Enhanced therapeutic and long-term dynamic vascularization effects of human pluripotent stem cell-derived endothelial cells encapsulated in a nanomatrix gel

Abstract

BACKGROUND

Human pluripotent stem cell (hPSC)-derived endothelial cells (ECs) have limited clinical utility due to undefined components in the differentiation system and poor cell survival in vivo. Here, we aimed to develop a fully defined and clinically compatible system to differentiate hPSCs into ECs. Further, we aimed to enhance cell survival, vessel-formation, and therapeutic potential by encapsulating hPSC-ECs with a peptide amphiphile (PA) nanomatrix gel.

METHODS

We induced differentiation of hPSCs into the mesodermal lineage by culturing on collagen-coated plates with a GSK3β inhibitor. Next, VEGF, EGF, and bFGF were added for endothelial lineage differentiation followed by sorting for CDH5 (VE-Cadherin). We constructed an extracellular matrix-mimicking PA nanomatrix gel (PA-RGDS) by incorporating the cell adhesive ligand Arg-Gly-Asp-Ser (RGDS) and a matrix metalloproteinase-2 degradable sequence. We then evaluated whether the encapsulation of hPSC-CDH5⁺ cells in PA-RGDS could enhance long-term cell survival and vascular regenerative effects in a hindlimb ischemia model using Laser Doppler perfusion imaging, bioluminescence imaging, real-time RT-PCR, and histological analysis.

RESULTS

The resultant hPSC-derived CDH5⁺ cells (hPSC-ECs) showed highly enriched and genuine EC characteristics and pro-angiogenic activities. When injected into ischemic hindlimbs, hPSC-ECs showed better perfusion recovery and higher vessel-forming capacity compared to media-, PA-RGDS-, or HUVEC-injected groups. However, the group receiving the PA-RGDS-encapsulated hPSC-ECs showed better perfusion recovery, more robust and longer cell survival (> 10 months), and higher and prolonged angiogenic and vascular incorporation capabilities than the bare hPSC-EC-injected group. Surprisingly, the engrafted hPSC-ECs demonstrated previously unknown sustained and dynamic vessel-forming behavior: initial perivascular concentration, a guiding role for new vessel formation, and progressive incorporation into the vessels over 10 months.

CONCLUSION

We generated highly enriched hPSC-ECs via a clinically compatible system. Further, this study demonstrated that a biocompatible PA-RGDS nanomatrix gel substantially improved long-term survival of hPSC-ECs in an ischemic environment and improved neovascularization effects of hPSC-ECs via prolonged and unique angiogenic and vessel-forming properties. This PA-RGDS-mediated transplantation of hPSC-ECs can serve as a novel platform for cell-based therapy and investigation of long-term behavior of hPSC-ECs.

INTRODUCTION

Ischemic cardiovascular diseases are the most common cause of morbidity and mortality in industrialized countries. As the loss of vascular supply is a main pathophysiologic feature of these diseases, therapies restoring this fundamental deficit should target growth of blood vessels, for which endothelial cells (ECs) play a crucial role.

Recently, hPSCs, which include human embryonic stem cells (hESCs) and human induced pluripotent stem cells (hiPSCs), have emerged as a promising candidate for vascular regeneration therapy because they have unquestionable EC differentiation capacity^1–5. However, there are two significant hurdles in their path towards clinical applicability: development of a clinically compatible system for generating ECs and low retention and survival of cells in vivo. In particular, despite its critical importance for clinical applicability, measures to increase cell survival in ischemic tissues were not previously addressed.

For hPSC-ECs to be used for clinical applications, the following criteria at least must be met. hPSCs need to be differentiated under defined conditions free of xenogeneic components. The culture system should yield ECs at high efficiency, which must be verified in multiple cell lines. The differentiated ECs should be highly pure and must be functional in vivo. Their therapeutic effects should be confirmed in animal models of cardiovascular disease. Finally, tumorigenic potential or adverse effects of implanted cells must be ruled out by long-term follow-up. While other studies have investigated one or several aspects of these points, no studies addressed all these issues, which has precluded clinical translation of hPSC-derived ECs.

Moreover, low survival of transplanted cells in ischemic tissues poses a critical barrier for cell therapy even with hPSC-derived cells. When adult cells were injected into ischemic hearts, > 50% of the cells disappeared immediately, and virtually all cells were lost within a month^{6, 7}. The survival of hESC-ECs injected into ischemic hearts was not much different: < 10% of the injected cells survived at 1 week and 1% or less remained at 8 weeks⁸. hPSC-derived cardiomyocytes showed poor survival in ischemic hearts, with most of them disappearing usually within a month⁹. Similar trends were observed when cells were injected into a hindlimb ischemia model. Monitoring of cell survival using bioluminescent imaging (BLI) demonstrated that mesenchymal stem cells (MSCs) disappeared by 6 days¹⁰ or 20 days¹¹ post cell transplantation. Survival of hPSC-ECs in a hindlimb ischemia model was reported by only one study, which showed no detectable bioluminescence signals after 14 days¹². These studies question the advantages of using hPSC-derived ECs over adult cells if they function through short-term paracrine effects without durable vascular incorporation. Moreover, due to short-term cell survival, there has been no way to investigate how hPSC-ECs behave and contribute to vessel formation in ischemic tissues.

To overcome low cell survival in vivo, various biomaterial-based strategies have been attempted^13–17. However, to date, there have not been any studies investigating the effects of biomaterial-mediated delivery of hPSC-ECs in any cardiovascular disease models. In this study, we investigated whether biomaterial-mediated delivery of hPSC-ECs could prolong cell survival and thereby improve therapeutic neovascularization, and if, due to increased cell survival, we could elucidate previously unknown behaviors of hPSC-ECs in vivo.

