Tissue-engineered vascular grafts transform into mature blood vessels via an inflammation-mediated process of vascular remodeling

hBMC-Seeded Biodegradable Scaffolds Transform into Functional Blood Vessels in the SCID/bg Mouse.

The first objective of this study was to determine if the SCID/bg mouse could be used as an animal model to study human TEVG development. Specifically, we looked to see if similar results could be obtained in the SCID/bg mouse host as compared with human recipients (i.e., that hBMC-seeded scaffolds transformed into functional venous conduits in high-flow low-pressure settings). TEVGs were constructed under similar conditions to the clinical study protocol (6 –8) and were followed over a 24-wk course as IVC interposition grafts. All hBMC-seeded scaffolds (n = 28) remained fully patent and functional as venous conduits (Fig. 1A). At 24 wk, TEVGs (Fig. 1 C and F) morphologically and histologically resembled the native mouse IVC (Fig. 1 D and G). As early as 10 wk, ECs lined the inner lumen and were invested by a surrounding layer of SMCs (Fig. 1 H and I). By 24 wk, a mature vascular architecture consisting of a confluent EC-lined intima and one to two layers of SMC media was clearly defined in the neovessel (Fig. 1J). The original scaffold material had degraded by 24 wk and was effectively replaced by a supportive adventitial layer of collagen fibrils (Fig. 1 K and L). No elastin was detected in the TEVGs or native mouse IVC.

Stem Cells Constitute a Minor Fraction of Seeded hBMCs.

To test if seeded hBMCs were differentiating into the mature vascular cells of the engineered neovessel, we first examined the phenotypes and relative abundances of cell populations within the hBMC population used for seeding (Table S1). We specifically searched for mature ECs and SMCs and for progenitor cells capable of differentiating into mature vascular cells. hBMCs consisted mainly of mature leukocyte populations, including monocytes (10 ± 4.7%), CD4⁺ T cells (7.0 ± 2.7%), CD8⁺ T cells (7.9 ± 2.5%), B cells (6.4 ± 2.1%), and natural killer (NK) cells (3.2 ± 1.4%). A relatively high percentage of CD146⁺CD31⁺CD45⁻ mature bone marrow microvascular ECs (0.050 ± 0.024%) was also present. The largest population of adult stems cells identified was CD34⁺ hematopoietic stem cells (1.8 ± 0.53%). AC133⁺KDR⁺CD45⁻ proangiogenic cells made up 0.0029 ± 0.0042% of the hBMC population, whereas CD90⁺CD73⁺CD105⁺CD45⁻CD34⁻ mesenchymal stem cells constituted the smallest percentage of hBMCs, with yields of only 0.0013 ± 0.00096%.

Static seeding of hBMCs into scaffolds resulted in a similar distribution of cell types as in the original population (i.e., there was no obvious preference for any particular subpopulation). Before implantation, the majority of hBMCs in the scaffold were CD45⁺ mature leukocytes (Fig. 2A). CD31⁺ mature ECs and CD34⁺ stem cells were also identified but in much lower numbers (Fig. 2 B and C). No α-smooth muscle actin (α-SMA) expression was detectable before implantation.

Seeded hBMCs Are Not Incorporated into the Developing TEVGs.

Next, to determine if seeded hBMCs were directly contributing to the vascular cell composition of the developing neovessel, we tracked the seeded hBMCs using human-specific markers and immunohistochemistry. TEVGs were explanted at various time points over a 24-wk time course and stained for various hBMC surface antigens. CD34⁺ human stem cells could no longer be detected in any of the TEVGs after implantation. A small number of CD68⁺ human monocytes and CD31⁺ human ECs could be found within the scaffold wall at 1 wk but were not detectable after that time point (Fig. 2 D and E). Lack of seeded hBMC retention was confirmed with quantitative RT-PCR, demonstrating no detectable human RNA in the scaffolds after 1 wk in the mouse (Fig. 2I).

TEVGs Are Composed of Recruited Murine Host Cells.

