Humanized large-scale expanded endothelial colony-forming cells function in vitro and in vivo

Abstract

Endothelial progenitor cells are critically involved in essential biologic processes, such as vascular homeostasis, regeneration, and tumor angiogenesis. Endothelial colony-forming cells (ECFCs) are endothelial progenitor cells with robust proliferative potential. Their profound vessel-forming capacity makes them a promising tool for innovative experimental, diagnostic, and therapeutic strategies. Efficient and safe methods for their isolation and expansion are presently lacking. Based on the previously established efficacy of animal serum-free large-scale clinical-grade propagation of mesenchymal stromal cells, we hypothesized that endothelial lineage cells may also be propagated efficiently following a comparable strategy. Here we demonstrate that human ECFCs can be recovered directly from unmanipulated whole blood. A novel large-scale animal protein-free humanized expansion strategy preserves the progenitor hierarchy with sustained proliferation potential of more than 30 population doublings. By applying large-scale propagated ECFCs in various test systems, we observed vascular networks in vitro and perfused vessels in vivo. After large-scale expansion and cryopreservation phenotype, function, proliferation, and genomic stability were maintained. For the first time, proliferative, functional, and storable ECFCs propagated under humanized conditions can be explored in terms of their therapeutic applicability and risk profile.

Figure 1

ECFC recovery from human steady-state PB in an animal serum-free humanized system. (A,B) Peripheral blood (PB) from healthy volunteers was density gradient-separated (+) to enrich for mononuclear cells () or immediately diluted (□) and seeded in EGM/10% pHPL in 75-cm² cell culture flasks. Culture surfaces were coated () with collagen only when indicated (+). PB from patients with stable cardiovascular disease was also immediately diluted and seeded in EGM/10% pHPL (▨). (A) The initial appearance of visible colonies was determined by daily culture observation. (B) Colony number was counted at the end of the primary 7- to 19-day culture period. Results are shown as mean plus or minus SEM of 6 independent experiments. * indicates statistically significant difference, P < .05. (C-H) Representative early colonies (day 8) and parts of large expanded colonies (day 13) from healthy volunteers are depicted with 40× initial magnification corresponding to different recovery strategies as indicated. (A composite picture of 1 representative large ECFC colony is shown in Figure S1D.) Images were captured with a DS-Fi1 camera on a Nikon (Lijnden, Netherlands) Diaphot 300 inverted microscope (original magnification 4×/0.13 NA objective) with the NIS-Elements D3.0 image acquisition software (Nikon). (J) Population doublings () and expanded cell number () determined after large-scale expansion of ECFCs from 6 healthy volunteers (healthy controls) and 3 CVD patients are shown. (K) Cumulative population doublings (mean ± SD) as obtained during large-scale expansion of ECFCs from 6 healthy volunteers after large-scale expansion are shown. Large-scale expansion-derived cells bear a history of mean 21 population doublings before initiating long-term culture at cell seeding densities of 10 (◆), 100 (■), 1000 (▲), and 10 000 cells/cm² (x). Cells were reseeded during long-term culture at indicated time points according to their initial seeding density.

Figure 2

Phenotypic characterization of large-scale expanded ECFCs. (A) Representative flow cytometry histograms of ECFCs after humanized large-scale expansion showing reactivity with EC-expressed marker molecules (right-shifted filled gray curves compared with black lined open curves of the appropriate isotype controls) and lack of reactivity with hematopoietic (CD14 and CD45) and stem cell-associated (CD90 and CD133) or activation (human leukocyte antigen class II type DR, HLA-DR) markers. (B-G) Representative cellular localization of (B) mesodermal cytoskeleton component vimentin, (C) CD146/melanoma and endothelial cell adhesion molecule detected by prototype antibody P1H12, (D) CD144/vascular endothelial-cadherin, (E) endothelial von Willebrand factor (VWF), (F) hematopoietic stem cell and vascular EC-specific CD31/platelet-endothelial cell adhesion molecule-1, compared with (G) one representative isotype control example. ECFCs were cryopreserved after large-scale expansion and thawed before seeding in chamber slides to obtain adherent cells for cytochemistry. Staining was done as specified in supplemental data. The brown color results from precipitation of the chromogen diaminobenzidine mediated by antibody binding to the target molecule. Images were captured with an Olympus (Hamburg, Germany) DP71 camera on an Olympus BX51 microscope (original magnification 100×/1.35 NA oil objective) with the Olympus cell D image acquisition software.

