The innate immune system contributes to tissue-engineered vascular graft performance

Abstract

The first clinical trial of tissue-engineered vascular grafts (TEVGs) identified stenosis as the primary cause of graft failure. In this study, we aimed to elucidate the role of the host immune response in the development of stenosis using a murine model of TEVG implantation. We found that the C.B-17 wild-type (WT) mouse (control) undergoes a dramatic stenotic response, which is nearly completely abolished in the immunodeficient SCID/beige (bg) variant. SCID mice, which lack an adaptive immune system due to the absence of T and B lymphocytes, experienced rates of stenosis comparable to WT controls (average luminal diameter, WT: 0.071 ± 0.035 mm, SCID: 0.137 ± 0.032 mm, SCID/bg: 0.804 ± 0.039 mm; P < 0.001). The bg mutation is characterized by NK cell and platelet dysfunction, and systemic treatment of WT mice with either NK cell–neutralizing (anti–NK 1.1 antibody) or antiplatelet (aspirin/Plavix [clopidogrel bisulfate]; Asp/Pla) therapy achieved nearly half the patency observed in the SCID/bg mouse (NK Ab: 0.356 ± 0.151 mm, Asp/Pla: 0.452 ± 0.130 mm). Scaffold implantation elicited a blunted immune response in SCID/bg mice, as demonstrated by macrophage number and mRNA expression of proinflammatory cytokines in TEVG explants. Implicating the initial innate immune response as a critical factor in graft stenosis may provide a strategy for prognosis and therapy of second-generation TEVGs.—Hibino, N., Mejias, D., Pietris, N., Dean, E., Yi, T., Best, C., Shinoka, T., Breuer, C. The innate immune system contributes to tissue-engineered vascular graft performance.

Currently used prosthetic vascular grafts such as polytetrafluoroethylene lack growth capacity and are a significant source of postoperative complications. Prosthetic conduits are nonviable and do not possess the ability to grow, repair, and remodel. Through tissue engineering, it is possible to construct a vascular graft that has structural and biochemical characteristics similar to those of a native vessel (1). Tissue-engineered vascular grafts (TEVGs) can be created by seeding cells onto a biodegradable scaffold. The scaffold is a source of postimplantation biomechanical integrity, and it provides sites for cell attachment and neotissue formation (2). Over time, the scaffold degrades, and the resulting neovessel represents an autologous, living vascular conduit.

We have previously reported the midterm results of a human trial in which TEVGs created by seeding autologous bone marrow–derived mononuclear cells onto a tubular biodegradable scaffold fabricated from a polyglycolic acid fiber coated with a 50:50 copolymer of l-lactide and ε-caprolactone were implanted for the repair of single-ventricle congenital heart disease (3). Although early results were satisfactory, the principal mode of TEVG failure was graft stenosis (4). Previous studies have noted that seeding TEVGs with autologous cells promotes vascular neotissue formation and decreases the rate of graft failure (1, 5, 6). It is necessary to elucidate the biochemical mechanisms of vascular neotissue formation in TEVGs to develop strategies for inhibiting TEVG stenosis in second-generation constructs.

We recently reported our results using a murine model to investigate neovessel formation, the process by which a scaffold seeded with bone marrow mononuclear cells transforms into a living vascular conduit with the ability to grow, repair, and remodel (7). In this study, we demonstrated that neovessel formation is an inflammatory cell–derived regenerative process. In the present study, we found that an exuberant inflammatory response could be a cause of stenosis. Advances in characterization and manipulation of mouse genetics make the murine host a good model for the study of these processes on a molecular level. In this study, we addressed the role of host immune function for excess neotissue formation in TEVGs using an immunodeficient mouse model.

MATERIALS AND METHODS

Scaffold fabrication

Scaffolds 0.8 mm in diameter were constructed from a nonwoven polyglycolic acid mesh (Concordia Fibers, Coventry, RI, USA) and a copolymer sealant solution of poly-l-lactide co-ε-caprolactone using previously described methods (8).

