Implanted Tissue-Engineered Vascular Graft Cell Isolation with Single-Cell RNA Sequencing Analysis

Abstract

The advent of single-cell RNA sequencing (scRNA-Seq) has brought with it the ability to gain greater insights into the cellular composition of tissues and heterogeneity in gene expression within specific cell types. For tissue-engineered blood vessels, this is particularly impactful to better understand how neotissue forms and remodels into tissue resembling a native vessel. A notable challenge, however, is the ability to separate cells from synthetic biomaterials to generate high-quality single-cell suspensions to interrogate the cellular composition of our tissue-engineered vascular grafts (TEVGs) during active remodeling in situ. We present here a simple, commercially available approach to separate cells within our TEVG from the residual scaffold for downstream use in a scRNA-Seq workflow. Utilizing this method, we identified the cell populations comprising explanted TEVGs and compared these with results from immunohistochemical analysis. The process began with explanted TEVGs undergoing traditional mechanical and enzymatic dissociation to separate cells from scaffold and extracellular matrix proteins. Magnetically labeled antibodies targeting murine origin cells were incubated with enzymatic digests of TEVGs containing cells and scaffold debris in suspension allowing for separation by utilizing a magnetic separator column. Single-cell suspensions were processed through 10 × Genomics and data were analyzed utilizing R to generate cell clusters. Expression data provided new insights into a diverse composition of phenotypically unique subclusters within the fibroblast, macrophage, smooth muscle cell, and endothelial cell populations contributing to the early neotissue remodeling stages of TEVGs. These populations were correlated qualitatively and quantitatively with immunohistochemistry highlighting for the first time the potential of scRNA-Seq to provide exquisite detail into the host cellular response to an implanted TEVG. These results additionally demonstrate magnetic cell isolation is an effective method for generating high-quality cell suspensions for scRNA-Seq. While this method was utilized for our group's TEVGs, it has broader applications to other implantable materials that use biodegradable synthetic materials as part of scaffold composition.

Impact statement

Single-cell RNA sequencing is an evolving technology with the ability to provide detailed information on the cellular composition of remodeling biomaterials in vivo. This present work details an effective approach for separating nondegraded biomaterials from cells for downstream RNA-sequencing analysis. We applied this method to implanted tissue-engineered vascular grafts and for the first time describe the cellular composition of the remodeling graft at a single-cell gene expression level. While this method was effective in our scaffold, it has broad applicability to other implanted biomaterials that necessitate separation of cell from residual scaffold materials for single-cell RNA sequencing.

Introduction

The field of tissue engineering and regenerative medicine seeks to provide better alternatives to existing treatments in a variety of organ systems, including skin, bone, cardiovascular, hepatic, renal, and gastrointestinal.^1–5 Broadly, the goal is to create tissue that provides the same structure and function of the native tissue it seeks to repair, replace, or restore. Biodegradable scaffolds are routinely utilized to provide physical support for cells during tissue growth and remodeling and may be composed of natural and/or synthetic materials.

Our investigational efforts have focused on addressing pediatric congenital heart defects through surgical intervention with tissue-engineered vascular grafts (TEVGs), heart valves, and cardiac patches.^6–8 The overarching challenges we face are similar to others: optimizing the growth rate with scaffold degradation, maximizing the quality of neotissue, and limiting pathological remodeling. Before implantation, our TEVGs are made by seeding scaffolds with bone-marrow-derived cells.^9,10 Our greatest translational barrier is graft stenosis, thrombosis, or a combination of both.^11,12 Elucidating the underlying mechanisms of graft occlusion has focused on using animal models to investigate the host response to implanted TEVGs and subsequent TEVG remodeling.

Our investigative methods in these studies have included immunohistochemistry (IHC), flow cytometry, reverse transcription-quantitative polymerase chain reaction (RT-qPCR), and bulk RNA sequencing to characterize cells within a remodeling TEVG. Single-cell RNA sequencing (scRNA-Seq) is an emerging tool allowing for nonbiased individual gene expression profiling of thousands of cells. Bioinformatics applied to scRNA-Seq data enables identification of cells and insight into their function. Incorporation of this tool can provide greater clarity to the cells responsible for turning an inert scaffold into functioning tissue.

A critical component of success to scRNA-Seq is generation of high-quality single-cell suspensions from tissue digests.¹³ High quality refers to cells that are clean of debris, viable, and minimally manipulated. While previous reports exist on documenting approaches with successful sequencing output from native tissues, they ultimately have poor transferability to biodegradable synthetic scaffolds.^14,15 Our TEVG scaffold is a porous polyester composite that degrades over a period of 3–12 months in vivo. In our initial efforts optimizing single-cell suspension generation from explanted TEVGs, we faced three experimental difficulties: cell dissociation, cell yield, and scaffold removal. The excess presence of scaffold debris resulted in column clogging at the 10 × Genomics stage during individual cell barcoding and library construction.

We present here a new approach aimed at optimizing generation of single-cell suspensions for downstream scRNA-Seq from implanted TEVGs. Similar to previously described approaches, we first used enzymatic and mechanical dissociation to separate cells from scaffold and extracellular matrix. Key to this study, we then utilized a novel application of antibody-labeled magnetic beads to selectively enrich dissociated cells from contaminating scaffold debris. These methods focus on efficient cell dissociation from the scaffold with removal of nondegraded scaffold material to prevent cell clumping and column clogging.

We applied this method generating scRNA-Seq results of explanted TEVGs. We provide IHC of explanted TEVGs to further correlate the gene expression profile of cell populations with cell protein markers and discuss the advantages/limitations of both methods. While these methods were applied to murine TEVGs, they have broad applicability to other experimental approaches utilizing implantable biomaterials that are plagued with significant scaffold debris following tissue digestion steps.

