Multiscale Imaging to Monitor Functional SHED-Supported Engineered Vessels

Abstract

Regeneration of orofacial tissues is hampered by the lack of adequate vascular supply. Implantation of in vitro engineered, prevascularized constructs has emerged as a strategy to allow the rapid vascularization of the entire graft. Given the angiogenic properties of dental pulp stem cells, we hereby established a preclinical model of prevascularized constructs loaded with stem cells from human exfoliating deciduous teeth (SHED) in a 3-dimensional-printed material and provided a functional analysis of their in vivo angiogenesis, vascular perfusion, and permeability. Three different cell-loaded collagen hydrogels (SHED-human umbilical vein endothelial cell [HUVEC], HUVEC with SHED-conditioned medium, and SHED alone) were cast in polylactic acid (PLA) grids and ectopically implanted in athymic mice. At day 10, in vivo positron emission tomography (PETscan) revealed a significantly increased uptake of radiotracer targeting activated endothelial cells in the SHED-HUVEC group compared to the other groups. At day 30, ex vivo micro-computed tomography imaging confirmed that SHED-HUVEC constructs had a significantly increased vascular volume compared to the other ones. Injection of species-specific lectins analyzed by 2-photon microscopy demonstrated blood perfusion of the engineered human vessels in both prevascularized groups. However, in vivo quantification showed increased vessel density in the SHED-HUVEC group. In addition, coinjection of fluorescent lectin and dextran revealed that prevascularization with SHED prevented vascular leakage, demonstrating the active role of SHED in the maturation of human-engineered microvascular networks. This preclinical study introduces a novel PLA prevascularized and implantable construct, along with an array of imaging techniques, to validate the ability of SHED to promote functional human-engineered vessels, further highlighting the interest of SHED for orofacial tissue engineering. Furthermore, this study validates the use of PETscan for the early detection of in vivo angiogenesis, which may be applied in the clinic to monitor the performance of prevascularized grafts.

Keywords: angiogenesis; collagen; magnetic resonance imaging; nuclear medicine; stem cells; tissue engineering.

Figure 1.

In vitro generation and characterization of prevascularized hydrogel yielded in polylactic acid (PLA) constructs. (A) Stereolithography design used for 3-dimensional (3D) printing of the PLA grids, containing 4 inner quadrants. Scale bar: 1 mm. (B) Scanning electron microscopy images of the printed PLA grid showing that the PLA surface was ruffled due to the deposition process. (C) Representative Z projections of the prevascularized hydrogels inside the PLA grids. Robust vessel networks stained with human CD31 were formed by day 4 in vitro in the stem cells from human exfoliating deciduous teeth/human umbilical vein endothelial cells (SHED-HUVEC) and HUVEC cultured with SHED-conditioned medium (CM-HUVEC) groups. No vessel formation was seen in the SHED group. Scale bar: 100 μm. The vertically resliced stacks demonstrate the hydrogel contraction in the presence of SHED. Scale bar: 200 µm. (D) Whole-mount constructs were imaged with 2-photon microscopy, and the 3D vessel parameter quantification was performed using the Skel2Graph3D tool (MATLAB; MathWorks). The 3D quantification results are presented for vessel volume, length, tortuosity, and mean diameter. Vessel volume (μm³) and total vessel length (μm) were normalized to the volume of the z-stack (μm³) to account for hydrogel contraction (n = 5 grids per group). (E) Live/dead cell viability assay was performed in triplicate. Representative Z projections of the prevascularized hydrogels at day 4. Live cells were stained with calcein AM (green) and dead cells with ethidium homodimer (red). The viable cell area was quantified and normalized to the total area (n = 4 grids per group in triplicate). No significant differences were observed between the groups. Scale bar: 100 µm. Data were obtained from 4 fields of view per grid. All error bars represent the standard deviation. *P < 0.05, **P < 0.01.

Figure 2.

