| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J Thorac Cardiovasc Surg 1995;110:988-997
© 1995 Mosby, Inc.
SURGERY FOR ACQUIRED HEART DISEASE |
Osaka and Nagano, Japan
Received for publication Sept. 30, 1994. Accepted for publication Dec. 23, 1994. Address for reprints: Takehisa Matsuda, PhD, Department of Bioengineering, NationalCardiovascular Center Research Institute, 5-7-1 Fujishirodai, Suita, Osaka 565, Japan.
Abstract
A layered-structure hybrid vascular graft, mimicking the hierarchic structure of the intima and media of a natural artery, is expected to exhibit antithrombogenicity and to accelerate neoarterial tissue formation. Two models of hybrid vascular grafts were prepared on knitted Dacron fabric grafts (inner diameter 4 mm, length 6 cm). Model I grafts consisted of an endothelial cell monolayer formed on collagenous matrix, and model II grafts consisted of an endothelial cell monolayer that formed on hybrid collagenous medial tissue in which smooth muscle cells were incorporated. Both models (n = 17 for each model) were implanted bilaterally in carotid arteries and left in place for up to 26 weeks. Although all of the implanted grafts were patent, the two models significantly differed in the degree of maturity of the regenerated neoarterial wall especially at earlier implantation periods. At 2 weeks, model II grafts showed a much higher degree of neointimal integrity than model I grafts: a smooth and organized neointimal layer was formed, which was free from leukocyte adhesion. On further increase of implantation time, the formation of neomedia (subendothelial smooth muscle cell layers) and circumferential orientation of both smooth muscle cells and collagenous extracellular matrix were much more advanced in model II grafts than in model I grafts. At 26 weeks after implantation, layered elastic laminae regenerated along circumferentially oriented smooth muscle cells in neomedia were observed only in model II grafts. Irrespective of model, little excessive smooth muscle cell proliferation occurred. A hierarchically structured hybrid graft eventually provided a more integrated neointimal layer and accelerated neoarterial tissue formation much more than model I grafts. The significance of the incorporation of smooth muscle cells into hybrid grafts is discussed. (J THORAC CARDIOVASC SURG 1995;110:988-97)
A natural artery has a hierarchically layered structure that consists of the intima, the media, and the adventitia. These layers contain endothelial cells (ECs), smooth muscle cells (SMCs), and fibroblasts, respectively. Homeostasis and physiologic function of the artery are maintained via intercellular regulation of these cells. It has been reported that EC seeding onto vascular grafts leading to the formation of a fully covered EC layer provides a nonthrombogenic potential similar to that of natural arteries. 
1,2 Although endothelialization techniques have been improved in the preparation of EC-seeded hybrid grafts, it has been reported that EC delamination in the acute phase of implantation is still problematic. 3 Once EC-seeded grafts enter the long-term phase after subendothelial layers are regenerated by ingrown SMCs, an EC monolayer with high hydrodynamic stability is attained in a dynamic environment.
If intercellular communication between ECs and SMCs operates to enhance the stability of the EC monolayer, a hybrid vascular graft biomimicking the hierarchic structure of an artery may provide intimal stability and concomitantly exhibit antithrombogenicity in the acute phase of implantation. Moreover, the incorporation of SMCs into hybrid grafts may alter the tissue regeneration process in a positive or negative manner as follows. Because SMCs secrete a variety of extracellular matrices such as collagen and elastin, it is expected that SMC-incorporated hybrid grafts may accelerate tissue regeneration. If incorporated SMCs, which are in the synthetic phenotype on in vitro culture, maintain this phenotype in vivo, neoarterial thickening resulting from excessive proliferation of SMCs occurs. This eventually leads to narrowing of the graft lumen, which results in occlusion in the long-term phase. In our previous study, 4 in vitro reconstruction of a hierarchically structured hybrid vascular graft that had an EC monolayer and SMC multilayers was reported.