Materials and Methods

The process of generating a single-cell suspension from the implanted scaffold was divided into three components: (1) scaffold enzymatic digestion and mechanical dissociation, (2) serial cell filtration, and (3) magnetic-activated cell sorting (MACS). The application of MACS to separate cells from scaffold debris represents a novel application of this technology to overcome the challenge of separating scaffold debris from cells for a scRNA-seq workflow. In the first stages, explanted grafts were finely minced and incubated with a collagenase/dispase solution. Following incubation and standard sample filtration through cell strainers, cell suspensions were separated from scaffold debris through MACS. The MACS workflow required use of mouse cell depletion kit (Miltenyi Biotec; Cat no. 130-104-694), QuadroMACS Separator (Miltenyi Biotec; Cat no. 130-091-051), and LS columns (Miltenyi Biotec; Cat no. 130-042-401).

Experiment

Animal welfare statement

Animal use was done under oversight of the Abigail Wexner Research Institute (AWRI) at Nationwide Children's Hospital Institutional Animal Care and Use Committee (IACUC no. AR20-00094). The AWRI is registered as a research facility with the United States Department of Agriculture (USDA) and is fully accredited by the Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC) (no. 31-R-0166). An animal welfare assurance is on file with the Office for Protection from Research Risks at the National Institutes of Health (OPRR-NIH). All procedures and experiments were conducted in accordance with the NIH and the American Veterinary Medical Association (AVMA) guidelines.

TEVG fabrication and scanning electron microscope imaging

TEVGs were obtained from Gunze Limited and fabricated from a nonwoven polyglycolic acid (PGA) felt and a copolymer sealant solution of poly-L-lactide and –e-caprolactone (P[CL/LA], PCLA) as previously described.¹⁶ Implanted scaffolds were 0.91 mm in diameter and 3–4 mm in length at time of implantation (Fig. 1a). Scanning electron microscope (SEM) imagining was provided by Gunze Limited (Fig. 1b).

FIG. 1.

Views of the TEVG at pre- and post-murine implantation. (a) Gross longitudinal and top-down view of a 20-gauge (0.91 mm) scaffold. Scale provided in millimeters. (b) SEM images of the scaffold wall and of the luminal surface before scaffold implantation. (c) Surgical images of the scaffold at 2 weeks postimplantation. The left image shows the scaffold (outlined in yellow) as an interposition inferior vena cava graft. The dashed black line marks the proximal portion of the graft inferior to the renal vessels. The white * marks caudal position of the distal end of the graft. Scale bar = 2 mm. The right image is a top-down view of the scaffold after being explanted. Scale bar = 1 mm. SEM, scanning electron microscope; TEVG, tissue-engineered vascular graft. Color images are available online.

Immunohistochemistry

Following explant, TEVGs were formalin fixed overnight at 4C and paraffin embedded. Grafts were cross sectioned through the middle with serial 4 μm sections made and heat fixed to glass slides. For IHC, our laboratory standard protocol was utilized as previously described.¹⁷ The following primary antibodies and respective concentrations were used: anti-Ki67 1:1000 (Abcam; ab15580), anti-CD68 1:2000 (Abcam; ab125212), anti-Ly6c 1:2000 (Abcam; ab15627), anti-F4/80 1:1000 (Bio-Rad; MCA497), anti-Myh11 1:3000 (Abcam; ab125884), anti-α-SMA 1:8000 (Abcam; ab124964), anti-Col1a1 1:250 (Millipore Sigma; ABT 257), anti-CD3 1:500 (Abcam; ab16669), anti-Pax5 1:2000 (Abcam; ab109443), anti-Lyve-1 1:1000 (Abcam; ab14917), anti-Calprotectin (S100a8 + S100a9) 1:20,000 (Abcam; ab288715), anti-vWF 1:1000 (Dako; A0082), anti-Calponin 1 1:1000 (Abcam; ab46794), and anti-CD31 1:250 (Abcam; ab28364).

Primary antibody binding was detected with biotinylated goat anti-rabbit IgG antibody (Vector; Cat no. BA-1000) and biotinylated goat anti-rat IgG (Vector; Cat no. BA-9401) followed by binding of VECTASTAIN Elite ABC-HRP Reagent (Cat no. PK-7100) and subsequent chromogenic development with 3,3-diaminobenzidine (Vector).

Microscopy imaging

Brightfield images were obtained on a Zeiss Axio Imager.A2 inverted microscope (Carl Zeiss, Oberkochen, Germany). Images were obtained utilizing a tiled approach with a 20 × objective lens. For quantified stains, images were processed in ImageJ (NIH) Version 2.3.0/1.5.3f using built-in features in addition to a color deconvolution plugin. Clustered hematoxylin-stained nuclei and IHC-stained cells were distinguished using watershed separation.

TEVG implantation and collection

Eight to 12-week-old female C57BL/6J mice underwent TEVG implantation as previously described.¹⁸ Two weeks post-TEVG implantation, mice were anesthetized, and an abdominal incision was made to expose the implanted TEVG (Fig. 1c). Grafts were excised below and above the proximal and distal anastomoses as marked by sutures, respectively, to ensure that no native inferior vena cava tissue was harvested. Grafts were immediately placed in room temperature calcium-free, magnesium-free phosphate-buffered saline (PBS) (Fig. 2).

FIG. 2.

Overview of the experimental approach to scRNA-Seq. Two weeks post-TEVG implantation, grafts were explanted, placed in PBS, and finely minced in a Petri dish. Enzymatic digestion cocktail was added to mechanically dissociated grafts and placed in a 37C incubator on a plate shaker for 1 h. Following enzymatic digestion, grafts digests were washed, centrifuged, and filtered through a 70 μm cell strainer. This was repeated with a 30 μm cell strainer. Digests were processed through the MACS cell separation protocol and counted before undergoing 10 × Genomics scRNA-Seq¹. ¹Parts of this figure were drawn by using pictures from Servier Medical Art. Servier Medical Art by Servier is licensed under a Creative Commons Attribution 3.0 Unported License (https://creativecommons.org/licenses/by/3.0/). MACS, magnetic-activated cell sorting; PBS, phosphate-buffered saline; scRNA-Seq, single-cell RNA sequencing. Color images are available online.

