–2168 j September 14, 2021 middle gel-loading channel of the microfluidic chip. Chips were In summary, we have demonstrated several advantages incubated at room temperature for 15 min before the addition of of this hiPSC-VoC model above existing models: (1) the EGM-2 supplemented with VEGF (50 ng/mL) to both flanking me- possibility of studying genetic vascular diseases with dia channels. The g-secretase inhibitor DAPT (10 mM) was also patient-specific lines; (2) evaluation of real-time changes added to the medium on day 1 for 24 h. Gravity-driven flow was in vascular structure and function by incorporating fluores- induced by the addition of 100 mL medium to the right media ports cently tagged cells; (3) the ability to measure and quantify and 50 mL media to left media ports. Medium was refreshed daily. effects of drugs affecting EC-mural cell crosstalk with view to modulating vessel stability and integrity. Nevertheless, Statistical analysis the model could be further improved by: (1) introducing Statistical analyses were performed using GraphPad Prism 9 soft- physiological (rather than gravity-driven) flow which ware. Normality of the data was evaluated by the D’Agostino-Pear- would recapitulate vessel shear forces and geometry in son test. One-way and two-way ANOVA with Tukey’s multiple healthy and disease environments and (2) the addition of comparison test was used for the analysis of three groups. For other hiPSC-derived cells such as monocytes or astrocytes paired or unpaired analysis of two groups, either Student’s t test which would allow mimicking of (isogenic) inflammatory or Wilcoxon-Mann-Whitney test was used. Analyses are indicated reactions or features of the brain. in the figure legends. The data are reported as mean ± SD. Statistical In conclusion, the hiPSC-VoC model described in this significance was defined as p < 0.05. paper will be useful to study and quantify changes in the vascular architecture and function during vascular devel- SUPPLEMENTAL INFORMATION opment or upon drug treatment. Clinically, this may trans- Supplemental information can be found online at https://doi.org/ late into a better understanding of vascular disease condi- 10.1016/j.stemcr.2021.08.003. tions and predicting drug efficiency. AUTHOR CONTRIBUTIONS EXPERIMENTAL PROCEDURES Conceptualization, V.V.O.; methodology, M.V.C., A.C., and V.V.O.; software, A.C.; validation, A.C. and M.V.C.; formal analysis, A.C. Full details are provided in supplemental experimental procedures. and M.V.C.; investigation, A.C., M.V.C., and F.E.v.d.H.; visualisa- tion, M.V.C. and A.C.; resources, A.A.F.d.V. and V.V.O.; writing – hiPSC lines original draft, M.V.C., A.C., C.L.M., and V.V.O.; writing – review Research on hiPSC was approved by the medical ethical committee & editing, M.V.C., A.C., S.A.J.L.O., C.L.M., and V.V.O.; supervision, at Leiden University Medical Center, the Netherlands. A detailed S.A.J.L.O, C.L.M., and V.V.O.; project administration, V.V.O.; fund- list of the hiPSC lines and batches used for each experiment is pro- ing acquisition, A.C., S.A.J.L.O., C.L.M., and V.V.O. vided in Table S2. CONFLICT OF INTERESTS Differentiation of hiPSC-ECs and hiPSC-VSMCs The authors declare no competing interests. hiPSC differentiation to ECs was performed as described previously (Orlova et al., 2014a, 2014b). hiPSC differentiation to VSMC was ACKNOWLEDGMENTS performed as previously described (Halaidych et al., 2019). The LUMC human iPSC Hotel for generation and characterization Setting up VoCs of hiPSC lines; LUMC’s confocal imaging facility; Cindy Bart, Sven hiPSC-ECs and mural cells were prepared prior to incorporation in Dekker and Juan Zhang for LV production; Oleh Halaidych for use- VoCs as described in supplemental experimental procedures. ful discussion on hiPSC-VSMCs. The Allen Cell Collection, avail- Commercially available microfluidic chips with one gel channel able from Coriell Institute for Medical Research, provided mate- and two media channels (AIM Biotech) were used. Cells were resus- rials. Images were generated using Biorender.com. pended and combined to obtain 10 3 10 hiPSC-ECs/mL and 2 3 This work was supported by the Netherlands Organisation for 10 mural cells/mL (5:1 ratio). Three different mural cell suspen- Health Research and Development (ZonMw): PTO 446002501 sions were tested in combination with hiPSC-ECs: (1) hiPSC- and VIDI 91717325; the Netherlands Organ-on-Chip Initiative VSMCs, (2) HBVSMCs, and (3) HBVPs. Cell were resuspended in which is an NWO Gravitation project (024.003.001) funded by EGM-2 supplemented with Thrombin (4 U/mL) and then gently the Ministry of Education, Culture and Science of the government mixed with fibrinogen (final concentration 3 mg/mL, Sigma) at of the Netherlands; European Research Council (ERCAdG 1:1 vol ratio. Cell/hydrogel mixture was quickly loaded into the 323182 STEMCARDIOVASC); the European Union’s Horizon (F and G) Quantification of normalized mean cell SM22 intensity (F) and number of SM22 + cells (G). Intensity was normalized to control condition. Data are shown as ±SD from N = 3, n = 9; three independent experiments with three microfluidic channels per experiment. Wilcoxon-Mann-Whitney test. *p < 0.05, **p < 0.001, ***p < 0.0001; ns, not significant. See also Figure S4. Stem Cell Reports j Vol. 16 j 2159–2168 j September 14, 2021 2167 2020 research and innovation program under the Marie Sklodow- Kerr, B.A., West, X.Z., Kim, Y.W., Zhao, Y., Tischenko, M., Cull, ska Curie grant agreement no. 707404. R.M., Phares, T.W., Peng, X.D., Bernier-Latmani, J., Petrova, T.V., et al. (2016). Stability and function of adult vasculature is sustained Received: March 1, 2021 by Akt/Jagged1 signalling axis in endothelium. Nat. Commun. 