TEVG implantation

TEVG implantations were performed using microsurgical techniques. The scaffolds were inserted into the infrarenal inferior vena cava (IVC) of 3- to-4-month-old, C.B-17-background wild-type (WT), SCID/beige mutation (bg), and SCID female mice (Taconic, Germantown, NY, USA). A total of 72 animals were implanted with TEVGs. All animal experiments were done in accordance with the institutional guidelines for the use and care of animals, and the institutional review board approved the experimental procedures described.

Histology

Explanted grafts were pressure-fixed in 10% formalin overnight and then embedded in paraffin or glycol methacrylate using previously published methods (7). Hematoxylin and eosin (H&E; Sigma-Aldrich, St. Louis, MO, USA) staining was conducted to evaluate graft morphometry. Graft luminal diameters were measured using ImageJ software (Image Processing and Analysis in Java; National Institutes of Health, Bethesda, MD, USA).

Immunohistochemistry

Primary antibodies included rat anti-mouse F4/80 (AbD Serotec, Kidlington, UK), mouse anti-human α-smooth muscle actin (α-SMA) (Dako, Carpinteria, CA, USA), rabbit anti-human von Willebrand factor (Dako), and mouse anti-mouse NK 1.1 (Abcam, Cambridge, MA, USA). Antibody binding was detected using appropriate biotinylated secondary antibodies, followed by binding of streptavidin–horseradish peroxidase and color development with 3,3-diaminobenzidine (Vector Laboratories, Burlingame, CA, USA). Nuclei were then counterstained with hematoxylin (Gills formula; Vector). For immunofluorescence detection, a goat anti-rabbit IgG-Alexa Fluor 568 (Invitrogen, Carlsbad, CA, USA) or a goat anti-mouse IgG-Alexa Fluor 488 (Invitrogen) was used with subsequent 4′,6-diamidino-2-phenylindole nuclear counterstaining.

Ultrasonography

Serial ultrasonography (Vevo Visualsonics 770, Toronto, Canada) was performed for surveillance of the TEVG. Before ultrasonography, mice were anesthetized with 1.5% inhaled isoflurane. Graft luminal diameter was determined sonographically at the indicated time points after implantation.

Microcomputed tomography angiography

In vivo patency and morphology of the TEVGs were evaluated using microcomputed tomography (μCT) angiography, as described previously (8).

Computer-assisted image analysis (quantitative immunohistochemistry)

Mouse macrophages, identified by positive F4/80 expression, were measured for each explanted scaffold. Two separate sections of each explant were counterstained with hematoxylin and imaged at ×400 magnification. The number of nuclei was then counted in 5 regions of each section and averaged.

Quantitative RT-PCR

Explanted tissue grafts were frozen in Tissue-Tek OCT (Sakura, Torrance, CA, USA) and were each sectioned into forty 10 μm sections using a Cryocut 1800 (Leica, Buffalo, NY, USA). Excess Tissue-Tek OCT was removed by centrifugation in water. RNA extraction was performed with an RNeasy Mini Kit (Qiagen VWR, Stockholm, Sweden), according to the manufacturer’s protocol. RNA concentration of each sample was determined by Quant-iT RiboGreen RNA Reagent and Kit (Invitrogen), according to the manufacturer’s protocol. Real-time RT-PCR was carried out with a 50 µl total volume containing 1 µg RNA, 5 µl 10× Taq-Man RT buffer, 11 µl MgCl₂ (25 mM), 10 µl dNTP mixture (10 mM), 2.5 µl Random Hexamer (50 mM), 1 µl RNase inhibitor (20 U/µl), and 1.25 µl reverse transcriptase (50 U/µl) (Applied Biosystems, Foster City, CA, USA). RNAse-free water was added to bring total sample volume to 50 µl. The thermal cycling parameters were incubation at 25°C for 10 minutes, reverse transcription at 48°C for 30 minutes, and inactivation at 95°C for 5 minutes. Predesigned and validated gene-specific TaqMan Gene Expression Assays from Applied Biosystems were used in duplicate for quantitative reverse transcriptase PCR (qRT-PCR) according to the manufacturer’s protocol. Primers for the following genes were used: Emr1 (Mm00802529_m1), chemokine (C-C motif) ligand 3 (CCL3) (Mm00441258_m1), iNOS (Mm00441258_m1), TNF-α (Mm00443258_m1), CX3CR1 (Mm00438354_m1), found in inflammatory zone 1 (Fizz1) (Mm00445110_g1), and matrix metalloproteinase 9 (MMP9) (Mm00600157_g1). Values were normalized to expression of hypoxanthine-guanine phosphoribosyltransferase (HPRT) (Mm00441258_m1).