We selected a novel injectable self-assembled peptide amphiphile (PA) nanomatrix gel, which has emerged as a highly promising biomaterial to enhance cell engraftment and survival^{16, 18, 19}. PAs are peptide-based molecules consisting of a short hydrophobic sequence linked to a hydrophilic, bioactive peptide sequence that self-assemble into high aspect ratio nanofibers at physiological conditions. By closely mimicking the physical and biochemical complexity of the natural extracellular matrix (ECM), PAs provide nourishing and protecting microenvironments for cells^{16, 18, 19}. For this study, we generated PA-RGDS by incorporating cell adhesive ligand Arg-Gly-Asp-Ser (RGDS) and matrix metalloprotease-2 (MMP-2) degradable sequence Gly-Thr-Ala-Gly-Leu-Ile-Gly-Gln (GTAGLIGQ) into the PA (Supplemental Figure 1)^{16, 18, 19}. RGDS, a fibronectin-derived cell adhesive ligand, is incorporated into PAs to promote cell adhesion and survival. The incorporation of MMP-2 degradable sequences allows cell-mediated degradation of the nanomatrix gel and migration of encapsulated cells into surrounding tissues.

In this study, we developed a fully defined, clinically-compatible cell culture system that can generate purified, functional, and therapeutically effective ECs. Next, we encapsulated hPSC-ECs within the nanomatrix gel and transplanted them into experimental hindlimb ischemia. These encapsulated hPSC-ECs remained engrafted for more than 10 months in ischemic tissues, and when compared to bare hPSC-ECs, they exerted higher and prolonged neovascularization and showed better vascular regenerative capacity. Particularly, this study for the first time demonstrated long-term in vivo behavior of hPSC-ECs in ischemic tissues and revealed sustained and dynamic incorporation of engrafted hPSC-ECs into the host vessels. This platform can serve as an effective modality for treating cardiovascular disease, and can be employed to investigate long-term behavior of the implanted cells.

METHODS

The use of hESCs was approved by the Emory University Human Embryonic Stem Cell Research Oversight Committee (ESCRO) and all animal protocols were approved by the Emory University Institutional Animal Care and Use Committee (IACUC).

Human pluripotent stem cell culture and differentiation

The hESCs (H1, H7, and H9) and hiPSCs (BJ1 and PGP1 obtained from Dr. George Daley) were cultured in mTeSR™ 1 (STEMCELL Technologies) on 5% matrigel at 37°C, 5% CO₂.²⁰. For directed differentiation, enzymatically dissociated clumps of hPSCs (lower than passage 60) were cultured on 0.01% collagen-coated plates in DMEM/F12 including 20% Serum Replacement (SR) with or without addition of specific differentiation factors for 7–10 days (Figure 1A).

Figure 1. Differentiation of hPSCs into ECs.

(A) Three stages of differentiation from hPSCs into ECs. Stage 1: Mesoderm induction. Stage 2: Endothelial differentiation. Stage 3: EC enrichment. (B) Flow cytometry analysis for KDR in three hPSC lines (H1, H9, or BJ1) cultured on collagen-coated plates with CHIR99021 treatment examined at indicated days. *P < 0.01, vs. Day 0, ^#P < 0.05, Day 3 vs. other days, two-way ANOVA followed by multiple comparisons with Tukey’s method. n = 5. (C) Flow cytometric analysis for EC-lineage markers in differentiating hESCs (H9) with or without DLL4 treatment determined at day 14. *P < 0.05, standard unpaired Student’s t-test. n = 4 to 5. (D) Double flow cytometric analysis showed enrichment of cells expressing KDR, TEK, and VWF in the CDH5⁺ cell fraction (shown is an example of H9). (E) mRNA expression of EC genes measured by qRT-PCR in endothelially differentiated hPSCs before and after sorting for CDH5 with MACS. Three independent experiments, each with technical triplicates. ^#P < 0.05, ^##P < 0.01, Unsorted vs. CDH5⁺. *P < 0.05, **P < 0.01, CDH5⁻ vs. CDH5⁺. One-way ANOVA followed by multiple comparisons with Tukey’s method. Representative examples from H9. (F) MACS-sorted hPSC-derived CDH5⁺ cells were subjected to immunocytochemistry after 24 hours. Concomitant expression of CDH5 and VWF was observed in hESC (H9)-derived CDH5⁺ cells and hiPSC (BJ1)-derived CDH5⁺ cells. (G) Detection of intracellular NO in post-sorted hESCs (H9)-CDH5⁺ cells, hiPSC (BJ1)-CDH5⁺ cells, and HUVECs measured by DAF-FM. (H) hPSC-derived CDH5⁺ cells formed tubular structures in Matrigel, took up DiI-Ac-LDL (red) and stained for FITC-UEA-1 lectin (green). (I) Confocal microscopic imaging of the sectioned Matrigel plug revealed that hPSC-CDH5⁺ cells expressed ILB4 and were incorporated into newly generated vessels within the Matrigel plug, indicating in vivo vasculogenic contribution of CDH5⁺ cells.

Quantitative RT-PCR

qRT-PCR assay was performed as described previously^{7, 21}. In brief, total RNA was isolated from cells using RNeasy (Qiagen, Venlo, Netherlands) according to the manufacturer’s instructions. Extracted RNA was reverse-transcribed using Taqman Reverse Transcription Reagents (Applied Biosystems, Foster, California) according to the manufacturer’s instructions. The synthesized cDNA was subjected to qRT-PCR using specific primers and probes (see Supplemental Table S1). Quantitative assessment of RNA levels was performed using an ABI PRISM 7500 Sequence Detection System (Applied Biosystems, Foster, California). Relative mRNA expression normalized to GAPDH expression was calculated as previously described²¹.