Figure 3

ECFC clonogenicity and genomic stability. ECFCs after large-scale expansion representing more than or equal to 20 population doublings of the colony-initiating cells were subjected to an endothelial colony assay in triplicate in a seeding density of 10 ECFCs/cm² in EGM/10% pHPL. (A) Colony assays were performed with ECFCs from 3 different donors each for 10 () or 14 days (). Colony plates were then fixed and stained before photo documentation. Precise cell numbers of all imaged colonies were counted in ImageJ software. (B) Examples of typical LPP and HPP colonies are depicted (Figure S2). Representative chromosome G-banding derived from ECFC nuclei after large-scale expansion of (C) female and (D) male ECFCs and corresponding sorted (E) female and (F) male karyograms are shown. Representative array CGH depiction of constitutional initial white blood cell–derived DNA compared with ECFC-derived DNA post large-scale expansion and after passage 4 of the same (G) female and (H) male volunteers as shown in panels C and E and D and F, respectively (further examples in Figure S4).

Figure 4

Telomere length and telomerase activity. Human ECFCs derived from PB (PB ECFCs; n = 4) and umbilical cord blood (UCB ECFC; n = 4) after different culture passages (p1-p3) were compared with each other and with mouse 3T3 fibroblasts as a positive control. (A-D) Flow cytometry fluorescence in situ hybridization was used to determine telomere length and (E) TRAP to measure telomerase activity. (A,B) Gating strategy of one representative example to select single cells in cell cycle phase G_0/1 based on size parameters (FSC-H indicates forward light scatter height; SSC-H, sideward light scatter height; region R1 = 88.6%) and DNA content determined after propidium iodine (PI) staining of hybridized cells (PI fluorescence width vs area, FL2-W, FL2-A; region R2 = 77.7% of region R1). (C) Corresponding histogram showing fluorescein isothiocyanate-tagged peptide nucleic acid (PNA-FITC) binding to telomeres (gray histogram) compared with background fluorescence after mock hybridization (open histogram). (D) Differences between PNA probe specific signal and background fluorescence height (ΔFL1-H; mean ± SD, n = 4 per passage) based on analyses of at least 10 000 single cells in G_0/1 phase per sample. (E) Telomerase activity displayed as cycle of threshold (Ct) in a real-time polymerase chain reaction-based TRAP assay. Statistically significant differences: *P < .05, **P < .01.

Figure 5

ECFCs function in vitro and in vivo after large-scale humanized expansion. (A-C) Representative transformed images of vascular network formation from 3 typical independent oligoclonal cultures and (D) 1 monoclonal ECFC culture compared with (E,F) 2 independent oligoclonal UCB-derived EPC networks under the same conditions on Matrigel. Nontransformed original phase-contrast microphotographs are documented in higher magnification (Figure S5). Images were captured with an Olympus Color View III camera on an Olympus IX51 microscope (original magnification 20×/0.4 NA objective) with the Olympus analySIS B acquisition software. (G) Serial image reconstruction of one representative complete vascular network created in a 0.4-cm² well of a 16-well glass chamber slide is shown (Videos S2,3). For in vivo neovasculogenesis, ECFCs from 2 donors were mixed with MSCs in Matrigel before injecting 0.2 mL of the composite subcutaneously in 4 nude mice per group (n = 24; 4 mice per ECFC source analyzed at 3 time points; a macroscopic view is shown in Figure S6D,E). (H) Topography of the histology is symbolized and (J) shown as a low magnification overview of a vimentin-labeled vascularized plug part. (K) Vimentin reactivity in the border area showing murine tissue (left half) in the direct vicinity of the human cell-containing area of the Matrigel plug as indicated in panel J. (M) Antihuman CD31, (N) antihuman VWF, (O,P) antimouse glycophorin A reactivity detected with antibody mTer119 within (O) human and (P) mouse vessels and (Q) representative isotype control reactivity of mouse red blood cell containing vasculature inside the plug.

Figure 6

Deranged vessel formation of pure ECFCs in vivo and vascular lineage commitment in vitro. (A-C) Representative pure ECFCs (10⁷) derived from large-scale single-step humanized expansion were resuspended in 0.2 mL phosphate-buffered saline and subcutaneously injected into nude mice. After one week, subcutaneous cell deposits (a macroscopic view is shown in Figure S6F) were processed for human CD31 histochemistry. (A) Overview appearance and (B,C) higher magnified regional view as indicated by schematic boxes with scale bars documenting magnification. ECFCs after large-scale expansion were also seeded into methylcellulose for hematopoietic colony-forming cell (CFC) assays. (D,G) Complete (10 cm²) assay plate overview, (E,H), 10× original magnification view, and (F,J) high magnification view that documents (D-F) complete lack of CFCs derived from ECFCs. In comparison, (G,H) regular red blood cell colony formation admixed with a hierarchy of differentially maturating white blood cell colonies and (J) less than 5% colony-forming units of granulocytes, erythrocytes, monocytes, and macrophages (CFU-GEMM) was derived from CD34⁺ hematopoietic progenitors.