Fully biologic endothelialized-tissue-engineered vascular conduits provide antithrombotic function and graft patency

Abstract

Tissue-engineered vascular conduits (TEVCs), often made by seeding autologous bone marrow cells onto biodegradable polymeric scaffolds, hold promise toward treating single-ventricle congenital heart defects (SVCHDs). However, the clinical adoption of TEVCs has been hindered by a high incidence of graft stenosis in prior TEVC clinical trials. Herein, we developed endothelialized TEVCs by coating the luminal surface of decellularized human umbilical arteries with human induced pluripotent stem cell (hiPSC)-derived endothelial cells (ECs), followed by shear stress training, in flow bioreactors. These TEVCs provided immediate antithrombotic function and expedited host EC recruitment after implantation as interposition inferior vena cava grafts in nude rats. Graft patency was maintained with no thrombus formation, followed by complete replacement of host ECs. Our study lays the foundation for future production of fully biologic TEVCs composed of hiPSC-derived ECs as an innovative therapy for SVCHDs.

Keywords: endothelial cell; flow bioreactor; human induced pluripotent stem cell; shear stress training; single ventricle congenital heart defect; tissue-engineered vascular conduit.

Figure 1.. Development of tissue engineered vascular conduits (TEVCs) using human induced pluripotent stem cells (hiPSCs) for implantation as inferior vena cava interposition grafts in nude rats.

(A) Schematic illustration of lumen endothelialization using perfusion bioreactors and decellularized human umbilical cord arteries (dHUAs; ~2.0 mm luminal diameter) with hiPSC-derived endothelial cells (hiPSC-ECs). HUAs were chosen because they are readily available and the luminal diameters match well with that of the rat inferior vena cava (IVC) for implantation. HUAs (~2.5 cm in length) were decellularized, and the luminal surface were coated with human fibronectin at 100 μg/mL overnight at 37°C, followed by two consecutive 4-hour seeding incubations of hiPSC-ECs (~2.0×10⁶ cells/cm²). A three-way connector with an injection port was used to introduce the fibronectin and both rounds of cell seeding to maintain sterility. An EC medium containing 10% FBS, 5 ng/mL VEGF-A, and 30 mg/mL dextran was next applied and shear stress was initiated at around 1 dyne/cm² for 12 hours. The shear training was then gradually increased towards 15 dynes/cm² over 24 hours, which was then maintained for an additional 3.5 days. Prior to implantation, shear stress was ramped down to 10 dynes/cm² for 12 hours and then to around 5 dynes/cm² for 12 hours in order to match the shear stress in the IVC of nude rats, where the TEVC is implanted as an interposition graft. (B) Immunostaining of endothelial markers (CD31 and eNOS) and human leukocyte antigen type A (HLA-A). Cross-sectional segments of grafts after luminal endothelialization with hiPSC-ECs were stained for CD31, eNOS, and HLA-A. DNA was counterstained by 4,6-diamidino-2-phenylindole (DAPI). Graft lumen is indicated by the asterisk *. Scale bar = 200 μm. (C) Percentage quantification of the luminal surface of TEVCs covered by hiPSC-ECs based on the immunostaining results in panel B. Quantification was performed based on representative sections of pre-implant of dHUA grafts endothelized using hiPSC-ECs for IVC implantation. Mean values and standard error of the mean indicated by the error bars are shown. Note that n=5 independent hiPSC-TEVCs were investigated. (D) A representative image of the luminal surface of TEVCs after endothelialization using hiPSC-ECs via scanning electron microscopy. Note that hiPSC-ECs adhered to the surface of decellularized TEVCs were aligned in the direction of medium flow and presented a typical elongated morphology under pulsatile, unidirectional flow. The arrow indicates the direction of luminal flow during in vitro culture. Scale bar = 20 μm. Note that n=3 independent TEVCs were investigated. (E) Images of statically EC-coated (static ECs) or shear flow-trained, EC-coated (15–5 dynes/cm²) TEVCs implanted as IVC interposition grafts before and after explantation. Grafts endothelialized with hiPSC-ECs (15–5 dynes/cm²) were trained in bioreactors towards an arterial-like shear stress of 15 dynes/cm² over 36 hours and then maintained for an additional 3.5 days, which was followed by a ramp-down to 10 dynes/cm² for 12 hours and then to around 5 dynes/cm² for 12 hours prior to implantation. Control grafts were coated with hiPSC-ECs statically in bioreactors for overnight pre-implantation. The static coating period was kept under 24 hours as the volume of culture medium filling the luminal space is limited and having no circulation or medium change may impact EC viability negatively in long term. Additionally, coating grafts with hiPSC-ECs statically in bioreactors under 24 hours pre-implantation mimics the prior clinical trial in which patient bone marrow mononuclear cells were seeded onto the polymeric scaffolds statically on the same day the vascular conduits were assembled and implanted. Statically EC-coated grafts were explanted 2 weeks after implantation. Shear flow-trained, EC-coated grafts were explanted 1 month after implantation. Dashed white rectangles indicate grafts implanted into the IVC. Black arrows indicate the formation of a thrombus. Scale bar = 1 mm (a graduation on the ruler). IVC, Inferior vena cava; AO, aorta. Note that n≥3 independent grafts for each group were utilized and implanted into respective nude rats for experiments. (F) Immunostaining of luminal endothelialized followed by 15–5 dynes/cm² flow trained grafts explanted 1 month after implantation. Cross-sectional segments of grafts were stained for eNOS and HLA-A. DNA was counterstained by DAPI. The graft lumen is indicated by the asterisk *. Scale bar = 200 μm. Note that n=4 independent TEVCs were investigated.