Magnetic activated cell sorting (MACS)

hPSCs were cultured on collagen coated plates for an additional 14 days. For sorting of CDH5⁺ with MACS, differentiated hPSCs were incubated with APC-conjugated mouse anti-human CDH5/CD144 (17-1449-42, eBioscience). After washing, the cell pellet was incubated with anti-APC beads (120-001-265, Miltenyi Biotec) and subjected to MACS sorting (Miltenyi Biotec).

Fabrication of the nanomatrix gel

Two PAs, C₁₆-GTAGLIGQRGDS (PA-RGDS) and C₁₆-GTAGLIGQS (PA-S), were synthesized via Fmoc-chemistry using an AAPPTec Apex 396 peptide synthesizer as previously described^{16, 18, 19}. The peptides were then alkylated at the N-termini via two 12 h reactions with palmitic acid in the presence of a mixture of o-benzotriazole-N,N,N’,N’-tetramethyluronium hexafluorophosphate (HBTU) and diisopropylethylamine (DiEA) dissolved in dimethylformamide (DMF). This was followed by cleavage from the resin and deprotection for 3 h, using a 40:1:1:1 cocktail of trifluoroacetic acid (TFA), deionized water, triisopropylsilane (TIPS), and anisole. The collected samples were subjected to rotary evaporation to remove excess TFA, precipitated in diethyl ether, and lyophilized. Successful synthesis of the PAs was confirmed via matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) mass spectrometry.

Stock solutions of the 2 PAs, PA-RGDS and PA-S [2% (weight/volume)] were individually prepared by dissolving lyophilized PA in distilled water and adjusting the pH to 7 using 1M sodium hydroxide (NaOH). The two PA solutions were then mixed in a 1:1 molar ratio, and self-assembly into three-dimensional hydrogels was induced by combining 50 µL of PA solution with a mixture containing 15 µL of 0.1 M CaCl₂ and 25 µL of cell suspension^{18, 19}

Intracellular nitric oxide detection

hPSC-derived ECs and HUVECs were treated with 4-amino-5-methylamino-2′,7′-difluorofluorescein (DAF-FM) diacetate (Invitrogen, Carlsbad, California) for 1h²². After the excess probe was removed, cells were incubated for an additional 20 minutes to allow for complete deesterification of the intracellular DAF-FM diacetate to the nonpermeable and nonfluorescent DAF-FM, which is converted to the highly fluorescent triazol form in the presence of NO. The fluorescence images were captured from at least 5 randomly selected cells per dish under fluorescence microscopy. The relative levels of intracellular NO were determined by quantifying the fluorescence intensity of DAF-FM using ImageJ software.

Transplantation of the cells into ischemic hindlimb

We performed hindlimb ischemia surgery and cell implantation as described previously^{21, 23}. Briefly, a ligation was made around the femoral artery and large branches were cauterized in 8 to 12 week old athymic male nude mice (Foxn1^nu) (Harlan, USA). Mice were randomly assigned to seven groups: Surgery only, medium, nanomatrix gel, HUVEC, hPSC-ECs, and HUVECs or hPSC-ECs encapsulated within the nanomatrix gel. 2 × 10⁵ cells were used in all cases. Immediately after the surgery, medium or bare cells were injected and encapsulated cells were implanted into hindlimb muscles. For histologic studies, cells were pre-labeled with Chloromethylbenzamido (CellTracker™ CM-DiI) for tracking the cells before cell implantation^{7, 21, 24}. According to IACUC guidelines, mice were kept in small cages (up to five animals in each cage) and fed sterile food and water. Pre-established exclusion criteria for surgery and follow-up analyses included systemic diseases, toxicity, respiratory distress, interference with eating and drinking and substantial weight loss (>15%). During the study period most of animals appeared to be in good health and were included in the appropriate analysis.

Blood flow measurement in ischemic hindlimb

Blood flow of the hindlimb was measured using a Laser Doppler perfusion imager (LDPI, Moor Instruments, United Kingdom) before and after surgery and every week for 4 weeks²¹. Mean values of perfusion were calculated from the stored digital color-coded images. The blood flow level of the ischemic (left) limb was normalized to the non-ischemic (right) limb to avoid data variations caused by ambient light and temperature.

Statistical Analysis

Statistical analyses and calculation of sample size were performed using Prism 5 software (GraphPad Software, La Jolla, CA, USA). All data demonstrated a normal distribution and similar variation between groups; hence only parametric tests were required. In each experiment, the sample size was determined using the expected difference and estimated standard deviation, with two-sided significance of 0.05 and power of 0.90. Statistical analyses were performed with the standard unpaired Student’s t-test for comparisons between two groups and an appropriate ANOVA followed by multiple comparisons with the Tukey method for comparison among more than 2 groups, as indicated in each figure legend. qRT-PCR and cell biological assays were performed at least in triplicate with three to six independent experiments. Flow cytometry analysis was performed with three to five independent samples. For the hindlimb ischemia studies, the estimated minimal number for having meaningful difference at 4 weeks is 5, so we included 5 to 12 animals in each group. Investigators were blinded to the assessment of the analyses of cell, animal, and histological experiments when comparison between the groups was made. All data were presented as mean ± standard error of the mean (s.e.m). Values of P < 0.05 were considered to denote statistical significance.