Graft mincing, digestion, and mechanical dissociation

Grafts were placed in 10 cm Petri dishes with residual PBS and mechanically minced with razor blades until finely dissected. Minced grafts were transferred to a 24-well dish with each well containing 250 μL of collagenase/dispase solution at a concentration of 1 mg/mL (Sigma Aldrich; Ref no. 10269638001). The plate was placed into a 37°C, 5% CO₂ humidified incubator on a plate shaker set to 300 rpm for 1 h. Following incubation, 1 mL of serum-free Dulbecco's modified Eagle's medium (DMEM) was immediately added to each well to neutralize enzymes. Grafts and media were transferred to a 15 mL conical tube, wells were rinsed with PBS, and the final conical tube volume was brought to 5 mL. Using a 5 mL serological pipette, graft samples were mechanically dissociated via vigorous pipetting until the solution reached a cloudy appearance (Fig. 2).

Sample filtration

For debris removal, samples were passed through a 70 μm fine mesh cell strainer into a 50 mL conical tube. Filtration was achieved by placing the serological pipette tip against the cell strainer and forcefully dispensing the cell containing solution. The 15 mL conical tube was rinsed with 10 mL of PBS and then passed through the cell strainer. An additional 1 mL of DMEM was passed over the cell strainer for rinsing. The final volume in the 50 mL conical was 15–16 mL. Samples were centrifuged at 300g for 15 min, supernatant was carefully aspirated, and cells were gently resuspended in 1 mL of ice-cold PBS.

In preparation for the MACS step, 30 μm Miltenyi cell strainers (Cat no. 130-041-407) were placed over 1.5 mL Eppendorf tubes and prewet with 200 μL of ice-cold PBS. Resuspended cells were passed through the filter, their conical tubes were rinsed with 300 μL of ice-cold PBS, and the rinse solution was used to rinse the 30 μm cell strainer. Following centrifugation at 300g for 10 min in a 4°C prechilled table-top centrifuge, samples were ready for MACS separation (Fig. 2).

MACS separation

All manufacture protocol steps were followed with no significant modifications. Care was given to the use of properly degassed 0.5% bovine serum albumin in PBS separation buffer that significantly reduced column clogging. Following separation and resuspension, cells were analyzed with trypan blue to confirm >90% cell viability before scRNA-Seq processing.

Clustering analysis

3′-end scRNA-Seq and data processing

GEM generation, barcoding, post-GEM-RT cleanup, cDNA amplification, and cDNA library construction were performed using Chromium Next GEM Single Cell 3′ v3.1 (10 × Genomics CG000315 Rev C). Libraries were sequenced on an Illumina NovaSeq6000 platform. Transcripts were matched to the mm10 reference (GRCm39) using Cell Ranger v6.0.0. Cell counting, suspension, GEM generation, barcoding, post-GEM-RT cleanup, cDNA amplification, library preparation, quality control, and sequencing were performed at the Institute for Genomic Medicine within the AWRI at Nationwide Children's Hospital in Columbus, OH.

Quality control

Quality control and downstream analysis of scRNA-seq data were performed in RStudio v 1.3.1093 using Seurat 4.0.4.^19–22 First, low-quality cells were removed. Initially, cells were retained that displayed greater than 500 RNA counts/cell, >200 genes/cell, and mitochondrial gene expression between 0% and 5%. Gene expression was normalized, and the top 2000 variable features were identified based on the variance-stabilizing transformation method, followed by performing principal component analysis (PCA). The top 36 PCs were used for clustering with the Louvain algorithm to generate cell clusters (FindClusters function, 20 clusters with resolution = 0.5). To represent high-dimensional data in two-dimensional space, we performed Uniform Manifold Approximation and Projection analysis using the PCA results. One cluster appeared anomalous, possessing relatively low total RNA reads, unique genes, and % mitochondrial gene expression per cell. Differential gene expression revealed that this cluster as neutrophils.

Due to the overlap in the quality control (QC) parameter values for neutrophils and low-quality cells, the Seurat object was then subset into two groups: neutrophils and all non-neutrophils so that refinement of QC parameters would not result in loss of neutrophils. Non-neutrophils that displayed <2500 RNA counts/cell or <500 genes/cell were determined to be enriched for low-quality cells or empty droplets containing background RNA and were excluded. Upon merging subsets, renormalization, and clustering, feature plots for these cell type-specific genes suggested the presence of doublets as many nonleukocyte clusters contained a small population of cells with high leukocyte gene expression. Doublets were identified and removed manually.

One subset contained endothelial cells (ECs) and smooth muscle cells (SMCs) and the other contained fibroblast and lymphatic ECs, respectively. For each subset, renormalization and clustering resulted in distinct clusters enriched for suspected doublets as identified by the high expression of leukocyte genes. After the removal of low-quality cells, 7223 cells remained for downstream bioinformatic analyses. The object was split by sample, renormalized using SCTransform, samples were reintegrated, and clustering was repeated.

Results

scRNA-Seq data from explanted TEVGs (n = 2) were analyzed to provide an overall cellular makeup breakdown. Analysis revealed a diverse population of cells dissociated from the explanted scaffold that broadly fell into the categories of endothelial cells, leukocytes, fibroblasts, and SMCs (Fig. 3a). The greatest subpopulations of all cells identified were macrophages and fibroblasts highlighting the inflammatory environment of the scaffold at time of explant (Fig. 3b, c).

FIG. 3.

Overview of scRNA-Seq results for explanted TEVGs. (a) UMAP plot showing the various clusters that were identified. Inset is a feature plot of CD45 expression, which differentiates leukocytes from nonleukocytes. (b) Visualization of relative size of generalized cell cluster identities. (c) Individual cluster identities and percent composition. (d) Cross-sectional image of an H&E-stained TEVG. Black box outlines the region from which IHC images were taken. Scale bar = 200 μm. H&E, hematoxylin and eosin; IHC, immunohistochemistry; UMAP, Uniform Manifold Approximation and Projection. Color images are available online.

Explanted scaffolds (n = 5) were utilized for histological analysis with quantification of hematoxylin and eosin (H&E) nuclei (Fig. 3d) to determine approximate total cells within the scaffold relative to the number of cells recovered following processing. Quantification of H&E images (n = 5) identified 8529 ± 3677 nuclei within a 4 μm tissue section (Supplementary Fig. S1). After adjusting scaffold length for tissue shrinkage during processing and undercounting of weakly stained nuclei, each TEVG contains an estimated ∼2 × 10⁶ cells on average.