7, Revised: August 5, 2021 Accepted: August 6, 2021 Orlova, V.V., Drabsch, Y., Freund, C., Petrus-Reurer, S., van den Hil, Published: September 2, 2021 F.E., Muenthaisong, S., Dijke, P.T., and Mummery, C.L. (2014a). Functionality of endothelial cells and pericytes from human REFERENCES pluripotent stem cells demonstrated in cultured vascular plexus and zebrafish xenografts. Arterioscler. Thromb. Vasc. Biol. 34, Belair, D.G., Whisler, J.A., Valdez, J., Velazquez, J., Molenda, J.A., 177–186. Vickerman, V., Lewis, R., Daigh, C., Hansen, T.D., Mann, D.A., et al. (2015). Human vascular tissue models formed from human Orlova, V.V., van den Hil, F.E., Petrus-Reurer, S., Drabsch, Y., Ten induced pluripotent stem cell derived endothelial cells. Stem Cell Dijke, P., and Mummery, C.L. (2014b). Generation, expansion Rev. Rep. 11, 511–525. and functional analysis of endothelial cells and pericytes derived from human pluripotent stem cells. Nat. Protoc. 9, 1514–1531. Berry, C., Sidik, N., Pereira, A.C., Ford, T.J., Touyz, R.M., Kaski, J.C., and Hainsworth, A.H. (2019). Small-vessel disease in the heart and Owens, G.K., Kumar, M.S., and Wamhoff, B.R. (2004). Molecular brain: current knowledge, unmet therapeutic need, and future di- regulation of vascular smooth muscle cell differentiation in devel- rections. J. Am. Heart Assoc. 8, e011104. opment and disease. Physiol. Rev. 84, 767–801. Campisi, M., Shin, Y., Osaki, T., Hajal, C., Chiono, V., and Kamm, Roberts, B., Haupt, A., Tucker, A., Grancharova, T., Arakaki, J., Fu- R.D. (2018). 3D self-organized microvascular model of the human qua, M.A., Nelson, A., Hookway, C., Ludmann, S.A., Mueller, I.A., blood-brain barrier with endothelial cells, pericytes and astrocytes. et al. (2017). Systematic gene tagging using CRISPR/Cas9 in hu- Biomaterials 180, 117–129. man stem cells to illuminate cell organization. Mol. Biol. Cell 28, Chen, T.W., Wardill, T.J., Sun, Y., Pulver, S.R., Renninger, S.L., Bao- 2854–2874. han, A., Schreiter, E.R., Kerr, R.A., Orger, M.B., Jayaraman, V., et al. Samuel, R., Duda, D.G., Fukumura, D., and Jain, R.K. (2015). (2013). Ultrasensitive fluorescent proteins for imaging neuronal activity. Nature 499, 295–300. Vascular diseases await translation of blood vessels engineered from stem cells. Sci. Transl. Med. 7, 309rv306. Cochrane, A., Albers, H.J., Passier, R., Mummery, C.L., van den Berg, A., Orlova, V.V., and van der Meer, A.D. (2019). Advanced Tronolone, J.J., and Jain, A. (2021). Engineering new microvascular in vitro models of vascular biology: human induced pluripotent networks on-chip: ingredients, assembly, and best practices. Adv. stem cells and organ-on-chip technology. Adv. Drug Deliv. Rev. Funct. Mater. 31, 2007199. 140, 68–77. van Dijk, C.G.M., Brandt, M.M., Poulis, N., Anten, J., van der Moo- Duval, K., Grover, H., Han, L.H., Mou, Y., Pegoraro, A.F., Fredberg, len, M., Kramer, L., Homburg, E., Louzao-Martinez, L., Pei, J., Kreb- J., and Chen, Z. (2017). Modeling physiological events in 2D vs. 3D ber, M.M., et al. (2020). A new microfluidic model that allows cell culture. Physiology (Bethesda) 32, 266–277. monitoring of complex vascular structures and cell interactions Halaidych, O.V., Cochrane, A., van den Hil, F.E., Mummery, C.L., in a 3D biological matrix. Lab Chip 20, 1827–1844. and Orlova, V.V. (2019). Quantitative analysis of intracellular van Duinen, V., Zhu, D., Ramakers, C., van Zonneveld, A.J., Vulto, Ca(2+) release and contraction in hiPSC-derived vascular smooth P., and Hankemeier, T. (2019). Perfused 3D angiogenic sprouting in muscle cells. Stem Cell Reports 12, 647–656. a high-throughput in vitro platform. Angiogenesis 22, 157–165. Halaidych, O.V., Freund, C., van den Hil, F., Salvatori, D.C.F., Rimi- nucci, M., Mummery, C.L., and Orlova, V.V. (2018). Inflammatory Van Norman, G.A. (2020). Limitations of Animal studies for pre- responses and barrier function of endothelial cells derived from dicting toxicity in clinical trials: Part 2: potential alternatives to human induced pluripotent stem cells. Stem Cell Reports 10, the use of animals in preclinical trials. JACC Basic Transl. Sci. 5, 1642–1656. 387–397. 2168 Stem Cell Reports j Vol. 16 j 2159–2168 j September 14, 2021 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Stem Cell Reports Unpaywall

Engineered 3D vessel-on-chip using hiPSC-derived endothelial- and vascular smooth muscle cells

Download PDF
10 pages

Loading next page...