Early angiogenesis detection within implanted constructs by positron emission tomography/magnetic resonance imaging (PET/MRI) and later imaging of the vasculature by micro–computed tomography (CT). (A) At 10 d of in vivo implantation, mice (n = 10 per group) were injected with ⁶⁴Cu-NODAGA-RGD, and a PET/MRI was performed 24 h later. Coronal, axial, and sagittal views of PET/MRI data in the 3 groups. Three volumes of interest (VOIs) were placed around polylactic acid (PLA) constructs for quantitative analysis. The uptake intensity is presented on the red scale. Red boxes delineate the implanted grids. Images are 3-mm-thick slices at the implantation site. Scale bar: 2 mm. (B) Quantification and comparison of the maximum standard uptake value (SUVmax) in the 3 groups. SUVmax was measured first for each VOI, and then the mean SUVmax of each grid was calculated (n = 20 grids per group). Data are presented as the mean ± standard deviation. (C) At 30 d of in vivo implantation, mice were injected with an intravascular contrast agent (Exitron; Nano12000). Reconstructed micro-CT acquisition showing segmented vessels in red inside the PLA grid in gray. (D) The total vascular volume was quantified within each grid (n = 12 grids per group). Scale bar: 1 mm. All error bars represent standard deviation. *P < 0.05, **P < 0.01, ****P < 0.0001.

Figure 3.

Stem cells from human exfoliating deciduous teeth/human umbilical vein endothelial cell (SHED-HUVEC) engineered vessels successfully anastomosed with host vessels and produced a dense vascular network in vivo. (A, B) Representative Z planes and orthogonal views of anastomoses in the SHED-HUVEC and HUVEC cultured with SHED-conditioned medium (CM-HUVEC) groups at 1 mo postimplantation. Red indicates human-engineered vessels perfused with UEA lectin, and cyan indicates mouse vessels perfused with IB4 lectin. Scale bar: 50 µm. (C) Z stack projections of lectin-stained vessels within the polylactic acid grids and quantification of the total vascular density (n = 4 grids per group). Scale bar: 100 µm. (D) Vascular morphological properties were analyzed on Z projections of the UEA I stained vascular skeleton. Vessel length (mm) and number of branches and branching points were normalized to the total area of the Z projection (mm²). The branching index refers to the ratio of the number of branches to vessel length. Quantification was performed on Z projections of 60-μm-wide substacks of the total constructs (n = 12 grid quadrants per group). Scale bar: 100 µm. All error bars represent standard deviation. **P < 0.01, ***P < 0.001, ****P < 0.0001.

Figure 4.

Functional perfusion of stem cells from human exfoliating deciduous teeth/human umbilical vein endothelial cells (SHED-HUVEC) engineered vessels. (A) Human vessels (stained with UEA lectin in red) were well perfused after intravenous injection of a 70-kDa dextran (green) at 2 wk after implantation, without vascular leakage of the dextran in the interstitial space (n = 3 mice per group). Scale bar: 100 μm. (B) Closer view of a SHED-HUVEC engineered vessel and its lumen perfused with dextran. Scale bar: 20 μm. (C) Less perfused vessels were observed in the HUVEC cultured with SHED-conditioned medium (CM-HUVEC) group. Scale bar: 100 μm. (D) Arrows point to vascular leakage of dextran into the interstitial space. Scale bar: 20 µm. All images were obtained from clarified whole-mount samples with 2-photon microscopy.

Figure 5.

Effects of stem cells from human exfoliating deciduous teeth (SHED) on in vivo vascular area, perivascular innervation, and cell apoptosis. (A) Human and mouse-specific CD31 staining in the prevascularized and non-prevascularized group, respectively, to quantify the vascular area. Scale bar: 20 µm. (B) Sensitive fibers within the constructs stained with calcitonin gene-related peptide (cGRP) in red and cGRP area quantification. Scale bar: 20 µm. (C) Apoptosis was assessed with a terminal deoxynucleotidyl transferase dUTP nick end labeling assay (TUNEL) assay. The number of apoptotic cells (depicted in green) was normalized to the total number of cells. Nuclei were stained with DAPI (blue). Scale bar: 50 µm. (D–F) Quantification of CD31 (n = 8 per group), cGRP, and TUNEL expression (n = 5 per group). All error bars represent standard deviation. *P < 0.05, ***P < 0.001, ****P < 0.0001.