Neointimal Smooth Muscle Cells in Mouse Vein Grafts Are Not Recruited from the Adjacent Artery Open Access

Intimal hyperplasia (IH), with accumulation and proliferation of smooth muscle cells (SMCs) in the tunica intima, is a uniform vessel wall response to different stimuli or injuries [1‒3]. IH can cause graft stenosis which is the most common reason for vein graft failure [4‒7]. The narrowing lesions are commonly localized to the anastomotic regions where pronounced neointimal formations are found at the border to the adjacent artery [8]. Altered hemodynamic forces, but also the surgical trauma and the compliance mismatch between artery and vein in the anastomosis, have been suggested to initiate IH at this location [8‒10]. The recruitment pathways of the SMC cells to the IH have been of interest and several studies have indirectly pointed out the SMCs from the media to be the main source [11, 12]. Accumulated evidence indicates that progenitor cells, not originating from the local vessel wall, contribute to the neointimal formation; however, the source of these cells has not yet been established. Circulating progenitor cells with hematopoietic origin have been suggested, but the hypothesis is controversial due to contradictory data [13‒15]. In recent years, a major change in the view of vascular repair has occurred and mice models, including genetically modified strains, have had a large impact on new studies giving the current view. Regarding vein graft IH, mouse studies have shown that a considerable amount of the neointimal SMC population has an origin apart from the graft, but, in these studies, a hematopoietic origin could not be demonstrated [16, 17]. Our own studies have shown that one source is the surrounding tissue [18] and others could be some other pool of circulating progenitor cells or SMC from the adjacent artery. The aim of the present study was to develop a mouse model using an artery-vein composite graft in which we could define the exclusive contribution of SMCs from the adjacent artery to the IH in vein grafts.

The in vivo experiments were performed after prior approval from the local ethics committee for animal studies at the administrative court of appeals in Gothenburg, Sweden, Dnr 1548_2003. As bedding material for the mice, wood shavings were used, and as nesting material, wood wool was provided. SM22α deficient mice with the reporter gene LacZ in the SM22α locus (SM22α-LacZ mice) [19] were bred and back crossed with C57BL/6 (Charles River, Germany) for at least seven generations. All mice in the study belonged to the C57BL/6 strain and were between 3 and 6 months old. C57BL/6 mice are named wild-type (WT) in this study. The mice were acclimatized at least 1 week before surgery and normal laboratory diet, and water were provided ad libitum.

A graft, composed of an artery anastomosed to a vein, was formed ex vivo. The artery was a segment of the descending aorta from a mouse of the SM22α-LacZ strain, whose SMCs can be stained blue with the substrate X-gal. The vein was the caval vein originating from WT mice. The artery/vein-composite graft was implanted into twelve WT mice in an inter-position into the common carotid artery. In a control experiment, both the artery and vein segments were of WT origin and the composite graft was implanted into a SM22α-LacZ mouse.

All animals were anesthetized with isoflurane inhalation. The descending thoracic aortas were harvested from donor SM22α-LacZ mice and caval veins were harvested from WT animals. A segment of the aorta was anastomosed with the vein with six to eight interrupted Prolene sutures (11-0) (Ethicon) using a microsurgical technique, and the graft was then kept on ice. This composite graft was operatively placed in an inter-position to the right CCA of a recipient mouse by a procedure similar to that described by Zou et al. [20]. Briefly, a midline incision was made ventrally on the neck and the right CCA was exposed from its bifurcation to the thoracic aperture. The artery was divided at its mid portion and a cuff was placed externally around both ends. The ends were everted and anchored to the outside of the cuffs with 8-0 silk ligatures (Ethicon). The composite graft was inter-positioned to the CCA by sleeving its ends over the cuffs. The graft was held in place with additional 8-0 silk ligatures to the cuffs. The vascular clamps were removed, and the graft and native vessels were checked for blood filling and pulsations. After the carotid blood flow was restored and eventual bleeding from aortic branches had been stopped by compression, the animals were injected with 3 IU of heparin (Leo Pharma) intravenously. The animals were then injected with 4 IU of dalteparinnatrium (Fragmin, Pfizer) subcutaneously once daily for 5 days.

Six weeks after surgery, the mice were sacrificed by exsanguination through a cut in the abdominal caval vein under isoflurane anesthesia. Simultaneously, a perfusion with saline solution was started from the heart and continued until the effluent from the caval vein was clear. The graft specimens were harvested and mounted in O.C.T. Compound (Sakura Finetek), snap frozen in liquid nitrogen and stored at −70°C. 8-µm longitudinal sections were made on glass (SuperFrost^®Plus, Menzel-Gläser) for future staining, see below.