 
/lp/unpaywall/engineered-3d-vessel-on-chip-using-hipsc-derived-endothelial-and-2S5g652r7W

References (28)

Publisher
Unpaywall
ISSN
2213-6711
DOI
10.1016/j.stemcr.2021.08.003
Publisher site
See Article on Publisher Site

Abstract

Stem Cell Reports Report Engineered 3D vessel-on-chip using hiPSC-derived endothelial- and vascular smooth muscle cells 1,2,4 1,4 1 3 Marc Vila Cuenca, Amy Cochrane, Francijna E. van den Hil, Antoine A.F. de Vries, 2 1 1, Saskia A.J. Lesnik Oberstein, Christine L. Mummery, and Valeria V. Orlova * Department of Anatomy and Embryology, Leiden University Medical Center, Einthovenweg 20, 2333ZA Leiden, the Netherlands Department of Clinical Genetics, Leiden University Medical Center, 2333ZA Leiden, the Netherlands Department of Cardiology, Leiden University Medical Center, 2333ZA Leiden, the Netherlands These authors contributed equally *Correspondence: [email protected] https://doi.org/10.1016/j.stemcr.2021.08.003 SUMMARY Crosstalk between endothelial cells (ECs) and pericytes or vascular smooth muscle cells (VSMCs) is essential for the proper functioning of blood vessels. This balance is disrupted in several vascular diseases but there are few experimental models which recapitulate this vascular cell dialogue in humans. Here, we developed a robust multi-cell type 3D vessel-on-chip (VoC) model based entirely on human induced pluripotent stem cells (hiPSCs). Within a fibrin hydrogel microenvironment, the hiPSC-derived vascular cells self-organized to form sta- ble microvascular networks reproducibly, in which the vessels were lumenized and functional, responding as expected to vasoactive stim- 2+ ulation. Vascular organization and intracellular Ca release kinetics in VSMCs could be quantified using automated image analysis based on open-source software CellProfiler and ImageJ on widefield or confocal images, setting the stage for use of the platform to study vascular (patho)physiology and therapy. multi-cell type 3D tissues and vessels-on-chip (VoC) (Tro- INTRODUCTION nolone and Jain, 2021). Typically, cells incorporated in Crosstalk between endothelial cells (ECs) and mural cells these microphysiological devices are derived from non-hu- (pericytes and vascular smooth muscle cells [VSMCs]) is man sources, human (tumor) cell lines, or directly from pri- pivotal for proper function of many blood vessels. mary human tissue. Primary human cells provide the Aberrant EC-mural cell crosstalk often leads to vascular closest mimic to human blood vessels but are of limited diseases that range from hypertension, atherosclerosis, availability and of variable genetic origin (Tronolone and vascular calcification, and coronary artery disease to Jain, 2021). While hiPSC derivatives are now regarded as stroke and other conditions (Owens et al., 2004). Animal an alternative, they have so far largely been used in combi- nation with primary cells in microfluidic chips. For models, including genetically modified mice, are widely example, human primary ECs, or hiPSC-ECs have been used to study vascular development and disease, but combined with human primary mural cells (Campisi these do not always capture patient phenotypes unless et al., 2018; van Dijk et al., 2020; van Duinen et al., 2019) the mutations are homozygous deletions, and differ- but not with mural cells derived from hiPSC, precluding ences associated with the genetic background and sus- opportunities to replicate (patient-specific) vascular dis- ceptibility are not evident in mice (Berry et al., 2019; eases originating in the mural cells. Van Norman, 2020). Human induced pluripotent stem Here, we overcome these limitations by incorporating cells (hiPSCs) generated from heathy individuals and pa- hiPSC-ECs with hiPSC-VSMCs in entirely hiPSC-based tients are a useful source of vascular cells and they do VoCs. We have previously described robust protocols to reflect the genetic background of the individual from derive ECs and VSMCs from multiple healthy hiPSC lines whom they are derived (Samuel et al., 2015). Several with little batch-to-batch variability (Halaidych et al., methods have been described to generate ECs and 2018, 2019; Orlova et al., 2014a, 2014b). The function- VSMCs from hiPSCs and some have already been ality of hiPSC-VoC was compared with similar VoCs con- used to model vascular disease-specific abnormalities taining hiPSC-ECs and human primary mural cells of the (Cochrane et al., 2019). Nevertheless, and despite recent advances, many current same development origin, namely human brain vascular in vitro models of blood vessels fail to emulate the inte- smooth muscle cells (HBVSMCs) and primary human grated, complex and multicell-type composition of the brain vascular pericytes (HBVPs). In all cases, we showed human vasculature and do not include a mimic of blood the microenvironment in the microfluidic device sup- flow (Duval et al., 2017). To address this, microfluidic de- ported the formation of a 3D perfusable, self-assembled vices have been engineered that do incorporate these fea- microvascular network. We optimised the culture condi- tures and provide the environment for the formation of tions and developed an automated quantification Stem Cell Reports j Vol. 16 j 2159–2168 j September 14, 2021 j ª 2021 The Author(s). 