In this study, we prepared EC-seeded hybrid grafts (model I grafts) and hybrid grafts with a layered structure of ECs and SMCs (model II grafts) and implanted these hybrid grafts in dogs for up to 26 weeks. The time course of the neoarterial formation process of implanted grafts was studied by histologic evaluation. In the comparison of these two models, this article focuses on (1) the stability of the EC monolayer in the acute phase of implantation and (2) the regeneration process of the vascular tissues in the long-term phase, with special emphasis placed on the significance of incorporation of SMCs into a hybrid graft.
MATERIALS AND METHODS
Harvest and culture of vascular cells
Eighteen adult mongrel dogs (weight range 25 to 35 kg) were used in this study in compliance with the "Principles of Laboratory Animal Care" (formulated by the National Society for Medical Research) and the "Guide for the Care and Use of Laboratory Animals" (NIH Publication No. 86-23, revised 1985). Animals were anesthetized with pentobarbital sodium (25 mg/kg) and intubated and the lungs were mechanically ventilated by means of a constant-volume ventilator. After systemic anticoagulation with intravenous heparin (100 U/kg) an 8 cm segment of the unilateral external jugular vein was excised.
ECs were harvested from the vein by the method of Jaffe 5 with some modifications. Briefly, the vein was filled with 0.1% collagenase (Collagenase N-2, Nitta Gelatin, Osaka, Japan) in Hanks' balanced salt solution (with calcium and magnesium; Nissui, Tokyo, Japan) for 14 minutes at 37° C, and harvested ECs were resuspended in 1.5 ml medium 199 (M199; Gibco, Grand Island, N.Y.) 6 supplemented with 20% heat-inactivated fetal bovine serum (FBS; HyClone, Logan, Utah), 25 ng/ml basic fibroblast growth factor (R&D, Minneapolis, Minn.), 25 µg/ml heparin (Wako, Osaka, Japan), 50 U/ml penicillin, 50 µg/ml streptomycin, and 2.5 µg/ml amphotericin (Fungizone; Flow Laboratories, Irvine, Scotland). ECs were then inoculated onto a 2 cm2 collagen-coated tissue culture dish (Corning Glass, Corning, N.Y.). SMCs were harvested from the media of the external jugular vein by the method of Ross 6 with some modifications. The media was cut into small fragments, placed onto culture dishes, and immersed in M199 supplemented with 10% FBS.
Each dish with ECs or SMCs was incubated at 37° C in a 5% carbon dioxide atmosphere and the medium was changed every other day. When confluency was reached, cells were subcultured several times at a ratio of 1:4, yielding four confluent 56 mm 2 dishes each with ECs or SMCs. The identification of ECs was made on the basis of their appearance (cobblestone morphologic features) at confluence observed by a phase-contrast microscope and also with the use of a fluorescent acetylated low-density lipoprotein (Biomedical Technologies, Stoughton, Mass.) observed under a fluorescent microscope. To identify SMCs, immunohistochemical staining with muscle-actin-specific monoclonal antibody (HHF35, Enzo Biochem, New York, N.Y.) 7 was done. A cell count was done on an aliquot of the cell suspension (model ZB1, Coulter Electronics, Hialeah, Fla.).
In vitro reconstruction of vascular wall
Knitted Dacron fabric vascular prostheses (Microknit, Golaski Laboratories, Inc., Philadelphia, Pa.; innerdiameter 4 mm, length 7 cm, water porosity 4000 ml/cm 2 per minute) were used. We previously reported that the complex gel of type I collagen and dermatan sulfate (weight ratio of 100:1) exhibited enhanced adhesion and growth of ECs and reduced adhesion of platelets. 8 We used this complex gel as artificial extracellular matrices for both ECs and SMCs in this study. Two models of hybrid grafts were constructed. The model II graft was prepared as follows. A cold mixed solution containing 6 x 10 6 cultured SMCs, 6 ml 0.5% acid-solved type I collagen (derived from bovine skin; Koken Corp., Tokyo, Japan), 0.3 mg dermatan sulfate (Seikagaku Ind., Tokyo, Japan), and 6 ml M199 was prepared. The graft, mounted on a specially designed stainless steel holder, was filled with 4 ml of the cold mixed solution under pressure and rotated around its longitudinal axis at 600 rpm for 10 minutes by a specially designed tubular spin coater. The residual solution in the graft was washed with Hanks' balanced salt solution. Three repetitions of this procedure and subsequent incubation at 37°C for 10 minutes enabled the formation of the hybrid media composed of multilayered SMCs embedded in a complex gel of type I collagen and dermatan sulfate.