Platelet inhibition

WT mice were treated with aspirin and Plavix (clopidogrel bisulfate; Bristol-Myers Squibb, Princeton, NJ, USA) for 10 weeks after TEVG implantation. Aspirin (30 mg/L) was administered via drinking water, which was replaced with fresh water every other day. Clopidogrel bisulfate (20 mg/kg) was started immediately after transplantation and injected intraperitoneally. These mice were humanely killed at the end of the 10-week treatment period, and the implanted scaffolds were fixed for histologic examination as above.

NK cell inhibition

WT mice were treated with 200 μl anti–NK 1.1 antibody (Ebioscience, San Diego, CA, USA) at 2 days and 24 hours before TEVG implantation and then weekly thereafter for 10 weeks. This dose of NK antibody was used to sustain depletion of NK cells from the spleen without reduction of cellular or humoral immunity. As above, these mice were humanely killed and the grafts collected for histologic analysis.

RESULTS

We have previously demonstrated the feasibility of creating sub–1 mm biodegradable tubular scaffolds that are functional vascular grafts (8). Here, we investigated the natural course of these grafts using the same mouse IVC implantation model. TEVGs implanted into SCID/bg mice were followed for 24 weeks. Histologic examination of the vessels at 1 day and at 1, 3, 6, 10, and 24 weeks (n = 4 for each time point) demonstrated progressive infiltration of the scaffold by macrophages, degradation of the polymer, and formation of a laminated neovessel (Fig. 1A). Up to 3 weeks after implantation, the wall thickness increased and luminal diameter decreased due to cell infiltration into the graft. Thereafter, the degradation of scaffold led to decreased wall thickness and increased luminal diameter, and both values closely approximated native IVC at 24 weeks (Fig. 1B). Immunofluorescence staining demonstrated von Willebrand factor–positive endothelial cells and an α-SMA-positive cell layer in the inner layer of the graft at 10 weeks after implantation (Fig. 1C). This staining confirmed the presence of a confluent endothelialized intima and a medial smooth muscle layer at the 10 week time point, suggesting that this was an appropriate end point to evaluate neotissue formation in our model.

Figure 1.

Characterizing neovessel formation in SCID/bg mice. A) H&E staining demonstrating infiltration of the scaffold wall with macrophages, and the serial degradation of the scaffold and neovessel formation. Scale bars, 200 μm. B) Luminal diameter and wall thickness changes in the neovessels over time from serial ultrasound monitoring. C) Immunofluorescence of mature neovessels showing positive staining for endothelial and smooth muscle cells. Scale bars, 200 μm.

To investigate the effect of host immune function on TEVG neotissue formation, we implanted 8 unseeded scaffolds into C.B-17 SCID/bg mice and 8 unseeded scaffolds into WT C.B-17 controls. Histologic examination at 10 weeks confirmed the development of stenosed grafts in the WT mice with severe intimal thickening (Fig. 2A). Sonography revealed stenosis of the grafts implanted in WT mice after 2 weeks, whereas the grafts of the SCID/bg group remained patent up to 10 weeks (Fig. 2B). All of the SCID/bg grafts, in contrast, remained patent and demonstrated significantly higher inner and outer diameters compared with the WT group (average luminal diameter: WT: 0.071 ± 0.035 mm, SCID/bg: 0.804 ± 0.039 mm; P < 0.001) (Fig. 2C, D).

Figure 2.

SCID/bg mice show markedly increased patency rates vs. WT controls. A) H&E stain of TEVG cross sections from WT and C.B-17 SCID/bg mice. Scale bars, 200 μm. B) Luminal diameters of TEVGs implanted in SCID/bg mice are higher than WT mice at 2, 4, 6, 8, and 10 weeks after implantation from serial ultrasound monitoring. WT scaffolds show uniform occlusion at 8 weeks after implantation. SCID/bg mice show increased luminal (C) and adventitial (D) diameters after 10 weeks vs. WT controls.