RESULTS

Generation of human pluripotent stem cell-derived ECs via a clinically compatible system

We developed a clinically compatible stepwise protocol which follows endothelial development (Figure 1A). To develop a fully defined system, KnockOut™ Serum Replacement substituted for animal serum and feeder cells. As a first step, we compared two coating materials, collagen and Matrigel™, and induced differentiation of hPSCs into the mesodermal lineage using CHIR99021, a GSK3β inhibitor which mimics Wnt activation²⁵. hESCs (H9) were plated onto dishes coated with 0.01% collagen or 10% Matrigel™ and were cultured for 3, 5, and 7 days in hESC medium with or without 3 µM CHIR99021. Real-time RT-PCR (qRT-PCR) showed that T (also known as Brachyury) and KDR transcripts were most highly expressed in conditions using collagen coating and CHIR99021 treatment for 3 days (Supplemental Figure 2A). The expression of definitive ectoderm (SOX1) and endoderm (FOXA2) markers were not significantly changed (Supplemental Figure 2A), and POU5F1 (OCT4) expression was reduced (Supplemental Figure 2B). Flow cytometry analyses confirmed that the percentage of KDR⁺ cells was highest (51.2 ± 4.3%) under these conditions (Supplemental Figure 2C). Another hESC line (H1) and a hiPSC line (BJ1) showed similar results (Figure 1B). These mesodermally differentiated hPSCs were cultured in medium containing VEGFA, FGF2, EGF, DLL4, and heparin for 5~9 days from day 4. qRT-PCR showed new expression or substantial increase of KDR, PECAM1, TEK (TIE2), CDH5, and VWF at day 14 compared to undifferentiated hPSCs and mesodermally-differentiated hPSCs at day 3. Addition of a Notch ligand, DLL4, further increased KDR, PECAM1, CDH5 and VWF expression (Supplemental Figure 3A). Flow cytometric analyses also confirmed increases in PECAM1, CDH5, and VWF (Figure 1C and Supplemental Figure 3B) and decreases in PTPRC (CD45), CD34, KIT (CD117), and PROM1 (CD133) (Supplemental Figure 4A–C). These data imply that DLL4 promotes EC differentiation and inhibits hematopoietic-lineage differentiation. Similar results were found in three hESC lines (H1, H7, and H9) and two hiPSC lines (BJ1 and PGP1) by flow cytometry and qRT-PCR (Supplemental Figure 5A and B). In double flow cytometry, 98.6% of CDH5⁺ cells expressed VWF, 79.0% TEK, and 66.3% KDR, suggesting substantial enrichment of the EC population in the CDH5⁺ cell fraction (Figure 1D). Immunocytochemistry confirmed expression of EC markers VWF and CDH5 as well (Supplemental Figure 6A). Nitric oxide (NO) production was demonstrated in these cells by a NO-sensitive fluorescent dye, 4-amino-5-methylamino-2′, 7′-difluorofluorescein (DAF-FM) diacetate (Supplemental Figure 6B). When plated on Matrigel™, the cells formed vascular-like structures (Supplemental Figure 6C). These results indicate that a combination of angiogenic factors together with EGF and DLL4 efficiently induced endothelial differentiation of hPSCs.

To enrich endothelial-lineage cells, we sorted the differentiated cells at day 14 for CDH5 with the magnetic-labeled cell separation system (MACS^®). The sorted CDH5⁺ cells exhibited 3- to 8-fold higher mRNA expression of KDR, NOS3 (eNOS), and VWF compared to CDH5⁻ cells (Figure 1E). Immunocytochemistry demonstrated that almost all CDH5⁺ cells expressed VWF and CDH5 (Figure 1F). The hPSC-CDH5⁺ cells showed typical cobblestone-like EC morphology and produced NO detected by DAF-FM diacetate (Figure 1G). Quantitatively, the DAF-FM intensities of the two lines of hPSC-CDH5⁺ cells were similar to that of HUVEC. These CDH5⁺ cells formed tubular structures in Matrigel and showed uptake of DiI-acetylated-LDL and binding to UEA-1 lectin (Figure 1H). In agreement with the in vitro results, we found co-localization of the injected cells with ILB4 positive vessels in vivo using a Matrigel plug assay (Figure 1I). These data suggest that the sorted hPSC-CDH5⁺ cells are highly enriched functional endothelial-lineage cells.

Pro-angiogenic properties of hPSC-derived ECs

To determine angiogenic activity of the sorted cells, we first measured expression of pro-angiogenic factors in pre- and post-sorted cells. qRT-PCR showed that mRNA expression of VEGFA, IGF1, ANGPT1, and FGF2 were significantly higher in the CDH5⁺ cells compared to the CDH5⁻ cells or unsorted cells (Figure 2A). Next, we conducted cell migration and tube-formation assays using conditioned media (CM) collected from cultured CDH5⁺ cells or HUVECs. In the cell migration assay with Boyden chamber, the number of migrated endothelial cells (ECs) was significantly higher in the CDH5⁺-cell group compared to the HUVEC group (Figure 2B). For the tube formation assay, endothelial cells were cultured on Matrigel-coated plates and the above two conditioned media were added. Tube length and branches assessed 8 hours later were significantly higher in the CDH5⁺-CM-treated group compared to HUVEC-CM-treated group (Figure 2C). These data indicate potent pro-angiogenic properties of hPSC-CDH5⁺ cells.

Figure 2. Pro-angiogenic properties of hPSC-CDH5⁺ cells.

(A) mRNA expression of representative angiogenic factors in undifferentiated hPSCs, endothelially differentiated hPSCs before sorting at day 14, hPSC-CDH5⁻ cells, and hPSC-CDH5⁺ cells measured by qRT-PCR. Data are presented as mRNA expression relative to GAPDH. Three independent experiments, each with technical triplicates. *P < 0.05, one-way ANOVA followed by multiple comparisons with Tukey’s method. (B, C) Sorted hPSC-CDH5⁺ cells and HUVECs were cultured in M199 media, and the conditioned media of CDH5⁺ cells (CM-CDH5⁺) and HUVECs (CM-HUVEC) were collected and used for migration (B) and tube formation (C) assays. M199 media containing 1% FBS was used as a negative control. Representative examples from H9 are shown. *P < 0.05, one-way ANOVA followed by multiple comparisons with Tukey’s method. n = 6.