Cultured ECs were then seeded onto the hybrid media-prelined graft as follows. After 1 ml of EC suspension in M199 supplemented with 20% FBS was poured into the hybrid media-prelined graft, the graft was allowed to lie horizontally and incubated at 37° C for 20 minutes. The graft was then rotated 90 degrees around its longitudinal axis and a new EC suspension was poured into it. This process was repeated four times to seed the ECs at a density of 6.6 x10 5 cells/cm 2 on the graft luminal surface. The seeded grafts were maintained in M199 supplemented with 20% FBS, 25 ng/ml fibroblast growth factor, and 25 µg/ml heparin for an additional 3 days before implantation.
The model I graft was prepared according to the same procedure as described, but without the incorporation of cultured SMCs into the complex gel.
Graft implantation
After systemic anticoagulation with intravenous heparin (200 U/kg) was achieved in the dog, both models of grafts (length 6 cm) were implanted bilaterally with the use of vertical end-to-end anastomoses with 6-0 polypropylene sutures in the carotid arteries of the same dog from which the vascular cells were harvested. The remaining 1 cm of both grafts was subjected to histologic examination of preimplanted grafts. Neither anticoagulant nor antiplatelet agents were administered, except for the intraoperative heparin.
Microscopic evaluation
When the projected implantation period was over, animals were subjected to systemic anticoagulation with intravenous heparin (200 U/kg), and the artery proximal and distal to each graft was cannulated. After in situ fixation was done with 2% paraformaldehyde, 0.5% glutaraldehyde, and 0.1% tannic acid in 0.1 mol/L sodium cacodylate, pH 7.4, at 100 mm Hg for 1 hour, grafts with their local arterial supplies were removed and further fixed by immersion in the same fixative for 8 hours. Samples for light microscopic evaluation were stained with hematoxylin-eosin stain, Masson's trichrome stain, and orcein (elastin) stain. To identify cell types in the graft, immunohistochemical staining was done. Specimens were stained by the biotin-avidin-peroxidase method with antibody against factor VIII–related antigen (Dako Laboratories, catalog no. A082, Glostrup, Denmark) for ECs and by HHF35 for SMCs. Samples for scanning electron microscopy, which were lyophilized and subsequently sputter-coated with platinum, were evaluated with a scanning electron microscope (S-4000, Hitachi, Tokyo, Japan).
EC coverage
EC coverage was measured for preimplanted and implanted grafts. The sampling sites, 10 sites in each graft, were designated before the planimetric analysis to ensure unbiased sampling. With scanning electron micrographs at 200xmagnification, EC coverage was obtained by a computer-assisted image processor (LA500, Pias Inc., Osaka, Japan).
Thickness of the neoarterial wall
The neoarterial wall of the graft was defined as the layer between the EC monolayer and the vascular prosthesis. Intimal thickness was obtained from averaged neoarterial thicknesses measured at 20 points in a transverse section of a midportion of a graft, which was enlarged 200 times by means of the image processor. Planimetric results were expressed as mean plus or minus the standard deviation. Comparisons of variables between groups were made by one-way analysis of variance. When the F test indicated a statistically significant difference between groups, comparisons between the two groups were done with the two-tailed Student's t test; significant differences were defined by p < 0.05.
RESULTS
Reconstruction of vascular wall and implantation
Two models of hybrid grafts were prepared on knitted vascular grafts and implanted bilaterally in carotid arteries of dogs. Model I grafts consisted of the EC monolayer, and model II grafts consisted of the EC monolayer and SMC multilayers. A mixed solution of collagen and dermatan sulfate was first uniformly coated on a knitted vascular prosthesis and subsequently thermally gelled to form collagenous meshes on the luminal surface. The formed gel served as an artificial basement membrane for ECs. When the solution was premixed with SMCs, the formed gel served as a hybrid media and as an artificial basement membrane for model II grafts. The number of SMCs actually incorporated into model II grafts was adjusted to be (4.2 ± 0.5) x 104 cells/cm 2 of the graft luminal surface. ECs adhered and spread well on the mixed collagenous meshes within several hours after high-density seeding (number of adhered ECs was (2.1 ± 0.2) x 10 5 cells/cm 2 for both models). The additional 3-day graft incubation before implantation resulted in the formation of a completely endothelialized luminal surface for both models, which was confirmed by scanning electron microscopic observation. The histologic sections of these grafts revealed that a swollen gel (thickness approximately 50 µm) existed beneath the EC monolayer. For model II grafts, SMCs embedded in the complex gel were randomly oriented.