The SCID/bg phenotype is due to defects in T- and B-cell function (SCID mutation) and NK cell and platelet dysfunction (bg mutation). To ascertain whether defects in these cell lines could be independently responsible for decreased intimal hyperplasia, we implanted scaffolds into mice that bore only the SCID mutation (SCID) (n = 8), as well as into WT C.B-17 mice that were treated with either an NK-cell depleting antibody (NK Ab) (n = 6), or with platelet-inhibiting aspirin and clopidogrel bisulfate (Asp/Pla) (n = 6). Ultrasonography demonstrated a difference in luminal diameter at 2 weeks after implantation (Fig. 3). The SCID mice developed graft stenosis at a rate equivalent to WT mice, while each of the treated groups exhibited luminal diameters that were halfway between SCID/bg mice and the WT group (WT: 0.071 ± 0.035 mm, SCID/bg: 0.804 ± 0.039 mm, SCID: 0.137 ± 0.032 mm, Asp/Pla: 0.452 ± 0.130 mm, NK Ab: 0.356 ± 0.151 mm; P < 0.001) (Fig. 3A, B). The remodeling of TEVG scaffolds into patent neovessels was confirmed in the SCID/bg and Asp/Pla groups by μCT angiography (Fig. 3C). Similarly, we identified the development of collateral vessels in cases of stenotic TEVGs in both the WT and SCID groups (Fig. 3C).

Figure 3.

A) H&E staining showing complete occlusion of TEVGs implanted into a SCID mouse and patent grafts implanted into mice treated with Asp/Pla or NK Ab. Scale bars, 200 μm. B) SCID mice developed graft stenosis at a rate equivalent to WT mice, while each of the treated groups exhibited luminal diameters that were halfway between SCID/bg mice and the WT group. C) μCT angiography of implanted TEVGs (ovals) demonstrating patent neovessels in the SCID/bg and Asp/Pla groups and the coincidence of collateral formation (arrows) in cases of stenosed TEVGs in the WT and SCID groups.

To further investigate the role of NK cells in intimal hyperplasia, we performed immunohistochemical (IHC) analysis of the stenotic WT grafts as well as the patent SCID/bg grafts using the anti-NK antibody NK 1.1 (Fig. 4A). In contrast to patent grafts, the stenosed grafts stained positive for NK cells, and in grafts that showed evidence of hyperplasia without complete stenosis, NK 1.1–positive cells were observed in the hyperplastic region (Fig. 4B, C).

Figure 4.

A) SCID/bg mice show no positive staining with NK 1.1. B, C) IHC staining with NK 1.1 antibody showing that NK cells are present at sites of neointimal hyperplasia (A, B, scale bars, 200 μm; C, scale bar, 50 μm).

Although SCID/bg mice do not typically exhibit a pronounced defect in the innate immune response, we nonetheless investigated the difference in macrophage recruitment between the SCID/bg mice (n = 8) and WT C.B-17 (n = 8) mice through semiquantitative assessment of the degree of macrophage infiltration. The SCID/bg mice showed significantly fewer macrophages per high-powered field (WT: 113 ± 12 /HPF, SCID/bg: 66 ± 18/HPF; P = 0.006) (Fig. 5A–C). This result is consistent with the result of PCR analysis of scaffolds at 3 days, which showed significantly less F4/80 mRNA production in SCID/bg grafts compared with WT explants (Fig. 5D). Although F4/80 expression does not differentiate between resident tissue macrophages and monocyte-derived macrophages (9), our previous report that F4/80⁺ macrophages found in the TEVGs at early time points originate in the bone marrow suggests that these macrophages are monocyte derived and therefore are a component of the innate immune response to TEVG implantation (10). To further investigate modulation of the macrophage response, we performed qRT-PCR to assess relative mRNA activity between the 2 groups at 3 and 28 days after implantation (n = 3 for each group, each time point) (Fig. 6). The expression of cytokines associated with the acute inflammatory response such as CCL3, iNOS, and TNF-α was higher in WT compared with SCID/bg mice at 3 days after implantation. At the 28-day time point, these inflammatory markers declined sharply in the WT group, while the levels in the SCID/bg mice remained constant or showed only modest declines (Fig. 6A). This profile suggests a blunted acute inflammatory response in the SCID/bg mice compared with WT. On the other hand, the expression of cytokines associated with later remodeling processes such as CX3CL1 (chemokine [C-X3-C motif] ligand 1), FIZZ1, and MMP9 were higher at 28 days than at 3 days in both groups, and the level of expression was comparable in both groups (Fig. 6B).