Cytoprotection of hPSC-ECs under oxidative stress by the nanomatrix gel encapsulation

We then investigated whether the PA-RGDS nanomatrix gel supports the viability of hPSC-derived ECs in regular culture conditions. hPSC-derived ECs were encapsulated within the nanomatrix gel (Figure 3A) and cultured for 7 days under normal endothelial cell culture conditions. Live/Dead staining at day 7 demonstrated that 96.5 ± 1.3% encapsulated hESC-ECs were viable (Figure 3B). Next we determined whether hPSC-ECs can survive under oxidative stress, which is a major cause of low cell survival in ischemic tissues ¹⁴. We thus exposed hPSC-ECs encapsulated within the nanomatrix gel to a high concentration of H₂O₂ (100µM). Live/Dead staining demonstrated that viability was 8.2 times better when hPSC-ECs were encapsulated than not (Figure 3C and D). These results indicate that encapsulation of hPSC-ECs within the nanomatrix gel can promote EC survival under oxidative stress.

Figure 3. Cytoprotective effects of the nanomatrix gel encapsulation against H₂O₂.

(A) Encapsulation of DiI-labeled hPSC-derived ECs. (B) Representative Live/Dead assay images of nanomatrix gel-encapsulated hPSC-ECs 7 days after culture in vitro. Quantification graph summarizes the results of Live/Dead assay. **P < 0.001, vs. Live cells. One-sample t-test. n = 5. (C, D) Encapsulation of hPSC-ECs within the nanomatrix gel increased cell survival after H₂O₂ (100 µM) treatment as determined by the Live/Dead assay. Representative images (C) and quantitative analysis (D) of Live/Dead assay. **P < 0.001, vs. hPSC-EC, standard unpaired Student t-test. n = 5. Representative examples from H9 are shown.

Degradation kinetics of the nanomatrix gel in vivo

To examine the in vivo degradation of the nanomatrix gel, the gel was pre-labeled with CM-DiI, a red fluorescent dye (Supplemental Figure 7A). We surgically induced hindlimb ischemia (HLI) in nude mice and the nanomatrix gel was implanted into the ischemic hindlimb. The mice were euthanized at 1, 3, or 6 weeks after injections, and the hindlimb tissues were harvested. Through histological evaluation under confocal microscopy, we found that the DiI-labeled nanomatrix gel was gradually decreased in muscle tissues over time and was almost completely degraded at 6 weeks (Supplemental Figure 7B).

Therapeutic effects of hPSC-ECs encapsulated within the nanomatrix gel on hindlimb ischemia

To determine the therapeutic effects of hPSC-derived ECs encapsulated within the nanomatrix gel on ischemic disease, we used a HLI model. We surgically induced HLI in nude mice and transplanted hPSC (BJ1-hiPSC) ECs (2 × 10⁵) encapsulated with or without the nanomatrix gel, HUVECs (2 × 10⁵) with or without the nanomatrix gel, the nanomatrix gel only, or culture medium into hindlimb muscle. Laser Doppler perfusion image (LDPI) analysis demonstrated that the hPSC-EC and hPSC-EC /nanomatrix gel group showed higher blood flow at 2, 3, and 4 weeks compared to other groups such as no treatment (surgery only), medium, the nanomatrix gel only, and HUVECs with or without the nanomatrix gel (Figure 4A and 4B). Morerover, the hPSC-EC/nanomatrix gel group showed higher blood flow than the bare hPSC-EC-group at 4 weeks. We further conducted similar experiments with another cell line, H9 hESCs, and found similar results (data not shown). We next determined density of vessels and muscles in tissues harvested at 4 weeks which were systemically perfused with FITC-conjugated isolectin B4 (ILB4). The functional vascular density was again higher in the hPSC-EC group compared to the other groups when omitting the encapsulated hPSC-EC group. But when the encapsulated and bare hPSC-EC groups were compared, the vascular density was significantly higher in the encapsulated group (Figure 4C and 4D). Furthermore, the muscle density was significantly higher in the hPSC-EC/nanomatrix gel group compared to the other groups (Figure 4E and 4F). Taken together, these results indicate that encapsulation of hPSC-ECs within the nanomatrix gel is most effective for repairing limb ischemia and augments functional and structural neovascularization.

Figure 4. Therapeutic effects of hPSC-ECs encapsulated within the nanomatrix gel in hindlimb ischemia.

(A, B) LDPI images (A) and quantitative analysis (B) showed blood flow in the mice receiving surgery only (No treatment), Medium, nanomatrix gel, HUVECs, HUVEC/nanomatrix gel, hPSC-EC, or hPSC-EC/nanomatrix gel over four weeks after treatment. *P < 0.05, **P < 0.001, hPSC-EC vs. five control groups (No treatment, Medium, nanomatrix gel, HUVEC, HUVEC/nanomatrix gel). ^#P < 0.05, ^##P < 0.001, hPSC-EC/nanomatrix gel vs. five control groups. ⁺⁺P < 0.05, hPSC-EC/nanomatrix gel vs. hPSC-EC, two-way Repeated measures ANOVA followed by multiple comparisons with Tukey’s method. n = 5–15 per group. (C, D) Vascular density in hindlimb muscles. Representative images of FITC-ILB4-perfused vessels (C) and quantitative analysis of vascular density (D). *P < 0.05, hPSC-EC vs. five control groups (No treatment, Medium, nanomatrix gel, HUVEC, HUVEC/nanomatrix gel), one-way ANOVA followed by multiple comparisons with Tukey’s method. **P < 0.05, hPSC-EC/PA-RGDS vs. hPSC-EC, n = 9 per group. Representative examples from BJ1 are shown. (E, F) Muscle density in ischemic hindlimbs. Representative images of ACTN2-positive skeletal muscles (E) and quantitative analysis of muscle density (F). *P < 0.05, hPSC-EC/nanomatrix gel vs. other groups (No treatment, Medium, nanomatrix gel (PA-RGDS), HUVEC, HUVEC/nanomatrix gel, hPSC-EC), one-way ANOVA followed by multiple comparisons with the Tukey’s method. n = 5 per group. Representative examples from BJ1 are shown.