These two model grafts were bilaterally implanted in canine carotid arteries. Among the 18 animals that received implantation, one animal was excluded from this study at 3 days after implantation because of bleeding from the anastomosis. Thus 17 animals were killed at the projected implantation periods: n = 5 for each model at 2 weeks, n = 2 at 4 weeks, n = 5 at 12 weeks, and n = 5 at 26 weeks. All the implanted grafts were patent regardless of model.
Microscopic appearance of luminal surfaces
Scanning electron microscopic observation with the aid of planimetric analysis revealed that the hybrid grafts, irrespective of model, were fully endothelialized after implantation. No defects caused by nonendothelialization or delamination were observed. At 2 weeks after implantation, there were some differences in luminal surface morphologic appearance between the two models. The luminal surfaces of model I grafts showed a slightly reduced degree of endothelialization (endothelialized area 89.4% ± 5.5%). As shown in Fig. 1, A, flat, elongated ECs at crests of crimps of the Dacron fabric vascular prosthesis were oriented along the direction of blood flow similarly to those of natural arteries, whereas ECs at troughs of crimps, which were somewhat polygonally shaped, lost their regular orientation. Leukocyte adhesion on ECs and migration into their interstices were occasionally observed at disorganized EC layers at the troughs (Fig. 1, B and C).
|
Neoarterial wall formation process
The neoarterial wall formation process of implanted grafts was studied during the entire period by light microscopic observation of histologic samples. At 2 weeks after implantation, histologic samples of model I grafts revealed a thin collagenous layer beneath the EC monolayer, into which cellular invasion from surrounding connective tissues was just initiated (Fig. 2, A). In model II grafts, the regenerating neoarterial wall consisted of populated cells and dense extracellular matrix (Fig. 2, B). These cells, distributed throughout almost the entire length of the grafts, were identified as SMCs by their positive reactivity to a muscle-actin–specific monoclonal antibody (HHF35) (Fig. 2, C). SMCs in the upper layer of regenerated neoarterial wall in transverse sections were spindle-shaped (Fig. 2, B), whereas those in longitudinal sections were spherical. The marked morphologic difference between the two sections showed that SMCs in the upper layer tended to be oriented circumferentially, whereas SMCs in the lower layer did not exhibit a regular orientation. A thin layer close to polyester fiber bundles was abundant in capillaries arising from the surrounding connective tissues.
|
|
|
|
Table I). There was no significant difference in the thickness of the SMC-containing layer except for in model I grafts at 2 weeks, irrespective of model and implantation period. The SMC-containing layers at 12 weeks' and 26 weeks' implantation were about 90 to 120 µm. The change in the total neoarterial wall thickness was attributed to the change in the thickness of non-SMC layers. For model I grafts, a considerable increase in the thickness of the non-SMC layers during the implantation period from 12 to 26 weeks was observed, whereas a considerable decrease during the same period was observed for model II grafts.
|
The concept of EC seeding onto small-caliber vascular grafts has been verified to reduce thrombogenicity of the graft luminal surface when these were implanted. However, delamination of seeded ECs because of hydrodynamic shear stress 3 and disruption of the endothelial monolayer by activated leukocytes, 9 both of which lead to reduced antithrombogenicity, were often observed in the acute phase of implantation. Because a natural artery has a hierarchic structure in which ECs, SMCs, and fibroblasts exist as separate layers, the homeostasis of a healthy artery must be maintained by interactions of the layers through autocrine or paracrine factors and through extracellular matrix. 10 Therefore ECs simply seeded on a vascular prosthesis are considered not to be as stable as natural ones in the acute phase of implantation because they lack interactions with a physiologic subendothelial matrix and other vascular cell types such as SMCs and fibroblasts.