Endothelial cells

Rapid and complete graft luminal endothelialization offers a significant protective barrier against stenosis and thrombosis.^23–25 Intramural vascularization by blood and lymphatic vessels support the developing neotissue. We identified two distinct cell clusters contributing to graft endothelialization: blood endothelial cells (BECs) and lymphatic endothelial cells (LECs). Identification of the BEC cluster was primarily based on expression of Pecam1 and the absence of Lyve1 (Fig. 4a, b). The LECs cluster was primarily identified by both Pecam1 and Lyve1 expression (Fig. 4c).^26,27 Pecam1 encodes expression of a cell-cell adhesion protein, also known as CD31. CD31 is nonspecific for distinguishing BECs from LECs but has been reported to be expressed less in LECs in congruence with our current results.²⁷ High Vwf expression is limited to a subset of BECs and megakaryocytes with relatively low expression in LECs.^28,29

FIG. 4.

Endothelial cells. Feature expression data plots and IHC images for (a) Pecam1/CD31, (b) vWF, and (c) Lyve1. Scale bar = 50 μm; dashed line = scaffold; stars = vessel negative for Lyve1 staining; arrow = stained vessel. A, aorta; L, lumen; LN, lymph node; NS, nonspecific staining; S, suture. Color images are available online.

We have previously utilized both anti-CD31 and anti-vWF antibodies for histological evaluation of neovessel formation and luminal endothelialization of our TEVGs.^6,17,30 The crossover expression of Pecam1 between BECs and LECs demonstrates that IHC targeting vWF and LYVE1 protein expression has greater specificity than IHC targeting CD31. Indeed, IHC staining of an explanted TEVG for vWF and LYVE1 demonstrates two distinct endothelial cell populations in support of our scRNA-Seq results (Fig. 4b, c). Additional gene expression profiles of the endothelial cell clusters are shown in Supplementary Figure S2.

Innate immune system cells

Host cell-mediated inflammatory response to implanted biomaterials represents a key driver of neotissue growth and remodeling. The character and extent of the inflammatory response ideally should provide a balanced environment where positive aspects of wound healing such as angiogenesis, stem cell proliferation, and timely scaffold degradation occur.^31,32

Part of the early innate immune response in wound healing, and by extension the presence of a biomaterial, is neutrophil trafficking and infiltration. scRNA-Seq data analysis identified a distinct neutrophil cluster based on the robust and specific expression of S100a8 (Fig. 5a). The S100a8 protein, which functions through Ca2+-mediated signaling, is similarly highly expressed in neutrophils making up almost 50% of cytosolic expressed proteins and IHC found this protein to be expressed in cells throughout the TEVG (Fig. 5a).³³ From a quantitative analysis comparison perspective, scRNA-Seq identified 7.2 ± 0.3% (range 7.0–7.5%, n = 2) of cells as neutrophils, compared to 2.3 ± 1.5% (range 0.6–4.8%, n = 5) identified by IHC (Fig. 3c, Supplementary Fig. S1).

FIG. 5.

Myeloid lineage cells. Feature expression data plots and IHC images for (a) S100A8/Calprotectin (S100A8 and S100A9), (b) Ly6c2/Ly6c, (c) Adgre1/F4/80, and (d) CD68. Scale bar = 50 μm. Color images are available online.

Following neutrophil response, chronic graft inflammation is principally mediated via monocytes/macrophages.³⁴ In our own grafts, we have found macrophages to constitute a major cellular population at 2 weeks postimplantation.³⁰ They represent a critical driver of neotissue formation. Elimination of the macrophage inflammatory response in mice TEVG recipients through host macrophage depletion results in failed engraftment.³⁵ Of the 36 cell clusters identified, the most abundant population was macrophages and monocytes (Fig. 3a–c). High Ly6c2 expression identified a monocyte cluster that was separate from the large macrophage cluster (Fig. 5b). By IHC, few LY6C+ cells were observed within the TEVG. Two markers, F4/80, the antigen encoded by the Adgre1 gene, and CD68, are routinely utilized for identification of tissue resident macrophages.³⁶

Indeed, expression of Adgre1 and Cd68 was differentially expressed between subpopulations of macrophages with some expression crossover (Fig. 5c, d). The distinct functional implications for F4/80 (high) and CD68 (low) expression macrophages, and vice versa, remain unclear in the literature. IHC analysis of scaffolds revealed significant expression of both proteins in congruence with expression data, but with distinct staining patterns. For example, CD68 demonstrated a more prominent staining pattern in the multinucleated giant cells surrounding the scaffold PGA fibers (Fig. 5c, d).

Expression of Il1b, H2-Aa, and Cd74 further identified macrophage populations distinct from Adgre1 and Cd68 (Supplementary Fig. S3). Expression of these genes has been previously reported to reflect phenotypically unique subpopulations of macrophages although the precise implications of these on TEVG remodeling remain unknown.^37,38 Additional leukocyte feature plots of gene expression included markers for natural killer cell and basophil subpopulations (Supplementary Fig. S4).

Adaptive immune system cells and proliferating cells

The adaptive immune system role in driving TEVG neotissue deposition and remodeling remains an area that has received less attention than its innate counterpart.³⁹ We were able to identify distinct T cell (Cd3 expression) and B cell (Pax5) populations within 2-week TEVGs (Fig. 6a, b). Expression of CD3 serves as a pan T cell marker capturing both CD4⁺ and CD8⁺ T cells.^40,41 Similarly, Pax5 is expressed in all B cells until differentiation into plasma cells and was expressed primarily within cluster 29 (Fig. 3a–c).⁴² The limited presence of these cells on scRNA-Seq is reflected on IHC staining of explanted scaffolds (Fig. 6a, b). The precise impact of T cell and B cell origin cells in the remodeling TEVG remains unclear. PCL, one of the two polymers comprising the scaffold, has been recently shown to specifically recruit an antigen presenting phenotype of B cells that drive collagen deposition.⁴³

FIG. 6.

Lymphocytes. Feature expression data plots and IHC images for (a) CD3e/CD3, (b) Pax5, (c) Mki67/Ki67. Scale bar = 50 μm. Color images are available online.