Prolonged engraftment of hPSC-ECs encapsulated within the nanomatrix gel in ischemic tissues

An important issue that was not addressed in prior studies is the long-term temporospatial behavior of the implanted hPSC-ECs in ischemic disease models. Thus, we monitored the fate of implanted cells by BLI over 21 weeks and histologically for 10 months (Figure 5). For BLI studies, hPSC-ECs were transduced with a reporter gene consisting Fluc-TdTomato (^{Fluc+TdTomato+}hPSC-CDH5⁺) before transplantation^{26, 27}. We confirmed that > 90% of the transduced hPSC-ECs expressed TdTomato by fluorescence microscopy and flow cytometry (Supplemental Figure 8A and B). We then optimized the exposure time to detect bioluminescence signals after encapsulating hPSC-ECs within the nanomatrix gel (Supplemental Figure 8C) and used these conditions for in vivo BLI experiments (Figure 5A). In mice receiving hPSC-ECs only, bioluminescence signals increased over 3 days, but dramatically reduced thereafter approaching an undetectable level at week 3. However, in mice receiving nanomatrix gel-encapsulated hPSC-ECs, the bioluminescence signals were gradually increased over one week, presumably because degradation of the nanomatrix gel over this period allowed quicker penetration of luciferin into the nanomatrix gel and hPSC-ECs. These BLI signals were well maintained between 2 and 14 weeks, but rapidly decreased to a minimally detectable level by 21 weeks. From one week and on, the BLI signals were consistently higher in the encapsulated group than in the non-encapsulated group (Figure 5A and B).

Figure 5. Survival of hPSC-ECs or hPSC-ECs encapsulated within the nanomatrix gel after implantation into ischemic hindlimbs.

(A) ^Fluc+Tomato+hPSC-ECs, alone or encapsulated within nanomatrix gel, were transplanted into ischemic hindlimbs and were tracked noninvasively by BLI over 21 weeks. Rainbow bar refers to the threshold to discriminate background noise from a positive BLI signal. (B) Quantitative analysis of BLI from ischemic hindlimbs implanted with bare or nanomatrix gel-encapsulated hPSC-ECs. n = 3 per group. *P < 0.05, Repeated measures ANOVA followed by multiple comparisons with Tukey’s method. (C) Confocal microscopic images of muscle sections implanted with bare or nanomatrix gel-encapsulated hPSC-ECs taken at indicated time points over 10 months. Representative examples from BJ1 are shown. (D) Quantification of engrafted cell density. n = 6. *P < 0.05. Two-way ANOVA followed by multiple comparisons with Tukey’s method.

For histological analysis, hPSC-ECs were pre-labeled with a red-fluorescent dye, DiI, and both nanomatrix gel-encapsulated and bare hPSC-ECs were transplanted into the hindlimb following HLI surgery. We harvested muscle tissues at 2 weeks, 4 weeks, 2 months, and 10 months and performed immunohistochemistry (Figure 5C). Confocal microscopic examination of the tissues demonstrated that the number of engrafted cells were gradually decreased over 10 months in both groups, but at all measured points it was significantly higher in the nanomatrix gel-encapsulated group compared to the bare hPSC-EC group (Figure 5C and D). Taken together, encapsulation of hPSC-derived ECs within the nanomatrix gel substantially increased engraftment and long-term survival of transplanted hPSC-ECs in ischemic hindlimbs.

In vivo neovascularization of nanomatrix gel-encapsulated hPSC-ECs

To investigate angiogenic effects of hPSC-EC transplantation, we performed qRT-PCR with muscles harvested at 2 and 4 weeks. We compared expression of six representative angiogenic genes (Vegfa, Angpt1, Fgf2, Igf1, Ccl2 (Mcp-1), and Pdgfb) between mice receiving the nanomatrix gel-encapsulated hPSC-ECs, hPSC-ECs, nanomatrix gel-encapsulated HUVECs, HUVECs, or the nanomatrix gel only. First, we compared mRNA expression between the hPSC-EC group and control groups (HUVEC, nanomatrix gel-encapsulated HUVECs, or nanomatrix gel) (Figure 6A). The hPSC-EC group showed higher expression of Vegfa and Angpt1 compared to the controls at 2 weeks and Mcp-1 at 4 weeks. Next, we compared mRNA expression between the nanomatrix gel- encapsulated hPSC-EC group and the bare hPSC-EC group. Both at 2 and 4 weeks, expression of all the genes except Angpt1 (2 weeks) and Ccl2 (4 weeks) was significantly higher in the encapsulated hPSC-EC group compared to bare hPSC-ECs. These data indicate that while bare hPSC-EC injection increased limited angiogenic and arteriogenic factors for 2 to 4 weeks, the encapsulated hPSC-ECs increased most representative angiogenic, arteriogenic, and cytoprotective factors more robustly and for longer than 4 weeks.

Figure 6. Contribution of hPSC-ECs encapsulated within the nanomatrix gel to neovascularization.

