The Histochemical Journal31: 723–727, 1999.© 1999Kluwer Academic Publishers. Printed in the Netherlands.

Smoothelin and intermediate filament proteins inhuman aortocoronary saphenous vein by-pass grafts

Bengt Johansson1,∗,2, Anders Eriksson2, Frans Ramaekers3 & Lars-Eric Thornell11Department of Anatomy, Ume˚a University, SE-901 87 Ume˚a, Sweden; Tel:+46 90 786 5790;Fax:+46 90 786 5480; E-mail: bengt.johansson@anatomy.umu.se2Department of Forensic Medicine, Ume˚a University, P.O.B. 6042, SE-907 12 Ume˚a, Sweden3Department of Molecular Cell Biology and Genetics, Cardiovascular Research Institute Maastricht,University of Maastricht, P.O.B. 616, NE-6200 MD, Maastricht, The Netherlands∗Author for correspondence

Received 23 March 1999 and in revised form 13 July 1999

Summary

The aim of this immunohistochemical investigation was to study the distribution of the novel cytoskeletal protein smoothelinand the intermediate filament proteins vimentin and desmin in normal human great saphenous vein and in human aortocoronaryby-pass vein grafts. Smoothelin was present in most smooth muscle cells in the media of the native vein. In the neointima of thevein grafts that had beenin situ for three months or more, smoothelin was, in general, present only in few smooth muscle cells.Desmin was distributed in the same pattern as smoothelin in the native great saphenous vein. When desmin and smoothelinwere present in the neointima, smoothelin was detected in more cells than desmin. Vimentin was present in most cells in allwall layers of both the native saphenous vein and the vein grafts. Vascular smooth muscle cells containing vimentin but notdesmin or smoothelin are the principal cells in the neointima of human aortocoronary vein grafts. In some grafts, however, allthree cytoskeletal proteins were detected in the neointima. The distribution of smoothelin and desmin in aortocoronary veingrafts support the postulate that these proteins are expressed mainly in the contractile smooth muscle cell phenotype.

Introduction

In aortocoronary by-pass grafting, the autologous greatsaphenous vein is commonly used as graft material. Compro-mised graft patency, due to neointima formation with subse-quent graft stenosis, is a limitation of the long-term benefitof this procedure (Davies & Hagen 1995, Wahrborg 1997).During intima thickening, smooth muscle cells of the mediamodulate from the contractile, differentiated phenotype to thesynthesising, dedifferentiated phenotype. They also prolifer-ate and migrate into the intima. Most smooth muscle cells inthe neointima of human vein by-pass grafts exhibit the syn-thetic phenotype, but the differentiated, contractile phenotypemay reappear (Yamadaet al. 1997). In smooth muscle cells inthe neointima, the content of the muscular type of intermedi-ate filament protein desmin decreased (Gabbianiet al. 1982).The other intermediate filament protein typical for vascularsmooth muscle cells, vimentin (Gabbianiet al. 1981), is stillexpressed by smooth muscle cells in the neointima (Gabbianiet al. 1982).

Recently, the novel cytoskeletal protein smoothelin hasbeen described. This protein exhibits similarities with thespectrin super-family but does not belong to any of the pre-viously known classes of cytoskeletal proteins (van der Loopet al. 1996). Based on its down regulation in cell culture

(van der Loopet al. 1996) and wide distribution in contractilevessels, smoothelin has been proposed as a protein unique forthe contractile smooth muscle cell phenotype (van der Loopet al. 1997). However, it has not been shown if this proposalis also valid for neointima formationin vivo.

In the present investigation, we have studied the distribu-tion of smoothelin and intermediate filament proteins in thehuman great saphenous vein and in autologous aortocoronaryvein grafts. The aim was to elucidate the possible changesin smoothelin and intermediate filament distribution duringneointima formation and to confirm if smoothelin is presentonly in the contractile smooth muscle cell in anin vivomodel.