Yue and associates 11 have shown that SMC seeding onto vascular grafts enhances the process of arterial wall regeneration including spontaneous neoendothelialization. Such evidence suggests that the incorporation of SMC layers in a hybrid graft increases the stability of the EC layer in the acute phase and accelerates neoarterial wall regeneration in the long-term phase. Under the working principle mentioned herein, we prepared EC-seeded and EC/SMC-seeded hybrid grafts and compared the degrees of endothelial stability and neoarterial wall regeneration potential between the two models for up to 6 months after implantation. Because our previous studies have shown that a mixed gel of collagen and dermatan sulfate serves as an extracellular matrix with high adhesion potential for ECs and reduced adhesion reactivity to platelets, 8,12 the mixed gel, which was uniformly formed by in situ thermal gelation, was used as a two-dimensional artificial basement membrane for ECs and a three-dimensional artificial medial extracellular matrix for SMCs. Complete endothelialization before implantation for both models and entrapment of SMCs in hybrid media in model II grafts were attained.
This study clearly differentiated the tissue regeneration potentials between the two hybrid grafts in vivo in terms of intimal stability, neoarterial regeneration, and phenotypic alteration of SMCs. Neoarterial tissues were conventionally divided into the following three subtissues according to the definition made by van der Lei and associates: 13 an endothelial lining as neointima, subendothelial SMC layers (stained by HHF35) as neomedia, and fibrohistiocytic tissue (not stained by HHF35) in the deepest layer as neoadventitia. As for the neointima, model II grafts at 2 weeks after implantation exhibited much higher degrees of stability and integrity than model I grafts. That is, as clearly observed by scanning microscopic observation studies, an integrated neointima with full endothelialization and high cellular orientation parallel to the direction of blood flow, resembling that of natural arteries, was formed in model II grafts (Fig. 1, D through F), whereas the neointima of model I grafts at this period showed a less integrated luminal surface with a reduced EC coverage probably derived from delamination, massive leukocyte adhesion, and irregular cellular alignment (Fig. 1, A through C). However, longer implantation did not produce a difference in luminal surface integrity between the two models.
As for neomedial regeneration, a marked difference between the two models was observed at 2 weeks' implantation. For model I grafts, the neoarterial thickness (approximately 30 µm in 
Table I) was smaller than that (approximately 50 µm) before implantation, indicating that there was no sign of neomedial regeneration. Only a thin collagenous layer with a few cells was observed. Surface waving because of the crimp structure of the Dacron fabric graft was observed. On the other hand, a markedly regenerated neoarterial tissue (thickness approximately 110 µm) mainly a result of the increase in the neomedial thickness was observed for model II grafts at this period, which resulted in the formation of a flat luminal surface without any crimp structure (Fig. 1, D).
Further implantation time led to the maturity of regenerated neoarterial tissue for both hybrid grafts. For the model I grafts, a rapidly increased thickness of the neoarterial wall was observed (Table I), which reached approximately 105 µm at 12 weeks mainly because of the increase in the neomedial thickness and 170 µm at 26 weeks after implantation. On the other hand, the thickness of model II grafts increased to 195 µm at 12 weeks mainly because of the increase in the neoadvential thickness, but decreased to 150 µm at 26 weeks because of degeneration of the neoadventitial tissue. As clearly shown in Table I, the thickness of the neomedia, which ranged from approximately 90 to 120 µm, once the neomedia was formed, did not significantly changed with time, irrespective of model. The change in the thickness of neoarterial wall in the long-term phase was found to be mainly determined by the thickness of the neoadventitia.
Local accumulation at subendothelial layers and circumferential orientation of SMCs were observed for model II grafts at 12 weeks, the degree of which was almost comparable to that of model I grafts at 26 weeks after implantation (Fig. 3, A and B). Proportionating to SMC orientation, collagen fibers exhibited a high degree of circumferential orientation (Fig. 3, C). Layered elastic laminae parallel to the orientation of SMCs in the neomedia were observed only in model II grafts at 26 weeks' implantation (Fig. 4, C). However, the regeneration of elastin structure in model I grafts at this period was much behind that in model II grafts.