(A) mRNA expression of angiogenic and arteriogenic factors in hindlimb tissues implanted with nanomatrix gel only, HUVECs only, encapsulated HUVECs within the nanomatrix gel, hPSC (BJ1)-ECs only, or encapsulated hPSC (BJ1)-ECs within the nanomatrix gel measured by qRT-PCR at 2 and 4 weeks. Three independent experiments, each with technical triplicates. *P < 0.05, hPSC-EC or hPSC-EC/nanomatrix gel vs. nanomatrix gel, HUVECs, and HUVECs/nanomatrix gel. ^#P < 0.05 or **P < 0.01, hPSC-EC/nanomatrix gel vs. hPSC-EC. One-way ANOVA followed by multiple comparisons with Tukey’s method. (B) Representative confocal microscopic images at 1, 4 weeks, 2 months and 10 months after implantation of bare or nanomatrix gel-encapsulated DiI-labeled hPSC-ECs. White arrows indicate hPSC-ECs that are incorporated into ILB4⁺ (green) vessels, and yellow arrows, perivascular localized hPSC-ECs. Blue: DAPI. (C) Representative microscopic images of perivascularly localized hPSC-ECs taken of muscle sections harvested at 4 weeks, 2 months, and 10 months. (D) Quantification graphs of perivascularly localized hPSC-ECs per engrafted cells over time. *P < 0.01. hPSC-EC/nanomatrix gel vs. hPSC-EC, ^#P < 0.001, vs. 1w, ^##P < 0.01, 10m vs. 2m, two-way ANOVA followed by multiple comparisons with Tukey’s method. (E) Reorganization of engrafted hPSC-ECs and vascular guiding effects over 10 months (yellow dotted line). (F) Orthogonal confocal images showing six different hPSC-ECs incorporated into vessels (white arrows). Samples from 4 weeks after transplantation of nanomatrix gel-encapsulated hPSC-ECs (BJ1). (G) Orthogonal confocal images showing twelve different hPSC-ECs incorporated into vessels. Samples from 2 months after transplantation of nanomatrix gel-encapsulated hPSC-EC (BJ1). (H) Confocal images showing hPSC-ECs incorporated into vessels (white arrows) and putative guiding roles of hPSC-ECs (cyan arrowhead and yellow dotted line). Samples from 10 months after transplantation of nanomatrix gel-encapsulated hPSC-ECs (BJ1). (I, J) Quantitative analysis showing the percentages of hPSC-ECs-incorporated vessels (DiI⁺ILB4⁺) per total vessels (ILB4⁺) (I) and intravascularly incorporated hPSC-ECs (DiI⁺ILB4⁺) per total engrafted hPSC-ECs (DiI⁺) (J). n = 6 to 8. *P < 0.05, or **P < 0.001. hPSC-EC/nanomatrix gel vs. hPSC-EC, ^#P < 0.05, 10m vs. other time points, two-way ANOVA followed by multiple comparisons with Tukey’s test.

Next, we examined vasculogenic effects or direct contribution of hPSC-ECs to vessel formation over 10 months via histological analyses. For these experiments, FITC-ILB4 was systemically injected before sacrifice to identify functional endothelium. Confocal microscopic examination of tissues demonstrated that engrafted hPSC-ECs were more frequently observed near the vasculature over time (Figure 6B–D). Quantitatively, the proportion of perivascularly localized hPSC-ECs per total engrafted hPSC-ECs were gradually increased over time in both encapsulated and bare hPSC-EC injected groups; however, the absolute numbers were significantly higher in the encapsulated hPSC-EC group (Figure 6C and D). Moreover, from 4 weeks to 10 months after implantation, some of the engrafted hPSC-ECs in both bare and encapsulated groups were linearly aligned, mimicking vascular structures, suggesting their guiding role for vascular growth during remodeling (yellow dotted lines in Figure 6E, H, and Supplemental Figure 9A).

Finally, we examined temporal changes in incorporation of hPSC-ECs into the vessels. In the unencapsulated hPSC-EC group, very few hPSC-ECs were incorporated into vessels at one week. However, the proportion of hPSC-EC-incorporated vessels (DiI⁺ILB4⁺) among total functional vessels (ILB4⁺) increased over time: less than 1% at 4 weeks and 9.4% at 10 months (Figure 6B upper panel and Figure 6I). In the nanomatrix gel-encapsulated hPSC-EC group, the proportion was similarly low at 1 week; however, it progressively increased between 1 and 10 months and at each time point (1, 2, 10 months) it was significantly higher compared to the unencapsulated group (Figure 6B lower panel, Figure 6F, G, H and I). At 10 months, the proportion of hPSC-EC incorporated vessels reached 45.9% (white arrows in Figure 6H, I, and Supplemental Video I), and 78.6% of engrafted or remaining hPSC-ECs were completely incorporated into vessels (white arrows in Figure 6H and J). Notably, vessels incorporated with multiple hPSC-ECs were observed, which was rarely seen at other time points. Moreover, some of unincorporated hPSC-ECs were linearly or curvilinearly aligned like vessels parallel to or continuous with host vessels (yellow dotted lines in Figure 6H and Supplemental Figure 9) and a part of them were faintly stained positive for ILB4 (cyan arrowheads in the yellow dotted lines in panel 9 in Figure 6H and Supplemental Figure 9B), indicating that hPSC-ECs play a guiding role for new vessel formation. Together, these data suggest that engrafted hPSC-ECs were initially moving toward vessels and gradually incorporated into vessels, and the nanomatrix gel, by enabling long-term cell survival, allowed sustained vasculogenic effects of hPSC-ECs.

We stained tissues with endodermal marker AFP and ectodermal markers SOX1 and Nestin and did not detect cells differentiated into other lineages (data not shown). Gross and microscopic examination showed no discernible tumors after injection with hPSC-ECs in 88 animals. Together, these data suggest that implantation of hPSC-ECs encapsulated with PA-RGDS recovered blood flow through enhancement of cell retention and neovascularization within ischemic hindlimb.