2159 This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). A Agglutinin + hiPSC-VSMCs Agglutinin + HBVSMCs Agglutinin + HBVPs + hiPSC-VSMCs + HBVSMCs + HBVPs (i) (i) (i) SOX17 Agglutinin SOX17 Agglutinin SOX17 Agglutinin (iii) (iii) (iii) (ii) (ii) (ii) D E Dextran (70 kDa) Agglutinin (i) Dextran (70 kDa) Agglutinin (iii) + hiPSC-VSMCs + hiPSC-VSMCs + HBVSMCs + HBVPs (ii) F H J K G I L M + hiPSC-VSMCs + HBVSMCs + HBVPs (i) Collagen IV CD31 (i) Collagen IV CD31 (i) Collagen IV CD31 (iii) (iii) (iii) (ii) (i) (i) (legend on next page) 2160 Stem Cell Reports j Vol. 16 j 2159–2168 j September 14, 2021 framework of the vascular network, mural cell hiPSC-EC sprouting (Figure 1Aiv). Vacuoles started to morphology, EC-mural cell interaction and extracellular appear in hiPSC-ECs 12–24 h after seeding (Figure S2A, matrix (ECM) composition. Finally, we demonstrated a white arrowheads) followed by proliferation and remodel- functional application: the automated quantification of ing up to 72 h. An interconnected microvascular network vasoactive responses. formed as early as day 2 (Figure S2A) and this spanned the complete microfluidic channel by day 7 (Figures 1B, S1A, and S2A). All combinations of mural cells with hiPSC-ECs RESULTS resulted in vascular lumen formation (Figure 1C). Further- more, these lumenized networks were perfusable by fluo- Characterization of the VoC model rescent beads (10 mm) or FITC-Dextran (70 kDa) (Figures Using the commercially available AIM Biotech 3D cell cul- 1D and 1E; Video S1) under gravity-driven flow. Microvas- ture chips, hiPSC-ECs (Halaidych et al., 2018; Orlova cular networks formed in the presence of hiPSC-VSMCs or et al., 2014a, 2014b) were combined with hiPSC-VSMCs primary mural cells showed similar morphologies, with (Halaidych et al., 2019), HBVSMCs or HBVPs (Figure 1Ai) no significant difference in vessel density (%, Figure 1F), in a fibrin hydrogel (Figure 1Aii) and the cell/gel mix was average vessel length (mm, Figure 1G), mean vessel diam- injected into the middle channel of the microfluidic chip eter (mm, Figure 1H), branching point (BP) density (BP/ (Figure 1Aiii). As a control, hiPSC-ECs in a fibrin hydrogel mm , Figure 1I), extravascular spaces (%, Figure 1J), or without mural cells were used (Figure S1). Endothelial number of hiPSC-ECs (Figure 1K). In contrast, microvas- growth medium-2 (EGM-2), supplemented with vascular cular networks formed by hiPSC-ECs alone were less orga- endothelial growth factor (VEGF) (50 ng/mL), was used nized than microvascular networks with hiPSC-VSMCs to support microvascular network formation. Microfluidic (Figure S1A). Although lumen formation was observed chips were perfused through gravity-driven flow by add- in microvascular networks formed using only hiPSC- ing 100 mL of medium to the inlet and 50 mL to the outlet ECs, these appeared irregular and broken (Figure S1B). of each medium channel. The gravity-driven flow was re- The instability of microvascular networks without established every 24 h allowing medium exchange in the hiPSC-VSMCs was also evidenced by increased leakage microfluidic chip (Table S1). On day 1, after hiPSC-ECs of FITC-Dextran from the vessel network (70 kDa; Fig- had begun to self-organize, g-secretase inhibitor DAPT ure S1I). There were, however, no significant differences (10 mM) was added to the medium for 24 h to promote in vessel density (%, Figure S1C), branching point (BP) Figure 1. Characterization of VoC (A) Schematic of the VoC protocol. hiPSC-ECs were cultured with hiPSC-VSMCs, HBVSMCs, or HBVPs (i). Cells were mixed in a fibrin hydrogel (ii) and injected into (AIM Biotech) microfluidic chips (iii). EGM-2 was supplemented with VEGF (50 ng/mL), chips were refreshed daily for 7 days. EGM-2 was also supplemented with DAPT (10 mM) on day 1 for 24 h (iv). (B) Representative immunofluorescence images of microvascular network showing hiPSC-EC (magenta; agglutinin) vessels spanning the complete length of the microfluidic channel. Images showing hiPSC-ECs cultured with hiPSC-VSMCs, HBVSMCs, or HBVPs, respectively (103). Scale bars, 200 mm. (C) Representative confocal images of microvascular network showing hiPSC-ECs (gray; agglutinin) and hiPSC-EC nuclei (cyan; SOX17). Images displaying xyz (i), xy (ii), and yz cross-sectional perspectives (iii). Images showing hiPSC-ECs cultured with hiPSC-VSMCs, HBVSMCs, or HBVPs, respectively (403). Scale bars, 100 mm. (D) Representative Immunofluorescence images showing hiPSC-ECs (magenta; agglutinin) and perfusion of 70 kDa FITC-Dextran (green). Images showing hiPSC-ECs cultured with hiPSC-VSMCs, HBVSMCs, or HBVPs, respectively (103). Scale bars, 50 mm. (E) Representative confocal images showing hiPSC-ECs (magenta; agglutinin) and perfusion of 70 kDa FITC-Dextran (green) in hiPSC-VoC. Images displaying xyz (i), xy (ii), and yz cross-sectional perspectives (iii) (403). Scale bars, 100 mm. (F–K) Quantification of vessel density (%) (F), average vessel length (mm) (G), mean diameter (mm) (H), branching point (BP) density (BPs/mm ) (I) extravascular spaces (%) (J), and number of hiPSC-ECs (K), from hiPSC-ECs cultured with hiPSC-VSMCs, HBVSMCs, or HBVPs, respectively, are shown. Data are shown as ±SD from N = 3, n = 6; three independent experiments with two microfluidic channels per experiment. (L) Representative confocal images of microvascular network showing hiPSC-ECs (magenta; CD31) and ECM (yellow; collagen IV). Images displaying xyz (i), xy (ii), and yz cross-sectional perspectives (iii). Images showing hiPSC-ECs cultured with hiPSC-VSMCs, HBVSMCs, or HBVPs, respectively (403). Scale bars, 100 mm. (M) Quantification of collagen IV density (%) from hiPSC-ECs cultured with hiPSC-VSMCs, HBVSMCs, or HBVPs, respectively, are shown. Data are shown as ±SD from N = 3, n = 8; three independent biological replicates with two to three microfluidic channels per experiment. One- way ANOVA with Tukey’s multiple comparison. See also Figures S1 and S2A–S2D and Video S1. Stem Cell Reports j Vol. 16 j 2159–2168 j September 14, 2021 2161 A SM22 Agglutinin + hiPSC-VSMCs SM22 Agglutinin + HBVSMCs SM22 Agglutinin + HBVPs + HBVSMCs + HBVPs + hiPSC-VSMCs (i) (i) SM22 Agglutinin SM22 Agglutinin (i) SM22 Agglutinin (iii) (iii) (iii) (ii) (ii) (ii) D E G H + hiPSC-VSMCs + HBVSMCs + HBVPs SM22 mCherry SM22 mCherry SM22 mCherry SM22 Intensity 0 40 (legend on next page) 2162 Stem Cell Reports j Vol. 16 j 2159–2168 j September 14, 2021 2 density (BP/mm , Figure S1F), extravascular space (%, Fig- significant difference in the cell length (mm, Figure 2D) or ure S1G), or number of hiPSC-ECs (Figure S1H) in micro- cell circularity (Figure 2E). Analysis of the contractile vascular networks formed either with or without hiPSC- marker