Sections were fixed in phosphate-buffered saline (PBS)-containing 0.2% glutaraldehyde and 2% formaldehyde for 5 min at 4°C, washed in distilled water, and air dried at room temperature. The slides were rinsed in PBS (Roche Diagnostics) and stained at 37°C overnight in PBS with 5 mm potassium ferricyanide, 5 mm potassium ferrocyanide, 2 mm magnesium chloride, and X-gal (1 mg/mL staining solution, Roche Diagnostics). The specimens were rinsed in PBS and washed in PBS supplemented with 0.1% Tween (Sigma-Aldrich) for 10 min. Descending aortas from SM22α-LacZ mice were used as positive controls and arteries from WT animals as negative controls. The number of cells was quantified using the ImageJ software.

Sections were fixed in 4% formaldehyde for 15 min in room temperature and subsequently stained with hematoxylin-eosin using standard protocols to observe the morphology of the tissue. To visualize elastin fibers in the media and to distinguish the arterial wall, sections were stained with Verhoeff-van Gieson. To identify SMCs in the media and to define the layer of IH, sections were stained with an antibody against α-smooth muscle actin (αSMA). To observe if any LacZ-positive cells were present in the IH layer, the αSMA staining was performed in specimens previously X-gal treated. For the αSMA staining: a monoclonal primary mouse antihuman smooth muscle α-actin antibody (1:4,000) (Cedarlane) was added and incubated overnight at room temperature followed by a secondary goat anti-mouse IgG2a antibody labeled with alkaline phosphatase (1:300 Southern Biotechnology Associates) for 30 min. Transverse sections of mouse aortas were used as positive controls and sections without primary antibody were used as negative controls.

To investigate cellular migration across the anastomosis, a new experimental setup with a carotid inter-positioned graft in mouse was developed. Composite grafts, consisting of a piece of aorta from SM22α-LacZ mice and a piece of vein from WT mice (shown in Fig. 1a), were inter-positioned on the common carotid artery (shown in Fig. 1b). One animal died during surgery and two grafts were occluded at the time of euthanasia.

Histology of the composite grafts and the anastomoses were analyzed in longitudinal consecutive sections throughout the composite grafts. The lumen could be followed all the way through the grafts excluding occlusions. The anastomosis was also recognized by identification of the suture line (shown in Fig. 2a, b). The artery and vein were conjoined with a continuous cellular formation across the anastomosis. IH was defined as thickening of the intima in which cells stained positive for αSMA were found. There was several cell layers thick IH which was very similar in both the vein and the arterial sections (shown in Fig. 3a). In several of the specimens, especially in connection to the anastomoses, there was a decrease of medial cells in the arterial segments.

In order to investigate whether SMCs from the arterial segment migrate to the IH of the adjacent vein, aortas from SM22-LacZ animals were anastomosed to veins from WT mice and inter-positioned to the common carotid artery (CCA) of WT mice (n = 8). After 6 weeks, longitudinal sections stained with X-gal revealed LacZ-positive cells in the arterial media but not in the neointima of the vein grafts. Moreover, the IH of the arterial section of the composite graft was negative for LacZ staining. Quantification of X-gal-positive stained cells revealed an average of 1,110 ± 84 cells, covering about 15 ± 1.9% of the tissue in the arterial part of the graft (shown in Fig. 3a), spreading of data is standard error of the mean. As positive and negative controls, aorta from SM22α-LacZ mice showed positive staining for LacZ SMC and carotid artery from WT mice showed no staining (shown in Fig. 3b). To ensure that the gene for SM22-LacZ was expressed in neointima formation of the vein, a control experiment was performed where both the artery and the vein of the composite grafts came from a WT mouse and were implanted into SM22-LacZ mice. Histological detection of X-gal revealed that the majority of the intimal cells in both the artery and vein were positive for LacZ which confirmed that SMCs in the neointima of the vein, as well as in the intimal thickening of the artery grafts, do express SM22α. Quantification of X-gal-positive stained cells revealed approximately 1,900 cells, covering about 22% of the tissue in the graft tissue (shown in Fig. 3c). Taken together, these observations indicate the recipient to be the major contributor of SMCs to the IH of both the artery and the vein. No recruitment of cells across the anastomosis to the vein graft from the connected artery was observed.