Figure 5.

IHC of WT C.B-17 mice (A) demonstrating more robust macrophage infiltration than in SCID/bg mice (B). C) SCID/bg mice show fewer macrophages than WT mice upon semiquantitative analysis. qRT-PCR demonstrating less F4/80 mRNA production in SCID-bg mice (D). Scale bar, 50 μm.

Figure 6.

The cytokines expressed by M1 macrophages such as CCL3, iNOS, and TNF-α (A) were higher in WT compared with SCID/bg (S/b) mice at only 3 days (d) after implantation. On the other hand, expression of cytokines associated with M2 macrophages such as CX3CL1, FIZZ1, and MMP9 (B) were increased at 28 days compared with 3 days in both groups, and the level of expression was comparable in both groups.

DISCUSSION

The results of this study highlight the critical role that the immune system plays in TEVG stenosis. We have recently proposed a mechanism for the natural history of neotissue formation in bone marrow mononuclear cell–seeded scaffolds in SCID/bg mice and surmised that it is an immune-mediated process. Therefore, we asserted that differences in host immune function could affect neotissue formation of TEVGs.

In previous studies, the magnitude of the inflammation that occurs after vessel injury appears to be linked to the subsequent intimal hyperplastic response. Sustained high levels of systemic inflammation markers such as C-reactive protein predict increased risk of human coronary artery stenosis after stenting (11). The number of monocytes recruited to the vessel wall has also been directly correlated with the volume of neointima formation (12). Studies have shown that monocyte chemoattractant protein 1 (MCP-1) is an essential chemokine in this process, and that blockade of MCP-1 abrogates both macrophage recruitment and intimal hyperplasia (13–15).

Inflammation has thus clearly been implicated in vascular hyperplasia in response to vessel injury. In the context of TEVG formation, inflammation is not purely pathogenic but in fact critical for neotissue generation, yet an overly exuberant inflammatory response causes excessive recruitment of smooth muscle cells resulting in intimal thickening (10). On the basis of the results of the present study, we propose that SCID/bg mice, by virtue of their genetic derangements, appear to have this near-optimum internal milieu. Though we hypothesized that an immunodeficient mouse would enjoy greater TEVG patency rates than an immunocompetent control, the contrast between the SCID/bg and WT mice is startlingly pronounced (Fig. 2A). WT mice develop uniformly stenosed grafts, and the immunodeficient mutants develop uniformly patent, though fully cellularized and mature, neovessels (Figs. 1A, 2). Additionally, the exuberant stenosis exhibited in WT C.B-17 mice is beyond that observed in humans, or indeed in other animal models (3). The exaggerated responses of this mouse species and that of its immunodeficient variant lie on the extreme ends of the spectrum of intimal hyperplasia in TEVGs. They provide a unique window through which the mechanisms that underlie this process can be viewed.

SCID/bg mice have not only widespread immune cell deficiency but also platelet dysfunction. The bg mutation is characterized mainly as a deficiency of granular cells with decreased antimicrobial efficacy of granulocytes (16), severe NK cell deficiencies (17), and defective platelet function (18, 19).

In investigating what characteristics make the SCID/bg mouse resistant to graft stenosis, we have attempted to separate the component cell line defects present. To this end, we made particular immune deficiency models: 1) SCID-only mice that lack an adaptive immune response due to defective T and B cells, 2) mice treated with an NK-neutralizing antibody to mimic NK cell defects, and 3) mice treated with antiplatelet agents aspirin and Plavix to imitate platelet dysfunction.

In this study, T and B cells appeared to play no role in abrogating TEVG intimal hyperplasia. This is in line with studies done in cyclosporine-suppressed rabbits (20), where suppression of humoral immunity had no effect on vessel stenosis (21). However, our data are at odds with results from studies in T-cell-depleted rats and Rag1^−/− knockout mice reconstituted with WT T and B lymphocytes in which these cells were shown to suppress neointimal hyperplasia (22–24). The mixed evidence of the role played by the adaptive immune system may be due to species-specific differences, as well as differences in the method of diminishing the adaptive immune response. Nonetheless, in our study, SCID mice demonstrated uniform occlusion of the graft, with no difference between these and WT C.B-17 mice.