Various cell types expressed the proliferation marker Mki67 and these proliferating cells were often distinguished as a separate cluster such as clusters 22 (macrophages), 33 (fibroblasts), and 34 (SMCs) (Figs. 3a, c and 6c). Cells expressing the Ki67 protein could be found throughout the TEVG. We performed quantitative analysis to compare sequencing results to IHC results. For Mki67, scRNA-Seq identified 3.0 ± 0.1% (range 2.9–3.1%, n = 2) of cells as expressing Ki67 compared to 14.9 ± 7.7% (range 8.3–28.0%, n = 5) identified by IHC (Supplementary Fig. S1). The somewhat limited expression of Mki67 may be reflective of an inherent limitation to scRNA-Seq losing multinucleated, foreign body giant cells, which have strong Ki67 expression, during the doublet exclusion process.³⁰

Fibroblasts and SMCs

Fibroblasts, in native blood vessels, are primarily located to the adventitial layer producing ECM for vascular structural support.⁴⁴ However, they have a more nuanced role, including involvement in reactive oxygen species production, vasoactivity, and growth factor production.⁴⁵ Identification of gene and protein markers highly specific for fibroblasts has been a historical challenge.⁴⁴ Recent work utilizing single-cell analysis of murine fibroblasts concluded that fibroblasts are heterogeneous with no pan fibroblast maker.⁴⁶ Our results demonstrated Col1a1 expression as the closest pan fibroblast marker in the TEVG cell population (Fig. 7a). Its protein expression was not limited to the adventitia, but rather found throughout (Fig. 7a). Acta2 expression in a subcluster of Col1a1 fibroblasts indicates a myofibroblast phenotype known for mediating the wound healing response (Fig. 7b).^47,48

FIG. 7.

Fibroblasts and SMCs. Feature expression data plots and IHC images for (a) Col1a1, (b) Acta2/αSMA, (c) Myh11, and (d) Cnn1/Calponin 1. Scale bar = 50 μm. SMCs, smooth muscle cells. Color images are available online.

Expression of Eln, Clu, Ly6a, and Cd34 was found to differentiate fibroblasts into additional subclusters (Supplementary Fig. S5). CD34⁺ fibroblasts have been described to reside as adventitial cells promoting stroma maintenance/angiogenesis, and expressing a proinflammatory transcriptomic profile with secretion of promonocyte recruitment proteins.^49–51 CD34 expression also overlaps with Ly6a/Sca1+ and may identify a population of adventitial fibroblasts (Supplementary Fig. S5) shown to proliferate and differentiate into new SMCs after severe vessel injury.⁵²

Proliferation and migration of vascular SMCs in a remodeling graft is necessary for generation of a blood vessel capable of regulatory functions, and for neovessel stabilization. Single-cell analysis of the remodeling graft localized SMCs to clusters 19, 21, 23, and 34 (Fig. 3a, c). All four clusters highly expressed Myh11 and Acta2, with cluster 19 having the highest Cnn1 expression (Fig. 7b–d). Myh11 and Acta2 expression is known to be highly specific to vascular SMCs. Acta2 is notable for expression in cells with contractile capabilities and as one of the earliest markers of progenitor cells differentiating into SMCs.^53–56 Cnn1 is specific to SMCs providing contractile function, but expressed at high levels only in a subset of SMCs.⁵⁷

IHC expression showed diffuse expression of αSMA (i.e., Acta2 gene product) and Calponin-1 (i.e., Cnn1 gene product) proteins (Fig. 7b, d). Myh11 protein expression observed in vSMCs surrounding the neovessels within the TEVG, but a confluent tunica media was not apparent (Fig. 7c).

Discussion

Despite the relatively recent introduction of scRNA-Seq to the scientific community in 2009, its widespread use and impact has been profound.⁵⁸ With this technology has come the ability to provide an in-depth analysis of the complex genomic identity of cells from in vitro cultures, native tissues, and tissue-engineered constructs. This present work sought to expand the potential use of this technology to a relatively underreported context: implantable tissue-engineered constructs fabricated from biodegradable synthetic materials.

For this, we utilized our group's TEVG fabricated from a composite PGA/PCLA scaffold. The size range of polymer fragments poised unique challenges in separating cells from scaffold material. We utilized a well-established separation technique in a novel application by incubating mechanically and enzymatically digested scaffolds with magnetic beads coupled to antibodies targeting mouse lineage cells. With this approach, we were able to remove enough scaffold debris to generate high-quality single-cell suspensions for 10 × Genomics processing. scRNA-Seq revealed for the first time the significant heterogeneity of cells populating the remodeling of TEVG. The TEVG was largely populated by macrophages driving a host inflammatory response.

This observation supports previous work by our group aimed at characterizing the host response to TEVGs utilizing IHC and flow cytometry.^30,59 Detailed transcriptomic analysis revealed distinct subpopulations within the macrophage cluster. Our group and others have attempted to better characterize the unique phenotypes of scaffold macrophages utilizing IHC and RT-qPCR approaches, but have faced limitations in granularity and specificity.^17,60,61 Similarly, our single-cell results showed phenotypically diverse subpopulations within the endothelial cell, leukocyte, fibroblast, and SMC clusters.

To provide insight into the populations described by scRNA-Seq in relationship to cells in the scaffold, we used qualitative and quantitative histological analysis. The fraction of cells expressing Ki67 was ∼5-fold greater by IHC than scRNA-Seq and, conversely, the fraction of cells expressing Calprotectin was ∼3-fold less by IHC than scRNA-Seq. There are many explanations for these discrepancies. Specific to these data, two samples were evaluated by scRNA-Seq and five by IHC, so the difference in values may be due to insufficient sampling. Another explanation is that the population of cells isolated from the remodeling scaffold before scRNA-Seq is not representative.

One such example is multinucleated giant cells that encapsulate the scaffold polymer and are the major cellular constituent within the scaffold at 2 weeks. These giant cells are not expected to easily dissociate from the scaffold intact or be retained through the final QC process that enriches for viable single cells. Since the estimated cell number in histological sections is based on the number of nuclei, multinucleated giant cells inflate the estimated cell number, decreasing the calculated frequency of cells identified by IHC. In turn, the absence of giant cells in the scRNA-Seq data artificially inflates the frequencies of all cells.