Figure 2.. Gradual shear flow training (15–5 dynes/cm²) enhances hiPSC-EC-derived vascular endothelium functions in IVC implanted grafts.

(A) Representative en face images of human fibrinogen adsorbed onto the luminal surface of statically endothelialized grafts exposed to an abrupt flow or of grafts coated with hiPSC-ECs under a gradual shear flow training regimen (15–5 dynes/cm²). Statically endothelialized grafts were prepared by seeding hiPSC-ECs onto dHUA luminal surface followed by an abrupt shear flow of around 5 dynes/cm² for 6 hours in bioreactors to mimic rat IVC shear flow for investigating the necessity of shear flow training for preventing thrombus formation. Scale bar = 25 μm. (B-C) Quantification of endothelial cell coverage and fibrinogen adsorption on graft luminal surface in panel A. Coverage of hiPSC-ECs (marked by tdTomatao; details see START Methods) on graft luminal surface was measured by tdTomatao positive areas (red) over total areas (B). Fibrinogen adsorption (yellow) on graft luminal surface was assessed by mean fluorescence intensity (MFI) of adsorbed fibrinogen (C). Segments (~5 mm in length) around mid-graft regions were immersed in a solution containing 50 μg/mL human plasma fibrinogen conjugated with 647 Alexa Fluorophore for 1 hour at room temperature, followed by PBS washing, formalin (10%) fixation, and cutting into imageable flat sheets, prior to imaging for quantifying EC coverage and fibrinogen adsorption. Mean values and standard error of the mean indicated by the error bars are shown. Multiple independent grafts (n=5) were quantified for EC coverage and fibrinogen adsorption. A nonparametric Mann-Whitney test was used to compare two groups of grafts (**p<0.01). (D) Representative en face scanning electron microscopy images from statically EC-coated grafts exposed to an abrupt flow and shear flow-trained (15–5 dynes/cm²), endothelialized grafts incubated with human whole blood. Segments (~5 mm in length) around mid-regions of grafts were immersed in human whole blood and incubated for 30 minutes at 37°C, followed by PBS washing, formalin (10%) fixation, and cutting into imageable flat sheets, which were further treated with 2.5% glutaraldehyde, ethanol, and carbon for scanning electron microscopy imaging (details see STAR Methods). An appreciable amount of polymerized fibrin and aggregated red blood cells was observed on the luminal surface of statically EC-coated grafts exposed to an abrupt flow compared to the gradually shear flow trained (15–5 dynes/cm²), endothelialized grafts. Scale bar = 25 μm. Multiple independent grafts (n=5) were investigated for whole blood assay. (E) Characterization of endothelialized grafts explanted at 1 day, 7 days, 2 weeks, or 2 months post-implantation via immunofluorescence. Cross-sectional segments of grafts were stained for αSMA, CD31, CD68, HLA-A, and MYH11. DNA was counterstained by DAPI. The graft lumen is indicated by the asterisk *. Scale bar = 200 μm. (F) Quantification of the percentage of endothelial cell coverage on the luminal surface of hiPSC-TEVCs by CD31 and HLA-A immunofluorescence analysis in panel E. Quantification was performed based on representative sections near mid-grafts, and three independent grafts were implanted into three respective nude rats for experiments (n=3). Mean values and standard error of the mean indicated by the error bars are shown. (G) Generation of hiPSC-TEVCs for future therapy for single ventricle defects using allogeneic universal hiPSC-TEVCs. Our novel strategy for generating endothelialized, functional venous conduits will help prevent thrombus formation and stenosis via hiPSC-EC-produced glycocalyx components, anti-coagulant proteins, and basement membrane components enhanced under high shear stress (15 dynes/cm²) training in bioreactors. Prior to implantation of vascular conduits as IVC interposition grafts in nude rats, shear stress was gradually decreased to 5 dyne/cm² to mimic that of the IVC. As a potential future therapy for treating single ventricle defects, off-the-shelf, decellularized donor vascular grafts can be endothelialized with allogeneic universal hiPSC-ECs that are immunocompatible with any patient. These endothelialized grafts could provide immediate endothelial functions, including preventing blood coagulation on biomaterial surface, expediting host EC recruitment, and enabling the remodeling and maturation of grafts into functional host venous conduits for single ventricle treatment.