SM22 in mural cells in microvascular networks VSMCs. By contrast, average vessel length (mm, Fig- showed a significantly lower normalized mean cell inten- ure S1D) and mean vessel diameter (mm, Figure S1E) sity in hiPSC-VSMCs compared with HBVSMCs (Figures were significantly reduced or increased respectively in 2F and 2G) while the total number of SM22 + cells was significantly lower in microvascular networks formed microvascular networks formed from hiPSC-ECs alone compared with microvascular networks formed in the with HBVSMCs (Figure 2H). Notably, we also observed presence of hiPSC-VSMCs. No significant differences that all mural cells displayed significantly higher SM22 were found in ECM deposition, evidenced by changes in staining intensity when in contact with hiPSC-ECs (Figures the relative density of collagen IV, between any cell com- 2F and 2I), which indicated that heterotypic cell-cell con- binations (Figures 1L, 1M, S1J, and S1K). Furthermore, the tact in VoC culture could further promote mural cell matu- presence of fibronectin was confirmed by immunostain- ration. Furthermore, long-term hiPSC-VoC culture resulted ing in microvascular networks formed with mural cells in an increase over time in the percentage of mural cells (Figure S2D). Finally, long-term culture of hiPSC-ECs in located close to the hiPSC-EC vessel wall (Figures S2E VoC with mural cells demonstrated that the microvas- and S2F). cular network architecture was stable and underwent 2+ continuous increase in vessel density over 21 days (Figures Assessment of hiPSC-VSMC Ca dynamics in the VoC S2A–S2C), although we observed an increase in density of model vessel networks formed with primary HBVSMCs on day To further assess the functionality of hiPSC-VSMCs in the 2+ 21 (Figures S2A–S2C). hiPSC-VoC, we measured intracellular Ca release in hiPSC-VSMCs engineered to express an ultra-sensitive 2+ Characterization of hiPSC-VSMCs and primary mural Ca sensor (GCaMP6f) (Chen et al., 2013). First, hiPSC- cells in the VoC model derived neural crest (NC) intermediates were transduced hiPSC-VSMCs, HBVSMCs, and HBVPs self-organized and with a lentiviral vector (LV) expressing GCaMP6f self-oriented toward the developing hiPSC-EC microvas- (Figure S3A). Transduction of hiPSC-NC cells with LVs en- cular networks, as early as day 1 (Figure S2E). On day 7, coding either enhanced green fluorescent protein or all mural cells were located at extravascular positions along GCaMP6f did not change expression of the surface the entire length of the microvascular network (Figure 2A), marker CD271 (Figure S3B). hiPSC-VSMCs engineered to 2+ surrounding hiPSC-EC lumens (Figure 2B). No significant express GCaMP6f showed no intracellular Ca release difference was observed in the percentage of hiPSC-VSMCs, upon perfusion with medium only (pre-stimulated, Figures 2+ HBVSMCs, or HBVPs associated with the hiPSC-EC lumen S3C and S3D) and similar intracellular Ca release upon (% mural cells localized at the vascular network, Figure 2C). stimulation with the vasoconstrictor endothelin-I (ET-I) Quantification of mural cell morphology also showed no (post-stimulated, Figures S3C and S3D) to Fluo-4-labeled Figure 2. Quantitative assessment of the structural proprieties of hiPSC-VSMCs and primary mural cells in VoC (A) Representative immunofluorescence images of microvascular network showing the hiPSC-ECs (magenta; agglutinin) and mural cells (green; SM22) derived vasculature unit spanning the complete length of microfluidic channel. Images showing hiPSC-ECs cultured with hiPSC-VSMCs, HBVSMCs, or HBVPs, respectively (103). Scale bars, 200 mm. (B) Representative confocal images of microvascular network showing hiPSC-ECs (gray; agglutinin) and mural cells (green; SM22). Images displaying xyz (i), xy (ii), and yz cross-sectional perspectives (iii). Images showing hiPSC-ECs cultured with hiPSC-VSMCs, HBVSMCs, or HBVPs, respectively (403). Scale bars, 100 mm. (C–E) Quantification of the percentage of mural cells associated with the hiPSC-EC lumen (% mural cells localized at the vascular network) (C), mean mural cell length (mm) (D), and mural cell circularity factor (the circle is 1) (E) in hiPSC-VSMCs, HBVPs, and HBVSMCs. Data are shown as ±SD from N = 3, n = 6; three independent experiments with two microfluidic channels per experiment. (F) (Top) Representative confocal images of microvascular network showing hiPSC-ECs (gray; mCherry) and mural cells (red; SM22). (Bottom) Representative surface-rendered objects of confocal images showing microvascular network (gray; mCherry) and mural cells (colour-coded scale representing SM22 intensity). Images showing hiPSC-ECs cultured with hiPSC-VSMCs, HBVSMCs, or HBVPs, respectively (403). Scale bars, 100 mm. (G–I) Quantification of normalized mean cell SM22 intensity (G), number of SM22 + cells (H), and normalized mean cell SM22 intensity of mural cells in contact with hiPSC-ECs (I) in hiPSC-VSMCs, HBVPs and HBVSMCs. Intensity was normalized to hiPSC-VSMCs. Data are shown as ±SD from N = 3, n = 6; three independent experiments with two microfluidic channels per experiment. One-way ANOVA (C–E and G–H) and two-way ANOVA (I) with Tukey’s multiple comparison. *p < 0.05, **p < 0.001, ***p < 0.0001. See also Figures S2E and S2F. Stem Cell Reports j Vol. 16 j 2159–2168 j September 14, 2021 2163 hiPSC-VSMCs (Halaidych et al., 2019). Next, intracellular assumed positions surrounding the vascular wall, support- 2+ Ca release in hiPSC-VSMCs in the microvascular network ing the vessel and maintaining its functionality in 3D. was examined prior to- (basal state) and after medium Although we did not find any morphological differences refreshment on day 7. Significantly higher GCaMP6f fluo- in microvascular network organization and mural cell rescence was observed across the entire microvascular morphology between hiPSC-VSMCs and primary mural network after medium refreshment (Figures 3A and 3B; cells, expression of the contractile marker SM22 did differ Video S2). We