Despite the diminishing presence of seeded hBMCs, overall cellularity of the TEVGs increased markedly during the first week of in vivo development (Fig. 3A). This increased cellularity stabilized after 1 wk and was primarily attributable to rapid infiltration of host mouse monocytes within the scaffold wall (Fig. 2F). Monocyte infiltration was subsequently followed by the influx of α-SMA-expressing cells around the inner luminal lining (Fig. 2G). By 3 wk, partial endothelialization along the inner lumen could be detected (Fig. 2H). By 24 wk, the scaffold had been replaced by a mature vascular structure (Fig. 1). Monocyte contribution to the cellularity of the TEVGs diminished concomitantly with scaffold degradation.

Seeded hBMCs Increase Early Monocyte Recruitment.

With direct evidence showing seeded hBMCs were not present after 1 wk in vivo, we next investigated if they had any effect on the vascular development of TEVGs by comparing the cellularity of hBMC-seeded scaffolds with that of unseeded scaffolds (Fig. 3A). At postimplantation days 1 and 3, no significant differences in cellularity could be detected between hBMC-seeded scaffolds (n = 6) and unseeded scaffolds (n = 4). However, by day 7, cellularity was significantly greater for hBMC-seeded scaffolds [n = 6; 270 ± 22 cells per high-power field (hpf)] compared with unseeded scaffolds (n = 4; 160 ± 40 cells per hpf; P < 0.001) (Fig. 3A). This increased cellularity was primarily attributable to a significantly increased infiltration of host mouse monocytes into hBMC-seeded scaffolds as compared with unseeded scaffolds (120 ± 20 monocytes per hpf vs. 60 ± 12 monocytes per hpf; P < 0.001) (Fig. 3B).

hBMCs Secrete Significant Amounts of MCP-1 When Seeded onto Scaffolds.

Based on these findings, we hypothesized that seeded hBMCs were functioning via a paracrine mechanism instead of directly differentiating into the vascular cells of the developing neovessel. To investigate this potential function, we examined the interaction hBMCs had with scaffold materials to determine if scaffold exposure activated hBMCs to secrete chemokines that could increase early monocyte recruitment. Scaffold exposure significantly increased the production of multiple cytokines by seeded hBMCs (Fig. 3 C–E). In particular, significantly high levels of MCP-1 were induced by hBMC seeding (Fig. 3 C and F), which reached quantities comparable to MCP-1 levels used in prior arteriogenesis studies (12, 13). Based on the levels obtained, MCP-1’s known function as a potent monocyte chemokine, and its established link to postnatal neovascularization (12 –15), we pursued further investigations into whether MCP-1 might be playing a role in TEVG development.

Local MCP-1 Secretion from Scaffolds Increases Early Monocyte Recruitment, Mimicking Seeded hBMC Effects on TEVG Development.

With in vivo data suggesting that seeded hBMCs increased early monocyte recruitment and in vitro studies indicating that hBMCs produced significant amounts of MCP-1 when seeded onto scaffold materials, we hypothesized that seeded hBMCs were increasing early monocyte recruitment via MCP-1 secretion. To test this hypothesis, we developed a system that enabled us to study the effects of secreted MCP-1, independent of other cytokines produced by seeded hBMCs. This was accomplished by creating hBMC analogues capable of only releasing MCP-1. To create these analogues, recombinant human MCP-1 was encapsulated into biodegradable alginate microparticles 1–20 μm in diameter, making them comparable in size to the heterogeneous population of hBMCs (Fig. 4A). The microparticles were then embedded into the scaffold to mimic the MCP-1 secretion by seeded hBMCs (Fig. 4 B and C). Embedded microparticles released ≈200 ng of MCP-1 from the scaffold over 72 h, which was similar to the duration of retention of seeded hBMCs in vivo (Fig. 4D).