We have recently shown that decellularized blood vessel grafts, reconditioned with the recipient’s own blood but empty from cells, are efficiently recellularized in vivo after implantation. Cells were repopulating vein grafts transplanted to vena cava in pigs already after 3 days [21], the veins were completely recellularized after 5 weeks [21] and had intact cellular morphology still after 1 year in vivo [22]. Similarly, decellularized and reconditioned carotid arteries implanted to sheep were efficiently recellularized in vivo [23] and urothelium was regenerated in a pilot study in ram [24]. In the arterial grafts, IH was recognized, but the origin of the recellularized cells, or the cells in the IH, were not analyzed in neither of these studies. Previous work has shown that neointimal SMCs of mouse vein grafts partly originate from the surrounding tissue or from a source outside the local vessel wall [16, 18, 19]. The cells could also be recruited from a circulating pool of SMC progenitors or from the adjacent artery. The importance of stem cells and progenitor cells for tissue regeneration in vessel wall pathology has evolved during recent years. Genetically modified mouse models, harboring marker genes, have been used to define the origin of SMCs in vascular IH. Models that comprise a vessel graft procedure have a benefit since either the graft donor or the graft recipient could be of a strain with a reporter gene. In studies where veins from donor mice were implanted into recipient mice, more than 40% of the SMCs in the vein graft IH were of recipient origin [16, 17].

In the present study, we provide data that neointimal SMCs in mouse vein grafts are not recruited from the adjacent artery part of a composite graft but instead from the recipient. We also demonstrate that the SMCs in the arterial IH, in connection to the anastomoses, do not originate from the pre-existing arterial cells but are recruited from the recipient. These data are crucial for understanding the biology behind vein graft stenosis and helps us interpret results from previous studies that have not been able to separate recruitment of neointimal SMCs from the adjacent artery part of the graft from a distant pool of progenitors. Previously, in a related model, we demonstrated considerable interindividual variability of the cellular origin between different grafts with the same amount of IH indicating that different cellular sources could compensate each other [18]. In addition, we showed that IH developed in arterial grafts even if the contribution from the vessel wall itself was totally abolished through preoperative decellularizing treatment which proves that a pool of cells outside the graft has the ability to independently form IH in vascular grafts [23]. The recruitment pathway of these cells has not yet been established. Fundamentally, the cells can enter the graft tissue via the luminal surface, trans-adventitially, or through longitudinal migration across the anastomosis. In the current study, we successfully developed an in vivo model of composite grafts consisting of a piece of aorta anastomosed to a piece of vein, which is then inter-positioned to the carotid artery of mice (shown in Fig. 1, 2). LacZ transgenic mice are widely used in biomedical research as a powerful tool for investigating gene expression, cellular dynamics and, as in this case, cell-trace in a developmental process. The LacZ gene encodes the enzyme β-galactosidase and can be visualized through X-gal staining with whole-mount or histological staining [25]. In the current study, mice expressing LacZ under the control of the SM22 promoter, specific for SMCs, was used to study SMC dynamics and their contributions to vascular remodeling and IH. While both hemizygous and homozygous SM22-LacZ transgenic mice can be used in research, we have used hemizygous mice due to a lower risk of unintended phenotypic consequences of homozygous loss of SM22 such as vascular remodeling, inflammation, and cellular plasticity, which could confound experimental outcomes. By using SM22α-LacZ transgenic mice, SMCs could be followed by X-gal staining. From the results, we could exclude trans-anastomotic migration as a recruitment pathway of vein graft neointimal cells in mice since no LacZ-positive cells could be identified in the vein graft IH when anastomosed to a SM22α-LacZ aorta (shown in Fig. 3a). We also confirmed that the absence of LacZ-positive cells was not due to the inability of SMCs to express SM22 when localized in veins. This was proven by clear X-gal staining of LacZ-positive SMCs in vein IH of WT composite grafts implanted into SM22-LacZ mice (shown in Fig. 3c) which indicates that many of the neointimal SMCs in our model have a recipient origin. It should be noted that we cannot rule out the possibility that the adjacent artery contributes to fibroblast-like cells in the vein graft that have lost SM22 promoter activity and therefore cannot be seen by X-gal staining.