then combined medium refreshment with between hiPSC-VSMCs and HBVSMCs. hiPSC-VSMCs and ET-1 stimulation (1 mM, Figure 3C; Video S2). To compare HBVPs showed lower SM22 staining intensities and higher 2+ the divergence in Ca responses, the average fluorescence total cell numbers in VoC culture. This could indicate that intensity of regions of interest over time (F/F ) was exam- hiPSC-VSMCs are less differentiated in VoC culture. ined by adapting previous methods (Halaidych et al., Notably, by incorporating and monitoring fluorescently 2+ 2019)(Figure 3D). Comparison of Ca kinetic parameters, tagged hiPSC-derived vascular cells (Roberts et al., 2017), measured at the half-maximum level (F/F0) , showed we showed that the most important steps of vascular max significantly higher amplitudes (F/F , Figure 3E), intensity network formation and remodeling occurred in the first over time (AUC [F*s/F0], Figure 3F) and duration, and 7 days of VoC culture. We also showed that these microvas- 2+ slower decay of the Ca transient without changes in the cular networks are stable for up to 3 weeks. Vessel density time to peak (s, Figures 3G–3I) upon ET-I stimulation. increased over time, although not much remodeling occur- ring beyond day 7 of culture. Specifically, microvascular networks formed with primary HBVSMCs showed the Modeling hiPSC-derived EC-VSMC crosstalk in the highest increase in vessel density, indicating that an in- VoC model crease in hiPSC-VSMC number observed on day 7 might To demonstrate the potential utility of hiPSC-VoC for dis- be advantageous for supporting long-term VoC culture. ease modeling, we examined the loss of EC-VSMC crosstalk Moreover, we demonstrated that hiPSC-VSMCs in the upon blocking NOTCH signaling using the small-molecule VoC were responsive to vasoactive stimulation by quanti- g-secretase inhibitor DAPT (10 mM). The addition of DAPT 2+ fying changes in Ca kinetic parameters. We noted that (10 mM, day 5–7) affected overall microvascular network ar- hiPSC-VSMCs were activated simply after medium refresh- chitecture (Figure S4A) with significant changes in the ment, an important consideration for proper experimental mean vessel diameter although vessel density was similar control. It seemed likely this was mediated by factors in the between the groups (Figures S4B and S4C). DAPT supple- fresh media since shear-stress during gravity-mediated flow mentation had no significant effect on the percentage of in the VoC was too slow to be stimulatory. In addition, we hiPSC-VSMCs associated with the hiPSC-EC lumen (% confirmed more pronounced and coordinated hiPSC- mural cells localized at the vascular network, Figures 4A 2+ VSMC intracellular Ca release following addition of the and 4B). However, hiPSC-VSMCs showed a significant vasoconstrictor ET-I. decrease in the cell length (mm, Figure 4C) with a signifi- Conventional preclinical models have shown low suc- cant increase in cell circularity (Figure 4D). Analysis of cess in accurately predicting drug efficacy and toxicity in the contractile marker SM22 in hiPSC-VSMCs showed a human trials (Van Norman, 2020). We considered it essen- significantly lower normalized mean SM22 intensity (Fig- tial therefore to provide some evidence that hiPSC-VoC re- ures 4E and 4F) while no change in the total number of sponded to drugs as expected and that this property was SM22 + cells was observed, following addition of DAPT to conserved across different healthy hiPSC lines and batches. the cultures (Figure 4G). This indeed was the case: under conditions of vessel matu- ration where the NOTCH signaling pathway coordinates heterologous cell-cell crosstalk and maintains vessel integ- DISCUSSION rity, we showed that these features were disrupted in the This report describes the generation of hiPSC-derived hiPSC-VoC after inhibition by DAPT on days 5–7. In partic- microvascular networks composed of hiPSC-ECs and ular, DAPT addition appeared to reduce acquisition of a hiPSC-VSMCs. We showed that hiPSC-ECs form an inter- contractile-like identity by hiPSC-VSMCs. This notion connected microvascular network with perfusable lumens was supported by reduced expression the contractile and ECM deposits as in previous microfluidic studies based marker SM22 and a less elongated morphology after EC- on other cell sources (Belair et al., 2015; Campisi et al., VSMC crosstalk inhibition by DAPT. We also observed 2018). We demonstrate that although hiPSC-ECs alone gradual reversal of vascular stability and signs of vessel can form a microvascular network, inclusion of mural cells regression, reflecting what is known about dysfunctional EC-VSMC crosstalk in developed vessels in vivo (Kerr facilitates hiPSC-EC self-organization and supports vessel et al., 2016). stability. Much like primary mural cells, hiPSC-VSMCs 2164 Stem Cell Reports j Vol. 16 j 2159–2168 j September 14, 2021 A Basal state mCherry GCaMP6f Media refreshment mCherry GCaMP6f C D Pre-stimulated Post-stimulated (ET-I 1μM) mCherry GCaMP6f mCherry GCaMP6f E F G H I 2+ Figure 3. Analysis of hiPSC-VSMCs Ca dynamics in VoC 2+ (A) Representative immunofluorescent images of intracellular Ca fluorescence showing hiPSC-ECs (gray; mCherry) and hiPSC-VSMCs (green; GCaMP6f) without- (basal state) and after EGM-2 refreshment on day 7 (103). Scale bar, 100 mm. (B) Normalized GCaMP6f intensity at day 7. GCaMP6f intensity was normalized to the condition prior to EGM-2 refreshment (basal state). Data are shown as ±SD of N = 3, n = 21; three independent experiments with seven microfluidic channels per experiment. 2+ (C) Representative confocal images of intracellular Ca fluorescence showing with hiPSC-ECs (gray; mCherry) and hiPSC-VSMCs (green; GCaMP6f) in pre- and post-stimulated (ET-I, 1 mM) states (203). Scale bar, 100 mm. (D) Normalized average fluorescence intensity F/F in hiPSC-VSMCs expressing GCaMP6f. Medium channels were gravity-flow perfused with EGM-2 alone or containing ET-I (1 mM). Stimulation time point is set as t = 5(s). Data are shown as ±SD of N = 3, n = 9; three independent experiments with three microfluidic channels per experiment. 