The inhibition of platelets and NK cells, in contrast, appeared to individually reduce stenosis rates to half of the rates expected of a SCID/bg mouse (Fig. 3). Previous studies have implicated platelet adherence to damaged endothelium and subsequent macrophage recruitment as being a critical factor in the formation of a hyperplastic intima. Platelet P-selectin expression is necessary for triggering monocyte arrest and adhesion, and P-selectin^−/− knockout mice demonstrate markedly decreased macrophage infiltration, as well as a subsequent decrease in intimal hyperplasia (25, 26). Direct blockade of P-selectin binding by sialyl-Lewis^X substrate replicates this response (27). Similarly, NK cells have been demonstrated to promote arteriogenesis in a murine hind limb ischemia model (28), and NK-cell-deficient mice show decreased rates of intimal hyperplasia and vascular remodeling in a mouse vein graft restenosis model (29).

The presence of large numbers of macrophages is one of the hallmarks of scaffolds implanted into the mouse IVC. Macrophages play a variety of roles in the process of tissue repair and generation, orchestrating both the acute inflammatory response whereby pathogens and cellular debris are cleared, as well as the later reparative response where cells are recruited to the site of injury and subsequently remodeled. Recently, strategies for categorizing the populations of macrophages that perform these distinct functions have been proposed (30–32), although it is recognized they can assume a spectrum of activation phenotypes (33). Our results demonstrated that proinflammatory cytokines, such as CCL3, iNOS, and TNF-α, were higher in the WT compared with SCID/bg group, with a concomitant loss of patency (Fig. 6). Although the relatively small panel of markers used to characterize TEVG explants cannot identify the molecular mechanism of monocyte and macrophage-mediated neovessel formation, it is sufficient to distinguish between inflammatory and regenerative processes occurring in the TEVG. Our results suggest that the excessive acute phase inflammation is an important factor in stenosis of TEVGs. These findings are consistent with our previous work investigating the role of macrophages on the formation of TEVG stenosis (10). Studies in murine atherosclerosis models have also reported accumulation of inflammation-promoting macrophages in stenotic areas of vessels. Although the mechanism of stenosis in our graft is likely different from atherosclerosis, these studies support the contribution of the inflammatory response toward vessel stenosis (34). In contrast, the cytokines associated with tissue remodeling were expressed in the chronic phase, and expression levels were comparable between the WT and SCID/bg groups.

We acknowledge that the findings presented here are observational. By dissecting the defective cell lines comprising the SCID/bg phenotype, we have identified the innate immune system as a key regulator of TEVG stenosis; however, further research is required to define the precise mechanisms by which platelets, NK cells, and monocytes/macrophages orchestrate TEVG remodeling. It is the initial burst of inflammatory activity that correlates with stenotic grafts in the WT C.B-17 mice, whether or not this is as a direct result of impaired platelet and NK cell activity. Our parallel study that shows a causative role of macrophage infiltration in neotissue formation and intimal hyperplasia is consistent with this. Further characterization of the molecular determinants of this initial response could, at the very least, provide a measurable predictor of graft patency. Furthermore, this has implications for the rational design of second-generation TEVGs.

Glossary

Asp/Pla

aspirin/Plavix (clopidogrel bisulfate)

bg

beige mutation

CCL3

chemokine (C-C motif) ligand 3

CX3CL1

chemokine (C-X3-C motif) ligand 1

FIZZ1

found in inflammatory zone 1

H&E

hematoxylin and eosin

IHC

immunohistochemistry

IVC

inferior vena cava

MCP-1

monocyte chemoattractant protein 1

MMP9

matrix metalloproteinase 9

NK Ab

NK-cell depleting antibody

qRT-PCR

quantitative reverse transcriptase PCR

TEVG

tissue-engineered vascular graft

WT

wild-type

α-SMA

α-smooth muscle actin

μCT

microcomputed tomography