In addition, scRNA-Seq detects mRNA expression, whereas IHC detects expressed proteins. Post-transcriptional regulation of protein expression may allow for the mRNA to be present without translation of the encoded protein or the expression of mRNA and protein may be out of phase with one another. For example, Ki67 mRNA expression changes throughout the cell cycle increasing in G1 and peaking in G2.⁶² Ki67 protein accumulates throughout the cell cycle, peaking in mitosis. After cell cycle exit, Ki67 protein is degraded with a half-life of 90 min.⁶³ Therefore, Ki67 protein may be present in quiescent cells possessing low levels of Ki67 mRNA with the level of the protein dependent on how long the cells has exited the cell cycle.

There are also many factors that influence the outcome of IHC staining and accuracy of quantification. These include IHC method and optimization as well as antibody clonality and specificity. Cell-specific challenges exist, including the location of expression (e.g., cell membrane, cytoplasm, nucleus), cell morphology, and frequency of homotypic cell-cell contacts. These factors affect the intensity and pattern of staining, which are critical for sensitivity and specificity of identifying cells during image processing that requires user defined threshold levels for staining intensity, cell size, and shape.

Beyond the challenges faced in direct quantitative comparison between scRNA-Seq and IHC, this current approach faces additional limitations. The process of scaffold decellularization consequentially sacrifices the ability to relate individual cell sequencing results to the cell's spatial positioning within the scaffold. For example, the cluster of endothelial blood cells identified on sequencing is unable to immediately be differentiated between luminal surface versus intramural. An additional limitation is the use of magnetic beads to separate cells from scaffold and the potential for its influence on the final population of cells undergoing scRNA-Seq. Reports of alternative approaches remain limited. Recent works by Greaney et al described a scraping approach to release cells from a rat trachea for sequencing, but this tissue-engineered scaffold was limited to an vitro platform.⁶⁴

Similar work looking at scRNA-Seq of tissue-engineered cardiac scaffolds was able to utilize standard cell dissociation techniques given the in vitro environment.⁶⁵ Other recent work utilizing scRNA-Seq to assess remodeling of 3D printed hydrogels in a rat critical size cranial bone defect model did not require additional methods to separate hydrogel from cells given the natural degradation of the biomaterial.⁶⁶ In a rat model, density-based centrifugation successfully separated cells from silk fibers to generate single-cell suspensions for scRNA-Seq allowing for the assessment of the host macrophage response.⁶⁵ Future work by our group and others in the tissue engineering space will focus on exploring other approaches to separate nondegraded synthetic biomaterial from cells and the potential influence of these approaches on final sequencing results.

Conclusion

The results presented here demonstrate that MACS separation is an effective approach for separating cells from ECM and nondegraded synthetic scaffold material in remodeling tissue-engineered constructs for a downstream scRNA-Seq workflow. These results of a diverse cell population are supported by historical work by our group and further supported by IHC analysis of distinct cell populations within the scaffold. While we focused largely here on cell identity and clustering as proof of concept, the potential impact from the wealth of data generated is significant. Insight into the cell origin and phenotype in a remodeling TEVG or other implantable scaffold will enable rational design improvements and therapeutic strategies to optimize regeneration of target tissue.

Footnotes

Acknowledgments

The authors thank the Institute for Genomic Medicine at the AWRI at Nationwide Children's Hospital, and in particular, Jesse Westphal, Dr. Katherine Miller, and Dr. Tracy Bedrosian for their assistance in scRNA-Seq. The authors thank the Animal Resource Core for their assistance in animal welfare and care as well as the Morphology Core for its assistance with TEVG processing, embedding, and sectioning.

Authors' Contributions

G.J.M.M.: conceptualization, formal analysis, investigation, methodology, project administration, writing—original draft, and writing—review and editing; J.C.B.: conceptualization, investigation, methodology, project administration, visualization, and writing—review and editing; C.K.B.: conceptualization, funding acquisition, project administration, resources, leadership, supervision, and writing—review and editing; J.W.R.: conceptualization, formal analysis, investigation, methodology, project administration, software, writing—original draft, and writing—review and editing.

Disclosure Statement

C.K.B. is an inventor on patent/patent applications (2015252805 [Australia], 2016565483 [Japan], 855,370, 9,446,175, 9,782,522, 10,300,082, 61/987,910, 62/266,309, 62/309,285, 62/209,990, 62/936,225) submitted by Yale University and/or Nationwide Children's Hospital that cover methods of improving the design, manufacturing, or performance of tissue-engineered vascular grafts. C.K.B. is a founder of Lyst Therapeutics.

Funding Information

This work was, in part, supported by the Additional Ventures organization 2020 innovation funds to author C.K.B. titled “Unlocking our Regenerative Capacity: Elucidating the Role of LYST on Neotissue Formation in Tissue Engineered Constructs.” J.C.B. was supported by NIH T32AI106704, NIH F32HL144120, and the Ohio State University President's Postdoctoral Scholar Program. J.W.R. was supported, in part, by the American Heart Association under award no. 18POST33990231.

Supplementary Material

References

Matsuzaki

, John

, Shoji

, et al. The evolution of tissue engineered vascular graft technologies: From preclinical trials to advancing patient care. Appl Sci, 2019; 9(7):1274.

Mazza

, Al-Akkad

, Rombouts

, et al. Liver tissue engineering: From implantable tissue to whole organ engineering. Hepatol Commun, 2018; 2(2):131–141.

Bitar

, Raghavan

Intestinal tissue engineering: Current concepts and future vision of regenerative medicine in the gut: Functional intestinal tissue engineering. Neurogastroenterol Motil, 2012; 24(1):7–19.

Chaudhari

, Vig

, Baganizi

, et al. Future prospects for scaffolding methods and biomaterials in skin tissue engineering: A review. IJMS, 2016; 17(12):1974.

Wragg

, Burke

, Wilson

. A critical review of current progress in 3D kidney biomanufacturing: Advances, challenges, and recommendations. Ren Replace Ther, 2019; 5(1):18.

