3D culturing of human pluripotent stem cells-derived endothelial cells for vascular regeneration

Abstract

Rationale: Human induced pluripotent stem cell-derived endothelial cells can be candidates for engineering therapeutic vascular grafts. Methods: Here, we studied the role of three-dimensional culture on their characteristics and function both in vitro and in vivo. Results: We found that differentiated hPSC-EC can re-populate decellularized biomatrices; they remain viable, undergo maturation and arterial/venous specification. Human PSC-EC develop antifibrotic, vasoactive and anti-inflammatory properties during recellularization. In vivo, a robust increase in perfusion was detected at the engraftment sites after subcutaneous implantation of an hPSC-EC-laden hydrogel in rats. Histology confirmed survival and formation of capillary-like structures, suggesting the incorporation of hPSC-EC into host microvasculature. In a canine model, hiPSC-EC-seeded onto decellularised vascular segments were functional as aortic grafts. Similarly, we showed the retention and maturation of hiPSC-EC and dynamic remodelling of the vessel wall with good maintenance of vascular patency. Conclusions: A combination of hPSC-EC and biomatrices may be a promising approach to repair ischemic tissues.

Keywords: angiogenesis tracking; endothelial cells; human pluripotent stem cells; multimodality imaging; tissue-engineered vascular grafts.

Figure 1

Differentiation of human embryonic stem cells, human induced pluripotent stem cell resulted in native-like endothelial phenotype. (A) Phase-contrast microscopy images of hESC-EC, hiPSC-EC and HUVEC demonstrating the classical endothelial cobblestone morphology in adherent monolayer culture on collagen. Representative immunofluorescent images for (B) von Willebrand factor, vWf and (C) CD31 indicated that these endothelial-associated proteins were highly expressed at day 19 of differentiation. Scale bars, 50 µm. (D) RNA-seq-based transcriptome comparison between hESC-EC, hiPSC-EC and HUVEC grown on collagen surface. Venn diagram showing the total number in each circle represents the amount of differentially expressed genes between the different comparisons. Only the annotated genes were considered. (E) Heat map of shared canonical pathways of the three cell types.

Figure 2

Functional endothelial cell-laden hydrogel implants. (A) Representative images showing tube formation angiogenic activity by capillary structures of hESC-EC, hiPSC-EC and HUVAC on growth factor-reduced Matrigel. (B) Hematoxylin-eosin staining of subcutaneous implants of hESC-EC, hiPSC-EC or HUVAC in Matrigel hydrogel in nude rats 14 days after subcutaneous implantation are shown. Scale bars, 100 µm. Presence of red blood cells in tubular structures suggests a functional link to recipient circulation, inset panels. See also Video S1 and S2 on 3DHisTech analysis of 3D vascular structures. (C) Magnetic resonance imaging (MRI, co-registration with (D) positron emission tomography (PET, showing (E) ⁶⁸Ga-NOTA-RDG2 uptake in the endothelial cell-implanted animals. PET and PET/MRI images feature white crosshairs at the plug implantation and increased vascularization sites. MRI image panel is shown without the same for ease of anatomical assessment. (F) [⁶⁸Ga]-NOTA-RDG2 expressed as maximum standardized uptake values (SUV_max) in donor hydrogel (solid bar) and surrounding recipient tissue are also measured (stripped bar), mean ± SEM, n = 4-6, Kruskal-Wallis nonparametric test.

Figure 3

In vivo multimodality imaging shows increased perfusion in response to hPSC-EC. Comparison of hydrogel-based endothelial plugs on ^98mTc-human serum albumin (HAS) to assess perfusion of newly formed vessels. (A) Representative CT scan to show whole animal. (B) Multimodality quantitative SPECT/CT imaging revealed increased perfusion after hiPSC-EC, (C) human cell-free Matrigel, (D) hESC-EC and (E) HUVEC plugs. White crosshairs showing the site of implants. (F) Bar graph showing SUV_max values were calculated, mean ± SEM. (G-H) Grouped bar diagrams show expression of arterial (ephrin B2, Notch1, Notch2), venous (EphB4) and common (CD31) endothelial marker genes at two weeks after subcutaneous transplantation of hESC-EC (G) and hiPSC-EC (H) in athymic rats. mRNA levels are normalized to those in pre-implanted control cells, n = 6. * P < 0.05, ** P < 0.01, *** P < 0.001, one-way ANOVA. See also Figure S2.

