Hypoxic Conditions Promote the Angiogenic Potential of Human Induced Pluripotent Stem Cell-Derived Extracellular Vesicles

Abstract

Stem cells secrete paracrine factors including extracellular vesicles (EVs) which can mediate cellular communication and support the regeneration of injured tissues. Reduced oxygen (hypoxia) as a key regulator in development and regeneration may influence cellular communication via EVs. We asked whether hypoxic conditioning during human induced pluripotent stem cell (iPSC) culture effects their EV quantity, quality or EV-based angiogenic potential. We produced iPSC-EVs from large-scale culture-conditioned media at 1%, 5% and 18% air oxygen using tangential flow filtration (TFF), with or without subsequent concentration by ultracentrifugation (TUCF). EVs were quantified by tunable resistive pulse sensing (TRPS), characterized according to MISEV2018 guidelines, and analyzed for angiogenic potential. We observed superior EV recovery by TFF compared to TUCF. We confirmed hypoxia efficacy by HIF-1α stabilization and pimonidazole hypoxyprobe. EV quantity did not differ significantly at different oxygen conditions. Significantly elevated angiogenic potential was observed for iPSC-EVs derived from 1% oxygen culture by TFF or TUCF as compared to EVs obtained at higher oxygen or the corresponding EV-depleted soluble factor fractions. Data thus demonstrate that cell-culture oxygen conditions and mode of EV preparation affect iPSC-EV function. We conclude that selecting appropriate protocols will further improve production of particularly potent iPSC-EV-based therapeutics.

Keywords: angiogenesis; extracellular vesicles (EV); hypoxia; hypoxia-inducible transcription factor (HIF); induced pluripotent stem cells (iPSC); regenerative medicine.

Figure A1

Preliminary results of 3D large scale culture for EV production. (A) Particle counts and (B) recovery in a pilot experiment using 3D human iPSC cultures in the DASbox Mini Bioreactor System (DASGIP-Eppendorf) as measured by tunable resistive pulse sensing (TRPS). Serum-free defined E8 medium showed particle counts at or below the limit of detection (LOD; dotted line) of the measurement system. This enabled accurate analysis of iPSC-derived particles, presumably representing extracellular vesicles, in iPSC-conditioned medium (CM) before and after enrichment by tangential flow filtration (TFF) with particle recovery comparable to the main study (technical triplicate analysis).

Figure A2

Angiogenic potential of iPSC-EVs. (A) Representative endothelial networks in the presence of iPSC EV obtained from iPSC cultures at different oxygen level as indicated compared to negative (−) and positive (+) control, respectively. (B) Endothelial networks in the presence of iPSC-EV derived from hypoxic 1% oxygen culture and isolated by TFF or TUCF and compared to addition of equivalent volume of soluble factors (SolF) corresponding to the TFF preparation. Representative images from three independent experiments showing network detection by imageJ plugin Angiogenesis Analyser.

Figure A3

No inhibition of T cell proliferation by iPSC EV generated from human iPSCs at. Mitogenesis of T cells was induced by adding phytohemagglutinin (PHA) and compared to the proliferation of untreated (UT) T cells. Human iPSC-EVs that were generated in iPSC cultures at different oxygen levels and analyzed in triplicates per experiment (iPSC-EVs 1%, n = 2; iPSC-EVs 5%, n = 4; iPSC-EVs 18%, n = 2) did not inhibit the PHA-induced T cell mitogenesis.

Figure 1

Large-scale production of induced pluripotent stem cell-derived extracellular vesicles (iPSC-EVs). (A) Schematic representation of large-scale EV production from iPSCs. After culture in six-well plates, iPSCs were transferred directly to T225 flasks and then to four-layered cell factories. Culture-conditioned medium (CM) was pooled and separated into soluble factors (SolF) and enriched EVs by tangential flow filtration (TFF). For further concentration and protein depletion, TFF-enriched EVs were submitted to ultracentrifugation (TUCF). Comparative EV analysis included particle, protein and EV identity determination as illustrated plus single EV phenotyping and function tests. (B) Identity and purity of human iPSCs determined by Tra-1-81, SSEA-4, Nanog, OCT-4 flow cytometry (representative data shown). (C) Particle concentration during EV production measured by tunable resistive pulse sensing (TRPS) in background (PBS + 0.05% Tween 20, 0.22 µm pore filtered), basic and supplemented defined serum-free medium (mTeSR™1) before and after culturing cells. Mean ± SEM of three independent experiments; individual results for CM are shown in Figure 2A. Dotted line represents limit of detection. Mean ± SD of three measurements (* p < 0.05).