These observations suggest that neoarterial wall regeneration until 6 months after implantation proceeds via the following three sequential stages: (1) synthetic phenotyped SMC proliferation and extracellular matrix (mainly collagen) production that form the neomedial tissue, which results in an increase in the neomedial thickness, (2) circumferential orientation of SMCs in the neomedia and regeneration of the neoadventitia, and (3) degeneration of the neoadventitia and generation of elastic laminae in the neomedia. Because an excessive proliferation of SMCs, which are totally in the synthetic stage as implanted, was not observed in the long-term phase in either model, SMCs in stage III must be the nonproliferating or contractile phenotype. Our detailed study on phenotypic modulation of SMCs in hybrid grafts, which was done by intracellular ultrastructural analysis with transmission electron microscopy, indicates that SMCs in model I grafts at 12 weeks' implantation are dominated by the synthetic phenotype, whereas those in model I grafts at 26 weeks and model II grafts at both 12 and 26 weeks after implantation are dominated by the contractile phenotype, similar to SMCs in nondiseased natural arteries. 14 It is speculated that model I grafts at 26 weeks' implantation are considered to be in stage II, whereas model II grafts at this period have already entered stage III. Therefore it can be said that model II grafts were much more advanced than model I grafts in terms of tissue regeneration, which includes matrix production and organization and cellular proliferation, segregation, orientation, and phenotypic alteration.
Our previous study 12 and those of others 13,15 have demonstrated that circumferential orientation of SMCs was markedly enhanced when pulsatile stress-responsive compliant grafts were used. In this study, we used a high-porosity knitted graft that is generally classified as noncompliant as compared with natural vessels. However, arterial pulsation produced a slight degree of pulsation of grafts because of their high porosity, which apparently induced cellular and molecular orientations in a circumferential configuration. Such mechanical stimulation has been reported as an important stimulus for the SMCs to produce extracellular matrix such as elastic laminae and collagen in vitro. 16,17 Our study may suggest that SMC circumferential layering responding to pulsatile-induced wall distention is a prerequisite for generation of well-assembled elastic laminae and phenotypic modulation, as was partly verified in our in vitro study. 14
In conclusion, the incorporation of subendothelial SMC layers into EC-seeded hybrid grafts significantly enhanced the integrity and stability of neointima to provide a nonthrombogenic potential in the acute phase of implantation and to accelerate regeneration and remodeling of the neomedia and neoadventitia in the long-term phase while a minimal increase in neoarterial thickness was maintained. It can be said that a hierarchically structured hybrid graft composed of ECs and SMCs provides considerable benefits to arterial wall replacement and regeneration.
Acknowledgments
We acknowledge the advice of Dr. Shigeko Takaichi of the National Cardiovascular Center during SEM studies. Dr. Miwa acknowledges the continuous encouragement of Professor Futoshi Iida of the Second Department of Surgery, Shinshu University School of Medicine, from which Dr. Miwa was on leave.
Footnotes
From the Department of Bioengineering, a National Cardiovascular Center Research Institute, Suita, Osaka, and the Second Department of Surgery, b Shinshu University Medical School, Matsumoto, Nagano, Japan. 
References
This article has been cited by other articles:
|
|
 |
|
  C. K. Hashi, Y. Zhu, G.-Y. Yang, W. L. Young, B. S. Hsiao, K. Wang, B. Chu, and S. Li Antithrombogenic property of bone marrow mesenchymal stem cells in nanofibrous vascular grafts PNAS, July 17, 2007; 104(29): 11915 - 11920. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. R. Hoenig, G. R. Campbell, B. E. Rolfe, and J. H. Campbell Tissue-Engineered Blood Vessels: Alternative to Autologous Grafts? Arterioscler Thromb Vasc Biol, June 1, 2005; 25(6): 1128 - 1134. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. He, T. Shirota, H. Yasui, and T. Matsuda Canine endothelial progenitor cell-lined hybrid vascular graft with nonthrombogenic potential J. Thorac. Cardiovasc. Surg., August 1, 2003; 126(2): 455 - 464. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