Figure 4

(A) Previously decellularised vessel walls were recellularised with either hESC-EC, hiPSC-EC or HUVEC and stained with anti-human CD31 antibody (BDBiosciences, green). Hoeschst for nuclear staining and respective negative controls are also shown. Nuclei are counterstained with Hoechst (blue). (B) Live cells stained with vital dyes calcein AM (green, excitation at 495 and emission at 515 nm) and necrosis marker TO-PRO3 (excitation at 642 nm and emission at 661 nm). Scale bars, 50 µm (A) and 100 µm (B).

Figure 5

RNAseq-based transcriptomics profiling of endothelial cells cultured either on collagen or decellularised vascular matrix. (A-D) Venn diagrams showing the differential expression profile of H7 hESC-EC, IMR hiPSC-EC and HUVEC cultured on collagen vs. decellularised vascular wall. (E-G) Volcano plots and (H) principal component analysis show differential expressions of genes in the three cell types. The first two principal components of the gene expression dataset are plotted here for each of the samples. (I-K) Enriched GO terms shows as dot plots. The 19 GO processes with the largest gene ratios are plotted in order of gene ratio. The size of the dots represents the number of genes in the significant differentially expressed gene list associated with the GO term and the colour of the dots represent the P-adjusted values. (L) RNA-seq-based heat map showing angiogenesis and cell-matrix adhesion protein profiling in 3D vs 2D endothelial cultures. (M) Functional association network diagram EC are generated by String DB pathway analysis. n = 3 biological replicates. (N) Angiogenesis and cell-matrix adhesion protein profiling in 2D and 3D endothelial cultures as well as native aortic and hiPSC-derived smooth muscle cells (SMC). Contractile and synthetic SMC are both shown. Further controls used are human microvascular endothelial cells (HMVEC) and human coronary artery endothelial cells (HCAEC-c). Heat map shows expression of angiogenesis-related proteins of hESC-EC and hiPSC-EC (2D and 3D cultures, respectively). Proteomics in hESC-EC and hiPSC-EC are normalised as z scores. n = 2 biological replicates, 4 technical replicates. Array membranes were visualized by chemiluminescent detection; pixel densities were quantified by ImageJ software. Protein quantities were equalized in each experimental setting; passages between 3 and 5 were used.

Figure 6

Surgical procedure and imaging of recellularised aortic graft implantation in canine model. (A) hiPSC-EC graft implantation into the peripheral artery by performing end-to-end anastomosis in dog infrarenal aorta, n = 4. (B-D) 3D and multiplanar reconstruction CT images of the implanted hiPSC-EC graft. Grey arrows showing the site of the graft. (E-F) Hematoxylin-eosin, (G-I) orcein and (J-L) Masson's staining before surgery (PRE) and at one-week follow-up (POST) **P < 0.01, paired Student's t-test. Immunohistochemistry of α-smooth muscle actin (red), human nuclei (green) and nucleus label DAPI (blue) in the luminal segment of the aortic wall at day 7 after surgery (M). Scatter-plot and bar graphs show mRNA levels of canine ACTA2 at implantation and at day 7 after surgery (N). Normalized mRNA levels are shown as 2-ΔΔCt, mean ± SEM, *P < 0.05, Wilcoxon matched-pairs signed-rank test, from n = 4 animals. y-axis, note log scale.

Figure 7

In vivo retention of hIPSC-EC on recellularized vessel wall. (A) Confocal fiberoptic endo-microscopy imaging to track fluorescently labelled cells seeded onto luminal surface of the decellularised vessel wall before, during, and 1 week after surgery. QTracker (QT) 525 and -655 vital dyes are used to improve cell signal to background and non-specific autofluorescence ratio. Resolution is 3 µm, with 8 Hz sampling rate. The field-of-view is 800 µm in diameter. Empty arrows point to hiPSC-EC with fluorescent signal in both channels, white arrows to cells that accumulated QT515 (green) only. (B) Human iPSC-EC were costained for human nuclei Ku80 (green) and anti-human CD31 (red) in the luminal segment of the aortic wall after recellularization by immunohistochemistry. Nuclei were counterstained with DAPI (blue). Scatter-plot and bar graphs show mRNA levels of (C) dog CD31 (D) human CD31, (E) human mesenchymal marker FSP1, (F) human versus dog GAPDH and (G-J) endothelial specification (Notch1, 2, ephrin B2, EphB4, at implantation and at day 7 after surgery. mRNA levels are shown as 2-dCt, mean ± SEM, n = 5-6 samples, from 4 animals. *P < 0.05, **P < 0.01, Student's t-test. (K) Ingenuity Pathway analysis of the genes tested.