Unseeded scaffolds embedded with MCP-1 microparticles developed and functioned similar to hBMC-seeded scaffolds when implanted as IVC interposition grafts in the SCID/bg mice. Host monocyte recruitment at 1 wk was significantly increased in MCP-1-eluting scaffolds (n = 3; 200 ± 60 monocytes per hpf) compared with both hBMC-seeded (n = 6; 120 ± 20 monocytes per hpf; P < 0.05) and unseeded (n = 4; 60 ± 12 monocytes per hpf; P < 0.05) scaffolds (Fig. 4E). Furthermore, after 10 wk in vivo, all MCP-1-eluting scaffolds (n = 6) were functioning as patent venous conduits and exhibited similar elements of vascular remodeling to hBMC-seeded scaffolds (n = 5). The MCP-1-eluting scaffold was fully infiltrated with host mouse monocytes. The internal lumen contained organizing vascular neotissue, consisting of an EC lining, SMC medial layer, and supportive collagen deposition (Fig. 4 F–K).

Biodegradable Polyglycolic Acid–Poly-L-Lactide and -ε-Caprolactone Scaffold Construction.

Scaffolds were constructed from a nonwoven polyglycolic acid (PGA) mesh (ConcordiaFibers) and a copolymer sealant solution of poly-L-lactide and -ε-caprolactone [P(CL/LA)] using previously described methods (11).

hBMC Isolation.

Unfractionated human bone marrow (10 donors) was purchased from Lonza. Bone marrow was diluted 1:1 in sterile PBS and filtered through a 100-μm nylon mesh. Mononuclear cells were isolated by density gradient centrifugation using Histopaque-1077 (Sigma).

Characterization of hBMCs.

Flow cytometry was used to identify and quantify subpopulations in the mononuclear cell fraction of human bone marrow. hBMCs from five different donors were used and tested in triplicate. Antibodies used were purchased from BD Biosciences [CD8 peridinin chlorophyll protein (PerCP; SK1), CD14 FITC (M5E2), CD31 FITC (WM59), CD34 phycoerythrin (PE) and PerCP-Cy5.5 (8G12), CD45 Allophycocyanin (APC)-Cy7 (2D1), CD56 PE (NCAM16.2), CD146 (P1H12), CD73 PE (AD2), CD90 FITC (5E10), and 7AAD], E-Bioscience [CD3 APC (HIT3a), CD4 FITC (L3T4), and CD19 (PE-Cy7)], Miltenyi Biotec [CD133 APC (AC133)], R&D Systems [VEGF-receptor 2 (R2) PE (89106)], and AbD Serotec [CD105 Alexa 647 (SN6)]. Cells were acquired on a FACSAria cell sorter (BD Bioscience), and results were analyzed using DIVA software (BD Bioscience).

hBMC Seeding onto PGA-P(CL/LA) Scaffolds.

A fibrin gel solution was used to attach hBMCs to the scaffold. hBMCs were suspended at 2 × 10⁷ cells per milliliter in sterile fibrinogen solution [100 mg/mL human fibrinogen (Sigma) in PBS]. Fifty microliters of BMC-fibrinogen solution (1 × 10⁶ BMCs) was statically seeded onto each scaffold. The cell solution was solidified onto the scaffold by adding sterile thrombin solution [100 U/mL human thrombin (Sigma) in 40 mM CaCl₂ in PBS].

Detection of Cytokine Secretion by hBMCs.

hBMCs were cultured at 2 × 10⁶ cells per milliliter on either PGA-P(CL/LA) scaffold or tissue culture plastic for 48 h in RPMI-1640 plus 10% (vol/vol) FBS. Cytokine levels were measured with Multiplex (Luminex) and ELISA (R&D Systems) kits per the manufacturers’ protocols.

Detection of Human RNA in TEVGs.