Drews

, Pepper

, Best

, et al. Spontaneous reversal of stenosis in tissue-engineered vascular grafts. Sci Transl Med, 2020; 12(537):eaax6919.

Zakko

, Blum

, Drews

, et al. Development of tissue engineered heart valves for percutaneous transcatheter delivery in a fetal ovine model. JACC Basic Transl Sci, 2020; 5(8):815–828.

Matsuzaki

, Miyamoto

, Miyachi

, et al. The evaluation of a tissue-engineered cardiac patch seeded with hips derived cardiac progenitor cells in a rat left ventricular model. Guo Y, editor. PLoS One, 2020; 15(6):e0234087.

Hibino

, McGillicuddy

, Matsumura

, et al. Late-term results of tissue-engineered vascular grafts in humans. J Thoracic Cardiovas Surg, 2010; 139(2):431.e2–436.e2.

10.

Roh

, Sawh-Martinez

, Brennan

, et al. Tissue-engineered vascular grafts transform into mature blood vessels via an inflammation-mediated process of vascular remodeling. Proc Natl Acad Sci U S A, 2010; 107(10):4669–4674.

11.

Patterson

, Gilliland

, Maxfield

, et al. Tissue-engineered vascular grafts for use in the treatment of congenital heart disease: From the bench to the clinic and back again. Regen Med, 2012; 7(3):409–419.

12.

Best

, Strouse

, Hor

, et al. Toward a patient-specific tissue engineered vascular graft. J Tissue Eng, 2018; 9:204173141876470.

13.

Denisenko

, Guo

, Jones

, et al. Systematic assessment of tissue dissociation and storage biases in single-cell and single-nucleus RNA-seq workflows. Genome Biol, 2020; 21(1):130.

14.

McDonald

, Shirali

, Aragón

, et al. Endothelial regeneration of large vessels is a biphasic process driven by local cells with distinct proliferative capacities. Cell Stem Cell, 2018; 23(2):210.e6–225.e6.

15.

Kalluri

, Vellarikkal

, Edelman

, et al. Single-cell analysis of the normal mouse aorta reveals functionally distinct endothelial cell populations. Circulation, 2019; 140(2):147–163.

16.

Tara

, Kurobe

, de Dios Ruiz Rosado

, et al. Cilostazol, not aspirin, prevents stenosis of bioresorbable vascular grafts in a venous model. ATVB, 2015; 35(9):2003–2010.

17.

Mirhaidari

GJM

, Barker

, Zbinden

, et al. Tissue engineered vascular graft recipient interleukin 10 status is critical for preventing thrombosis. Adv Healthcare Mater, 2020; 9(24):2001094.

18.

Roh

, Nelson

, Brennan

, et al. Small-diameter biodegradable scaffolds for functional vascular tissue engineering in the mouse model. Biomaterials, 2008; 29(10):1454–1463.

19.

Hao

, Hao

, Andersen-Nissen

, et al. Integrated analysis of multimodal single-cell data. Cell, 2021; 184(13):3573.e29–3587.e29.

20.

Stuart

, Butler

, Hoffman

, et al. Comprehensive integration of single-cell data. Cell, 2019; 177(7):1888–1902.e21.

21.

Butler

, Hoffman

, Smibert

, et al. Integrating single-cell transcriptomic data across different conditions, technologies, and species. Nat Biotechnol, 2018; 36(5):411–420.

22.

Satija

, Farrell

, Gennert

, et al. Spatial reconstruction of single-cell gene expression data. Nat Biotechnol, 2015; 33(5):495–502.

23.

Sánchez

, Brey

, Briceño

. Endothelialization mechanisms in vascular grafts. J Tissue Eng Regen Med 2018;12(11):2164–2178.

24.

Rajendran

, Rengarajan

, Thangavel

, et al. The vascular endothelium and human diseases. Int J Biol Sci, 2013; 9(10):1057–1069.

25.

Hurlimann

The relationship between the endothelium and the vessel wall. Eur Heart J Suppl, 2002; 4:A1–A7.

26.

Prevo

, Banerji

, Ferguson

DJP

, et al. Mouse LYVE-1 is an endocytic receptor for hyaluronan in lymphatic endothelium. J Biol Chem, 2001; 276(22):19420–19430.

27.

Podgrabinska

, Braun

, Velasco

, et al. Molecular characterization of lymphatic endothelial cells. Proc Natl Acad Sci U S A, 2002; 99(25):16069–16074.

28.

Pan

, Dinh

, Rajaraman

, et al. Patterns of expression of factor VIII and von Willebrand factor by endothelial cell subsets in vivo. Blood, 2016; 128(1):104–109.

29.

Xiang

, Hwa

Regulation of VWF expression, and secretion in health and disease. Curr Opin Hematol, 2016; 23(3):288–293.

30.

Reinhardt

, Rosado J de

, Barker

, et al. Early natural history of neotissue formation in tissue-engineered vascular grafts in a murine model. Regen Med, 2019; 14(5):389–408.

31.

Franz

, Rammelt

, Scharnweber

, et al. Immune responses to implants—A review of the implications for the design of immunomodulatory biomaterials. Biomaterials, 2011; 32(28):6692–6709.

32.

Saleh

, Bryant

. The host response in tissue engineering: Crosstalk between immune cells and cell-laden scaffolds. Curr Opin Biomed Eng 2018;6:58–65.

33.

Wang

, Song

, Wang

, et al. S100A8/A9 in inflammation. Front Immunol, 2018; 9:1298.

34.

Carnicer-Lombarte

, Chen

, Malliaras

, et al. Foreign body reaction to implanted biomaterials and its impact in nerve neuroprosthetics. Front Bioeng Biotechnol, 2021; 9:622524.

35.

Hibino

, Yi

, Duncan

, et al. A critical role for macrophages in neovessel formation and the development of stenosis in tissue-engineered vascular grafts. FASEB J, 2011; 25(12):4253–4263.

36.

Gordon

, Plüddemann

, Martinez Estrada

Macrophage heterogeneity in tissues: Phenotypic diversity and functions. Immunol Rev, 2014; 262(1):36–55.

37.