Figure 2

Monitoring iPSC-EV purification efficacy. (A) Particle concentration (red bars) as measured by tunable resistive pulse sensing (TRPS) and protein concentration (Bradford; blue bars) of conditioned media (CM), EVs after tangential flow filtration (TFF), cell-secreted soluble factors (SolF) separated in the TFF permeate, EVs ultra-centrifuged after TFF (TUCF). Mean ± SD of three independent experiments (* p < 0.05; compared with CM and (+ p < 0.05; compared with TFF). (B) The particle/µg protein ratio (orange bars) was calculated by dividing the total particle count by total protein amount. Mean ± SD of three independent experiments. Efficiency of EV production (particle recovery; green bars) comparing the number of particles in the starting material (CM = 100%) to the particle number after TFF, in the SolF and after TUCF. (C,D) Illustration of particle size after (C) TFF and (D) TUCF purification. Lines represent mean data from three independent experiments ± SD. Dotted line indicates particles size peak mode of 109 nm in (C) and 103 nm in (D).

Figure 3

Characterization iPSC-EVs. (A) Human iPSC EVs in conditioned medium (CM), after enrichment by tangential flow filtration (TFF), and sub-sequent ultracentrifugation (TUCF) were compared to their parental iPSCs by immunoblotting for tetraspanins CD81, CD9 and CD63, for the EV marker alix and the cell-compartment contamination marker calnexin. Representative Western blots of two independent experiments. (B) Quantification of band intensity normalized to total protein loaded. Bars represent results from three independent replicates (**** p < 0.0001, *** p < 0.001, ** p < 0.01, * p < 0.05). (C) Representative super-resolution microscopy images of TFF- vs. TUCF-enriched iPSC-EVs showing co-localization of VEGF (green), and CD63 tetraspanins (red) as indicated. Scale bar 100 nm. Significantly more VEGF signal co-localization was determined on CD63 tetraspanin-labeled iPSC-EVs isolated by TFF than after TUCF as indicated.

Figure 4

Conditioning of iPSCs at different oxygen levels during EV production. (A) Representative images of iPSC cultures at different oxygen levels as indicated. Scale bar 1000 µm. (B) Formation of oxygen adducts measured by flow cytometry with the hypoxyprobe pimonidazole. Color code as indicated; black line shows unstained sample. (C) Particle concentration and particles secreted per cell per 24 h, by human iPSCs at indicated oxygen levels. Bars represent replicates from three independent experiments. (D) Representative immunoblotting for HIF-1α levels in human iPSCs at different oxygen levels. CoCl₂ was used as a positive control for HIF-1α stabilization. Protein quantification (relative band intensity) normalized to total protein loaded (right graph in (D)). Bars represent mean ± SD results from three independent replicates (**** p < 0.0001).

Figure 5

Angiogenic potential of iPSC-EVs obtained after iPSC conditioning at different oxygen levels. (A) Total length of the endothelial networks in the absence (white bar, positive control; grey bar negative control) or presence of iPSC EVs and corresponding soluble factors (SolF) obtained at defined oxygen levels as indicated and separated by tangential flow filtration (TFF), respectively (**** p < 0.0001, *** p < 0.001, ** p < 0.01, * p < 0.05). (B) Angio-profiler Western blot array of TFF-enriched iPSCs (duplicate analysis on one randomly selected EV preparation) normalized by equal EV number input. (C) Total length of endothelial networks in the absence or presence of iPSC-EVs derived from hypoxic iPSC culture and produced by TFF or TFF plus ultracentrifugation (TUCF) or by adding EV-depleted soluble factors (SolF) obtained as TFF retentate from 1% oxygen culture conditions. Bars represent pooled results from three independent experiments (**** p < 0.0001, ** p < 0.01). (D) Electropherogram analysis of different RNA samples isolated from TFF- (top) and TUCF-purified EVs (bottom) derived from different oxygen conditions as indicated and run on a Bioanalyzer RNA pico chip (mean ± SD; n = 3). Grey curves represent the small RNA profile of the same parental iPSCs in both histograms for comparison.