The adjacent artery is discussed as a contributor of SMCs when using a grafting procedure in mice. In the studies of Zhang et al. [17] and Wu et al. [26] analyses of the recipient contribution in different sections along the vein graft revealed that the center of the vein grafts had fewer recipient-derived cells compared with peri-anastomotic graft segments. Both authors interpreted their data to be a result of cell migration from the adjacent artery. We consider that the higher proportion of graft extrinsic cells, close to the anastomosis in those studies, could be a result of a more profound loss of graft intrinsic SMC due to the surgical trauma. Extensive loss of cells of the vessel wall in the conjunction to anastomoses has been demonstrated previously [11] and may result in a compensatory recruitment of cells from remote sources.

Our results contradict the main finding of the study by Liang et al. [27], where they claim that SMC precursors migrate from the adjacent artery into the vein graft to form IH. Their surgical model is similar to ours but differs in one important aspect: only the studied vessel segment has a separate origin compared with the adjacent artery and the recipient animal. In our study, the arterial and venous segments, as well as the recipient animal, originated from three different individual animals. In their model, the recipient animal expresses a reporter gene in tissue with neural crest origin, which includes the carotid artery. The majority of the cells of the neointimal formation are positive for the neural crest origin gene and their interpretation is that those cells migrate from the adjacent artery. In our opinion, they cannot exclude that those cells have come from a source of progenitor cells separate from the artery, but with an origin from neural crest.

Dilley et al. [11] report in a rat model that SMCs in the arterial media proliferated and migrated through the anastomosis into the expanding vein graft intima. Different recruitment pathways in mice and rats could explain the contradictory conclusions. However, we believe that their experimental setup does not consider whether SMCs are recruited from the adjacent artery or a pool of circulating progenitors. They report a decline in the proportion of labeled SMCs in the vein graft at 14 days compared with 7 days after surgery with the explanation that unlabeled arterial SMCs proportionally increase by time, which decreases the labeling index. In our opinion, a more likely mechanism involves the continuous accumulation of circulating precursor cells in the graft IH.

Our results indicate that the SMCs of the IH in the arterial segment of the graft did not originate from its underlying media since the intimal SMCs (positive for αSMA staining) did not express LacZ when the artery had SM22α-LacZ origin (shown in Fig. 3a). This is in concordance with the study of transplant arteriopathy by Shimizu et al. [15] and Li et al. [28]. They demonstrated that no neointimal SMCs originated from the underlying media when a section of aorta was inter-positioned into an allogenic recipient mouse using a technique of interrupted sutures. Similar results were presented by Hu et al. [13] where allogenic sections of aorta were grafted to the carotid artery by cuff anastomoses. In those studies, control animals with isografts did not develop any IH.

We are aware that the experimental setup in the present study diverges from both other animal vein graft models and from the situation in humans, since the artery is not only dissected but also explanted and anastomosed to the vein ex vivo, and we cannot exclude that this treatment resulted in an impaired migratory ability of the arterial SMCs. On the other hand, we and others have shown that nutrient vasa vasorum are not present in the descending aorta in mice [29] which indicate that any change in cellular behavior ought, at least, not to be secondary to lack of blood supply. Every anastomotic procedure, including end-to-end anastomoses, demands separation of the artery from the perivascular tissue to some extent and it always results in discontinuity of vessel wall structures. These events are present in all vein graft models as well as in operations in humans, and consequently, we assume that the ex vivo procedure in the present study is of minor importance for our results.

The limited sample size, particularly the use of only one control animal, is a clear limitation of the study and reduces the statistical strength of our conclusions. This low number of animals necessitates caution in generalizing the findings and in interpreting the results as definitive. However, despite this constraint, we believe the data provide meaningful biological insights and represent an important step in understanding the origin of neointimal cells in this graft model. In conclusion, we have developed a new model for morphological studies of sutured end-to-end anastomoses in mice and observed that vein graft neointimal SMCs are not recruited from the adjacent artery through trans-anastomotic migration, in mice.

We thank Anna Hallén for technical support.

This study was reviewed and approved by the Local Ethical Committee for Animal Studies in Gothenburg, Sweden. The in vivo experiments were performed after prior approval from the Local Ethics Committee for Animal Studies at the administrative court of appeals in Gothenburg, Sweden, Dnr 1548_2003.

The authors have no conflicts of interest to declare.

This work was supported by grants from the Sahlgrenska University Hospital (ALF, 3234), the Swedish Medical Research Council (2004-2042-24314-43), and Göteborg Medical Society.

K.Ö. and E.M. designed and performed the study. K.Ö., E.M., and J.H. analyzed the data. All authors contributed with drafting the manuscript and approved the final version.

All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author.