Ipseiz

, Pickering

, Rosas

, et al. Tissue-resident macrophages actively suppress IL-1beta release via a reactive prostanoid/IL-10 pathway. EMBO J [Internet], 2020; 39(14):e103454; doi: 10.15252/embj.2019103454

38.

Przybyl

, Haase

, Golic

, et al. CD74-downregulation of placental macrophage-trophoblastic interactions in preeclampsia. Circ Res, 2016; 119(1):55–68.

39.

Hibino

, Mejias

, Pietris

, et al. The innate immune system contributes to tissue-engineered vascular graft performance. FASEB J, 2015; 29(6):2431–2438.

40.

Havran

, Poenie

, Kimura

, et al. Expression and function of CD3-antigen receptor on murine CD4 + 8+ thymocytes. Nature, 1987; 300(12):170–173.

41.

Chetty

, Gatter

CD3: Structure, function, and role of immunostaining in clinical practice. J Pathol, 1994; 173(4):303–307.

42.

Torlakovic

, Torlakovic

, Nguyen

, et al. The value of anti-Pax-5 immunostaining in routinely fixed and paraffin-embedded sections: A novel pan pre-B and B-cell marker. Am J Surg Pathol, 2002; 26(10):1343–1350.

43.

Moore

, Maestas

, Cherry

, et al. Biomaterials direct functional B cell response in a material-specific manner. Sci Adv, 2021; 7(49):eabj5830.

44.

Stenmark

, Yeager

, El Kasmi

, et al. The adventitia: Essential regulator of vascular wall structure and function. Annu Rev Physiol, 2013; 75(1):23–47.

45.

Wang

, Rätsep

, Chapman

, et al. Adventitial fibroblasts in vascular structure and function: The role of oxidative stress and beyond. Can J Physiol Pharmacol, 2010; 88(3):177–186.

46.

Muhl

, Genové

, Leptidis

, et al. Single-cell analysis uncovers fibroblast heterogeneity and criteria for fibroblast and mural cell identification and discrimination. Nat Commun, 2020; 11(1):3953.

47.

Desmoulière

, Chaponnier

, Gabbiani

Tissue repair, contraction, and the myofibroblast. Wound Repair Regen 2005;13(1):7–12.

48.

Hinz

The role of myofibroblasts in wound healing. Curr Res Transl Med, 2016; 64(4):171–177.

49.

Hartlapp

, Abe

, Saeed

, et al. Fibrocytes induce an angiogenic phenotype in cultured endothelial cells and promote angiogenesis in vivo. FASEB J, 2001; 15(12):2215–2224.

50.

Mizoguchi

, Slowikowski

, Wei

, et al. Functionally distinct disease-associated fibroblast subsets in rheumatoid arthritis. Nat Commun, 2018; 9(1):789.

51.

Barth

, Westhoff

CD34+ fibrocytes: Morphology, histogenesis and function. CSCR, 2007; 2(3):221–227.

52.

Tang

, Wang

, Huang

, et al. Arterial Sca1+ vascular stem cells generate de novo smooth muscle for artery repair and regeneration. Cell Stem Cell, 2020; 26(1):81.e4–96.e4.

53.

Chasseigneaux

, Moraca

, Cochois-Guégan

, et al. Isolation and differential transcriptome of vascular smooth muscle cells and mid-capillary pericytes from the rat brain. Sci Rep, 2018; 8(1):12272.

54.

Yuan

SM.

α-smooth muscle actin and ACTA2 gene expressions in vasculopathies. Rev Brasil Cir Cardiovasc [Internet], 2015; 30(6):644–649; doi: 10.5935/1678-9741.20150081

55.

Sorokin

, Vickneson

, Kofidis

, et al. Role of vascular smooth muscle cell plasticity and interactions in vessel wall inflammation. Front Immunol, 2020; 11:599415.

56.

Hungerford

, Owens

, Argraves

, et al. Development of the aortic vessel wall as defined by vascular smooth muscle and extracellular matrix markers. Dev Biol, 1996; 178(2):375–392.

57.

Liu

, Jin

. Calponin isoforms CNN 1, CNN 2 and CNN 3: Regulators for actin cytoskeleton functions in smooth muscle and non-muscle cells. Gene 2016;585(1):143–153.

58.

Tang

, Barbacioru

, Wang

, et al. mRNA-Seq whole-transcriptome analysis of a single cell. Nat Methods, 2009; 6(5):377–382.

59.

de Dios Ruiz-Rosado

, Lee

, Mahler

, et al. Angiotensin II receptor I blockade prevents stenosis of tissue engineered vascular grafts. FASEB J, 2018; 32(12):6822–6832.

60.

Blum

, Roby

, Zbinden

, et al. Sex and Tamoxifen confound murine experimental studies in cardiovascular tissue engineering. Sci Rep, 2021; 11(1):8037.

61.

Tan

, Dharmadhikari

, Liu

, et al. Tracheal macrophages during regeneration and repair of long-segment airway defects. Laryngoscope, 2022; 132(4):737–746.

62.

Sobecki

, Mrouj

, Colinge

, et al. Cell-cycle regulation accounts for variability in Ki-67 expression levels. Cancer Res, 2017; 77(10):2722–2734.

63.

Heidebrecht

, Buck

, Haas

, et al. Monoclonal antibodies Ki-S3 and Ki-S5 yield new data on the ‘Ki-67’proteins. Cell Prolif, 1996; 29(7):413–425.

64.

Greaney

, Adams

, Brickman Raredon

, et al. Platform effects on regeneration by pulmonary basal cells as evaluated by single-cell RNA sequencing. Cell Rep, 2020; 30(12):4250.e6–4265.e6.

65.

Lam

, Keung

, Chan

, et al. Single-cell transcriptomics of engineered cardiac tissues from patient-specific induced pluripotent stem cell–derived cardiomyocytes reveals abnormal developmental trajectory and intrinsic contractile defects in hypoplastic right heart syndrome. JAHA, 2020; 9(20):e016528.

66.

Zhang

, Jiang

, Xie

, et al. Msx1+ stem cells recruited by bioactive tissue engineering graft for bone regeneration. Nat Commun, 2022; 13(1):5211.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.22 MB

0.41 MB

1.03 MB

0.50 MB

0.67 MB

0.00 MB