DISCUSSION

This study addressed two major roadblocks for cardiovascular cell therapy: 1) generation of clinically compatible, high-purity, therapeutically effective, and safe hPSC-derived ECs, and 2) enabling long-term survival of hPSC-ECs in ischemic tissue via encapsulation within the nanomatrix gel, thus enhancing their therapeutic effects. Furthermore, this study uncovered long-term in vivo behavior of hPSC-ECs in ischemic tissues: sustained and dynamic incorporation of engrafted hPSC-ECs into the host vessels and a guiding role of hPSC-ECs for new vessel formation.

Employing two molecules, CHIR99021 and DLL4, and sorting with CDH5, we established a highly efficient EC differentiation system. Of note, the purified hPSC-CDH5⁺ cells produced nitric oxide, a critical marker for functional ECs, which was not demonstrated in previous studies. Our system and the resultant hPSC-ECs have unique characteristics that favor clinical translation. First, this protocol employed fully defined conditions and showed no cell line variability among the lines we tested. Because a main advantage of hiPSC technology is its potential for autologous therapy, a protocol that can be universally applicable is required. Second, hPSC-derived ECs generated by our protocol demonstrated proangiogenic potential and direct vessel-forming effects. These dual characteristics have respective benefits, as each contributed to therapeutic neovascularization at different time points. Third, the implanted hPSC-ECs, even encapsulated, did not induce tumorigenic or other side effects during long-term follow-up.

The PA nanomatrix gel dramatically increased survival of hPSC-ECs and induced robust and longstanding vascular regenerative effects. The unique structural and functional characteristics of the nanomatrix gel provide important insight into how encapsulated hPSC-ECs exert therapeutic effects for tissue ischemia. Our results suggest that, initially the ECM-mimicking structure incorporating adhesive ligands allows easy adhesion of the nanomatrix gel with host ECM, stabilization of encapsulated hPSC-ECs, and transport of nutrients and growth factors by diffusion, thereby promoting the viability of the encapsulated hPSC-ECs and exerting proangiogenic effects on the host cells during the critical early phase of cell survival. Degradation of the nanomatrix gel then exposes and allows migration of encapsulated hPSC-ECs into ischemic areas and structural contribution of hPSC-ECs to vessel formation. This vasculogenic effect gradually becomes the main role for vascularization.

Long-term retention of hPSC-ECs via the nanomatrix gel allowed us to examine previously unreported temporal reorganization of engrafted cells and their relation to new vessel formation. First, we discovered that it takes several weeks for a majority of implanted cells to migrate toward vessels and to directly incorporate into vessels. As reported in prior studies^{5, 28} and ours, when hPSC-ECs alone were implanted, most cells died within several weeks, so that most effects are paracrine. But when hPSC-ECs were delivered within the nanomatrix gel, many cells were protected and thus had an opportunity to migrate toward vascular areas later. Second, the proportion of engrafted hPSC-ECs incorporated into the vessels increased steadily over 10 months. Initially, hPSC-ECs were more localized in the perivascular areas; however, during host vessel reorganization and new vessel formation, more hPSC-ECs were incorporated into the vessels. At 2 months, there were more engrafted cells and more perivascularly localized cells in the nanomatrix gel + cells group compared to the bare cell group, mainly due to higher survival of implanted cells. Between 2 to 10 months, these surplus perivascularly localized hPSC-ECs appeared to be more incorporated into the vessels during the vessel growth and remodeling as shown in Figures 6I and J. Thus at 10 months, the number of perivascularly localized cells in the nanomatrix gel group was substantially decreased to the level of bare cell group. Why the number of the perivascularly localized cells are similar remains to be determined. We can speculate that a certain balance is needed between the number of perivascularly localized cells per vessels or these similar numbers might be just co-incidental. To address this more appropriately, even a longer follow-up is needed. Third, long-term engraftment is a critical factor for ongoing vasculogenesis by hPSC-ECs. Third, hPSC-ECs showed a guiding role for vessel growth, which enables multiple cellular incorporation during vessel formation. This is an unprecedented and the most notable finding in our study. Even at 10 months, we observed a guiding role and sustained and robust vessel growth, attributed to surviving hPSC-ECs. These data offer the very optimistic prospect that hPSC-ECs can be useful for clinical therapy for ischemic vascular disease, since most human cardiovascular diseases are chronic and require sustained vessel formation for effective treatment.

In summary, we for the first time developed a cell therapy with nanomatrix gel-encapsulated hPSC-ECs and demonstrated its therapeutic effects and safety. Since both the cell generation protocol and the nanomatrices are biocompatible, this platform can represent a highly effective therapeutic option for ischemic vascular disease. Furthermore, these hPSC-ECs and this bioengineered platform can help investigate the long-term fate of the implanted cells, disease investigation, and drug discovery.

Clinical Perspective.

What is new?

• *We developed a novel, fully-defined cell culture system to generate highly enriched ECs from hPSCs, which is compatible with clinical application.
• *The hPSC-derived ECs showed genuine EC characteristics and pro-angiogenic activities and exerted favorable therapeutic effects for repairing limb ischemia.
• *Encapsulation of hPSC-derived ECs in the biocompatible PA-RGDS nanomatrix gel improved long-term survival of hPSC-ECs in an ischemic environment, showing the longest survival of hPSC-ECs in ischemic tissue (10 months), and improved vessel-forming properties.

What are the clinical implications?

• *The hPSC-ECs will be useful for treatment of ischemic cardiovascular disease, drug discovery, and pathophysiological investigation of vascular diseases.
• *The PA-RGDS-mediated transplantation of hPSC-ECs can serve as a novel platform for cell-based therapy for cardiovascular disease and investigation of long-term behavior of hPSC-ECs.