TEVGs harvested 1, 3, 7, or 21 days postimplantation or immediately before implantation were snap-frozen in liquid nitrogen, and RNA was extracted by mechanical crushing over dry ice followed by incubation in RLT lysis buffer (Qiagen). Samples were passed through Qiashredder columns and processed using RNeasy mini-kits according to the manufacturer's protocols (Qiagen). RT with random hexamer and oligo-dT primers was performed according to the Multiscribe RT system protocol (Applied Biosystems). PCR reactions were prepared with TaqMan 2× PCR Master Mix and predeveloped assay reagents from Applied Biosystems (human GAPDH, Hs99999905 mL; mouse hypoxanthine phosphoribosyl transferase , Mm00446968 mL). Species specificity of the human and mouse probes was confirmed on human and mouse control artery segments. To determine the limits of human RNA detection, standard curves were generated by measuring human GAPDH levels obtained from 10-fold serial dilutions of hBMCs in culture and hBMCs seeded onto scaffolds. The limit of detection of this assay was between 10 and 100 cells.

Infrarenal IVC Interposition Surgery.

TEVG implantations in the mice were performed as previously described (11). All scaffolds were sutured into the infrarenal IVC of 3–4-month-old female C.B-17 SCID/bg mice (Taconic Farms). A total of 55 animals were implanted with either hBMC-seeded (n = 28), unseeded (n = 18), or MCP-1-eluting (n = 9) scaffolds. At 1 day, 3 days, 1 wk, 3 wk, 6 wk, 10 wk, and 24 wk, animals were killed and grafts were explanted for analysis. All animal experiments were done in accordance with the institutional guidelines for the use and care of animals.

Micro-Computed Tomography Angiography.

In vivo patency and morphology of the TEVGs were evaluated using micro-computed tomography angiography, as previously described (11).

Histology.

Explanted grafts were fixed in 10% (vol/vol) formalin and embedded in paraffin or glycol methacrylate using previously published methods (11, 43). Sections were stained with H&E, Gomori trichrome, and Verhoeff–van Gieson stains.

Immunohistochemistry.

Species-specific antibodies were used to distinguish between human and mouse cells. Human-specific primary antibodies included mouse-anti-human CD31 (Dako), CD68 (Dako), CD34 (Abcam), and CD45 (AbD Serotec). Mouse-specific primary antibodies included rat-anti-mouse Mac-3 (BD Bioscience), F4/80 (AbD Serotec), and goat-anti-mouse VEGF-R2 (R&D Systems; minimal cross-reactivity). Primary antibodies that cross-reacted with both species were mouse-anti-human α-SMA (Dako), VEGF (Santa Cruz), and rabbit-anti-human von Willebrand factor (Dako). Antibody binding was detected using biotinylated secondary antibodies, followed by binding of streptavidin-HRP and color development with 3′,3-diaminobenzidine (Vector). Nuclei were counterstained with hematoxylin. For immunofluorescence detection, a goat-anti-rabbit IgG-Alexa Fluor 568 (Invitrogen) or a goat-anti-mouse IgG-Alexa Fluor 488 (Invitrogen) was used with subsequent 4′,6-diamidino-2-phenylindole nuclear counterstaining.

Quantitative Cellularity Analysis.

Relative cellularity was measured for each explanted scaffold. Two separate sections of each explant were stained with H&E and imaged at a magnification of ×400. The number of nuclei were counted in five regions of each section and averaged. Mouse monocytes, identified by positive F4/80 expression, were quantified in explanted scaffolds using similar methods.

Synthesis of MCP-1-Eluting Microparticles.

Recombinant human MCP-1 (R&D Systems) was encapsulated into biodegradable alginate microparticles using previously published methods (44). Microparticles were incorporated into the scaffold by directly mixing them into the P(CL/LA) sealant at 50 μg/μL. Scaffolds were constructed using methods described previously (11). Size and shape distributions of MCP-1 microparticles were determined by imaging with an XL-30 scanning electron microscope (FEI Company). The release profile of MCP-1 from these scaffolds was measured by ELISA (R&D Systems). MCP-1-eluting scaffolds were each immersed in 1 mL of M199 medium and incubated in a 37 °C orbital shaker. At 1, 2, 4, 8, 12, 24, 48, 72, 96, 120, 144, and 168 h, medium was collected and replaced with 1 mL of fresh M199. Samples were stored at −20 °C until analysis.

Statistical Analysis.

Statistical differences were measured using a Student's t test or ANOVA. P < 0.05 was considered statistically significant.