2+ (E–I) Ca transient parameters: amplitude F/F (E), duration (s) (F), area under the curve (AUC, F*s/F ) (G), time to peak (s) (H), and 0 0 decay (s) (I) of channels gravity-flow perfused with EGM-2 alone or containing ET-I (1 mM). Data are shown as ±SD of N = 3, n = 9; three independent experiments with three microfluidic channels per experiment. Paired (B) Student’s t test. Wilcoxon-Mann-Whitney test (E–I). *p < 0.05, **p < 0.001, ***p < 0.0001; ns, not significant. See also Figure S3 and Video S2. Stem Cell Reports j Vol. 16 j 2159–2168 j September 14, 2021 2165 A B C FG Figure 4. Modeling loss of EC-VSMC crosstalk in VoC (A) Representative confocal images of microvascular network showing in hiPSC-ECs (gray; GFP) and hiPSC-VSMCs (green; RFP) in control and DAPT (10 mM) supplemented conditions. Images displaying xyz (i), xy, (ii), yz cross-sectional perspectives (iii), and enlargements of white framed areas (iv) (403). Scale bars, 100 mm in (i–iii) and 50 mm in (iv). (B–D) Quantification of the percentage of hiPSC-VSMCs associated with the hiPSC-EC lumen (% mural cells localized at the vascular network) (B), mean hiPSC-VSMCs length (mm) (C), and hiPSC-VSMC circularity factor (the circle is 1) (D) in control and DAPT (10 mM) supplemented conditions at day 7. Data are shown as ±SD from N = 3, n = 27; three independent experiments with nine microfluidic channels per experiment. (E) (Left) Representative confocal images of microvascular network showing hiPSC-ECs (gray; mCherry) and hiPSC-VSMCs cells (red; SM22). (Right) Representative surface-rendered objects of confocal images showing microvascular network (gray; mCherry) and hiPSC-VSMCs (colour-coded scale representing SM22 intensity) in control and DAPT (10 mM) supplemented conditions (403). Scale bars, 100 mm. (legend continued on next page) 2166 Stem Cell Reports j Vol. 16 j 2159–2168 j September 14, 2021 middle gel-loading channel of the microfluidic chip. Chips were In summary, we have demonstrated several advantages incubated at room temperature for 15 min before the addition of of this hiPSC-VoC model above existing models: (1) the EGM-2 supplemented with VEGF (50 ng/mL) to both flanking me- possibility of studying genetic vascular diseases with dia channels. The g-secretase inhibitor DAPT (10 mM) was also patient-specific lines; (2) evaluation of real-time changes added to the medium on day 1 for 24 h. Gravity-driven flow was in vascular structure and function by incorporating fluores- induced by the addition of 100 mL medium to the right media ports cently tagged cells; (3) the ability to measure and quantify and 50 mL media to left media ports. Medium was refreshed daily. effects of drugs affecting EC-mural cell crosstalk with view to modulating vessel stability and integrity. Nevertheless, Statistical analysis the model could be further improved by: (1) introducing Statistical analyses were performed using GraphPad Prism 9 soft- physiological (rather than gravity-driven) flow which ware. Normality of the data was evaluated by the D’Agostino-Pear- would recapitulate vessel shear forces and geometry in son test. One-way and two-way ANOVA with Tukey’s multiple healthy and disease environments and (2) the addition of comparison test was used for the analysis of three groups. For other hiPSC-derived cells such as monocytes or astrocytes paired or unpaired analysis of two groups, either Student’s t test which would allow mimicking of (isogenic) inflammatory or Wilcoxon-Mann-Whitney test was used. Analyses are indicated reactions or features of the brain. in the figure legends. The data are reported as mean ± SD. Statistical In conclusion, the hiPSC-VoC model described in this significance was defined as p < 0.05. paper will be useful to study and quantify changes in the vascular architecture and function during vascular devel- SUPPLEMENTAL INFORMATION opment or upon drug treatment. Clinically, this may trans- Supplemental information can be found online at https://doi.org/ late into a better understanding of vascular disease condi- 10.1016/j.stemcr.2021.08.003. tions and predicting drug efficiency. AUTHOR CONTRIBUTIONS EXPERIMENTAL PROCEDURES Conceptualization, V.V.O.; methodology, M.V.C., A.C., and V.V.O.; software, A.C.; validation, A.C. and M.V.C.; formal analysis, A.C. Full details are provided in supplemental experimental procedures. and M.V.C.; investigation, A.C., M.V.C., and F.E.v.d.H.; visualisa- tion, M.V.C. and A.C.; resources, A.A.F.d.V. and V.V.O.; writing – hiPSC lines original draft, M.V.C., A.C., C.L.M., and V.V.O.; writing – review Research on hiPSC was approved by the medical ethical committee & editing, M.V.C., A.C., S.A.J.L.O., C.L.M., and V.V.O.; supervision, at Leiden University Medical Center, the Netherlands. A detailed S.A.J.L.O, C.L.M., and V.V.O.; project administration, V.V.O.; fund- list of the hiPSC lines and batches used for each experiment is pro- ing acquisition, A.C., S.A.J.L.O., C.L.M., and V.V.O. vided in Table S2. CONFLICT OF INTERESTS Differentiation of hiPSC-ECs and hiPSC-VSMCs The authors declare no competing interests. hiPSC differentiation to ECs was performed as described previously (Orlova et al., 2014a, 2014b). hiPSC differentiation to VSMC was ACKNOWLEDGMENTS performed as previously described (Halaidych et al., 2019). The LUMC human iPSC Hotel for generation and characterization Setting up VoCs of hiPSC lines; LUMC’s confocal imaging facility; Cindy Bart, Sven hiPSC-ECs and mural cells were prepared prior to incorporation in Dekker and Juan Zhang for LV production; Oleh Halaidych for use- VoCs as described in supplemental experimental procedures. ful discussion on hiPSC-VSMCs. The Allen Cell Collection, avail- Commercially available microfluidic chips with one gel channel able from Coriell Institute for Medical Research, provided mate- and two media channels (AIM Biotech) were used. Cells were resus- rials. Images were generated using Biorender.com. pended and combined to obtain 10 3 10 hiPSC-ECs/mL and 2 3 This work was supported by the Netherlands Organisation for 10 mural cells/mL (5:1 ratio). Three different mural cell suspen- Health Research and Development (ZonMw): PTO 446002501 sions were tested in combination with hiPSC-ECs: (1) hiPSC- and VIDI 91717325; the Netherlands Organ-on-Chip Initiative VSMCs, (2) HBVSMCs, and (3) HBVPs. Cell were resuspended in which is an NWO Gravitation project (024.003.001) funded by EGM-2 supplemented with Thrombin (4 U/mL) and then gently the Ministry of Education, Culture and Science of the government mixed with fibrinogen (final concentration 3 mg/mL, Sigma) at of the Netherlands; European Research Council (ERCAdG 1:1 vol ratio. Cell/hydrogel mixture was quickly loaded into the 323182 STEMCARDIOVASC); the European Union’s Horizon (F and G) Quantification of normalized mean cell SM22 intensity (F) and number of SM22 + cells (G). Intensity was normalized to control condition. Data are shown as ±SD from N = 3, n = 9; three independent experiments with three microfluidic channels per experiment. Wilcoxon-Mann-Whitney test. *p < 0.05, **p < 0.001, ***p < 0.0001; ns, not significant. See also Figure S4. Stem Cell Reports j Vol. 16 j 2159–2168 j September 14, 2021 2167 2020 research and innovation program under the Marie Sklodow- Kerr, B.A., West, X.Z., Kim, Y.W., Zhao, Y., Tischenko, M., Cull, ska Curie grant agreement no. 707404. R.M., Phares, T.W., Peng, X.D., Bernier-Latmani, J., Petrova, T.V., et al. (2016). Stability and function of adult vasculature is sustained Received: March 1, 2021 by Akt/Jagged1 signalling axis in endothelium. Nat. Commun. 7, Revised: August 5, 2021 Accepted: August 6, 2021 Orlova, V.V., Drabsch, Y., Freund, C., Petrus-Reurer, S., van den Hil, Published: September 2, 2021 F.E., Muenthaisong, S., Dijke, P.T., and Mummery, C.L. (2014a). Functionality of endothelial cells and pericytes from human REFERENCES pluripotent stem cells demonstrated in cultured vascular plexus and zebrafish xenografts. Arterioscler. Thromb. Vasc. Biol. 34, Belair, D.G., Whisler, J.A., Valdez, J., Velazquez, J., Molenda, J.A., 177–186. Vickerman, V., Lewis, R., Daigh, C., Hansen, T.D., Mann, D.A., et al. (2015). Human vascular tissue models formed from human Orlova, V.V., van den Hil, F.E., Petrus-Reurer, S., Drabsch, Y., Ten induced pluripotent stem cell derived endothelial cells. Stem Cell Dijke, P., and Mummery, C.L. (2014b). Generation, expansion Rev. Rep. 11, 511–525. and functional analysis of endothelial cells and pericytes derived from human pluripotent stem cells. Nat. Protoc. 9, 1514–1531. Berry, C., Sidik, N., Pereira, A.C., Ford, T.J., Touyz, R.M., Kaski, J.C., and Hainsworth, A.H. (2019). Small-vessel disease in the heart and Owens, G.K., Kumar, M.S., and Wamhoff, B.R. (2004). Molecular brain: current knowledge, unmet therapeutic need, and future di- regulation of vascular smooth muscle cell differentiation in devel- rections. J. Am. Heart Assoc. 8, e011104. opment and disease. Physiol. Rev. 84, 767–801. Campisi, M., Shin, Y., Osaki, T., Hajal, C., Chiono, V., and Kamm, Roberts, B., Haupt, A., Tucker, A., Grancharova, T., Arakaki, J., Fu- R.D. (2018). 3D self-organized microvascular model of the human qua, M.A., Nelson, A., Hookway, C., Ludmann, S.A., Mueller, I.A., blood-brain barrier with endothelial cells, pericytes and astrocytes. et al. (2017). Systematic gene tagging using CRISPR/Cas9 in hu- Biomaterials 180, 117–129. man stem cells to illuminate cell organization. Mol. Biol. Cell 28, Chen, T.W., Wardill, T.J., Sun, Y., Pulver, S.R., Renninger, S.L., Bao- 2854–2874. han, A., Schreiter, E.R., Kerr, R.A., Orger, M.B., Jayaraman, V., et al. Samuel, R., Duda, D.G., Fukumura, D., and Jain, R.K. (2015). (2013). Ultrasensitive fluorescent proteins for imaging neuronal activity. Nature 499, 295–300. Vascular diseases await translation of blood vessels engineered from stem cells. Sci. Transl. Med. 7, 309rv306. Cochrane, A., Albers, H.J., Passier, R., Mummery, C.L., van den Berg, A., Orlova, V.V., and van der Meer, A.D. (2019). Advanced Tronolone, J.J., and Jain, A. (2021). Engineering new microvascular in vitro models of vascular biology: human induced pluripotent networks on-chip: ingredients, assembly, and best practices. Adv. stem cells and organ-on-chip technology. Adv. Drug Deliv. Rev. Funct. Mater. 31, 2007199. 140, 68–77. van Dijk, C.G.M., Brandt, M.M., Poulis, N., Anten, J., van der Moo- Duval, K., Grover, H., Han, L.H., Mou, Y., Pegoraro, A.F., Fredberg, len, M., Kramer, L., Homburg, E., Louzao-Martinez, L., Pei, J., Kreb- J., and Chen, Z. (2017). Modeling physiological events in 2D vs. 3D ber, M.M., et al. (2020). A new microfluidic model that allows cell culture. Physiology (Bethesda) 32, 266–277. monitoring of complex vascular structures and cell interactions Halaidych, O.V., Cochrane, A., van den Hil, F.E., Mummery, C.L., in a 3D biological matrix. Lab Chip 20, 1827–1844. and Orlova, V.V. (2019). Quantitative analysis of intracellular van Duinen, V., Zhu, D., Ramakers, C., van Zonneveld, A.J., Vulto, Ca(2+) release and contraction in hiPSC-derived vascular smooth P., and Hankemeier, T. (2019). Perfused 3D angiogenic sprouting in muscle cells. Stem Cell Reports 12, 647–656. a high-throughput in vitro platform. Angiogenesis 22, 157–165. Halaidych, O.V., Freund, C., van den Hil, F., Salvatori, D.C.F., Rimi- nucci, M., Mummery, C.L., and Orlova, V.V. (2018). Inflammatory Van Norman, G.A. (2020). Limitations of Animal studies for pre- responses and barrier function of endothelial cells derived from dicting toxicity in clinical trials: Part 2: potential alternatives to human induced pluripotent stem cells. Stem Cell Reports 10, the use of animals in preclinical trials. JACC Basic Transl. Sci. 5, 1642–1656. 387–397. 2168 Stem Cell Reports j Vol. 16 j 2159–2168 j September 14, 2021

Journal

Stem Cell Reports – Unpaywall

Published: Sep 1, 2021

There are no references for this article.