Materials and methods

Nine aortocoronary saphenous vein grafts from 8 individuals(7 males and 1 female aged 60–77 years) were obtained atautopsy 6 h to 5 days post-mortem. The grafts had beenin situfor 1 week to 8 years [1 week(n = 2), 3 months(n = 1),3 years(n = 1) and 8 years(n = 5)] and most of them werepositioned between the ascending aorta and the left anteriordescending artery. There was no occluding stenosis in anygraft. Samples from three normal great saphenous veins werecollected after coronary artery by-pass surgery. The tissues

724 B. Johanssonet al.

were embedded in Tissue Tec OTC compound (Miles inc,Elkhart, IN, USA) and rapidly frozen in liquid isopentaneprecooled in liquid nitrogen.

After storage at−80◦C, the specimens were cut in5µm sections in a Reichert-Jung Frigocut (Leica, Nussloch,Germany) or Leica 3000 CM (Leica) cryostat at−25◦C.The sections were processed for immunohistochemistryaccording to the peroxidase–antiperoxidase (PAP) technique(Sternberger 1979), and the sections intended for smoothe-lin staining were pre-treated with Triton X-100 for 10 min.In brief, the tissue sections were air-dried for 20–30 minand then incubated with 5% normal rabbit serum (DakoA/S, Copenhagen, Denmark). The sections were incubatedwith the appropriate primary antibodies for 60 min at+37◦Cor overnight at+4 ◦C. In the latter procedure, the primaryantibodies were diluted 1 : 10. The vimentin monoclonal anti-bodies, clone V9 raised against human vimentin, were pur-chased from Sanbio, Uden, Netherlands. The desmin cloneD33 monoclonal antibodies, raised against human muscledesmin, were purchased from Dako A/S. The smoothelinmonoclonal antibody R4N, raised against a chicken gizzardextract, was described by van der Loopet al. (1996). Indesmin and smoothelin double stainings, the rabbit poly-clonal antibody to desmin, A0611 (Dako A/S), was used. Allantibodies were diluted in phosphate-buffered saline (PBS)with 1% bovine serum albumin (BSA). After washing in0.01 M PBS, the sections intended for PAP staining were incu-bated with rabbit anti-mouse immunoglobulin G (Dako A/S)for 30 min at room temperature. The sections were washedagain in PBS, followed by incubation with mouse PAP (DakoA/S) for 30 min. Development of peroxidase staining wasobtained by applying 3,3′-diaminobenzidine (Sigma Chemi-cal Company, St. Louis, MO, USA) and H2O2 for 10–15 min,followed by rinsing in running water for 5 min. Some sec-tions were counterstained in hematoxylin for 10 s to visu-alise nuclei. Finally, the sections were dehydrated in gradedconcentrations of ethanol followed by xylene and mountedwith DPX (BDH Laboratory Supplies, Poole, UK). In con-trol staining procedures, the primary antibody was omitted. Insome specimens, double stainings with desmin and smoothe-lin were performed using secondary antibodies complexedwith fluorescein isothiocyanate (FITC) (Dako A/S) and Cy3(Jackson Immunoresearch laboratories inc, West Grove, PA,USA). The antibody staining was evaluated in a Zeiss Axio-phot photomicroscope (Carl Zeiss, Oberkochen, Germany)and documented with a MTI 3 CCD video camera (Dage-MTI Inc, Michigan City, IN, USA). The photographs weredigitally stored and processed using the Adobe Photoshop5.0™ (Adobe Systems Inc, Mountain View, CA, USA) imageprocessing software.

Results

Great saphenous vein

In the normal great saphenous vein, vimentin was distributedin all wall layers and stained most cells, i.e. smooth muscle

cells throughout the wall and fibroblasts in the adventitia(Figure 1A). The desmin and smoothelin antibodies labelledmost smooth muscle cells in the media, intima, and in lon-gitudinally oriented smooth muscle cells in the adventitia(Figure 1B,C). Smooth muscle cells in the vasa vasorum (notshown) but not fibroblasts in the adventitia were detectedwith the desmin and smoothelin antibodies (Figure 1B,C).Double stainings revealed that desmin and smoothelin weredistributed mainly in a similar, but not identical pattern(Figure 2A,B).

Aortocoronary saphenous vein grafts

In all grafts, there were various degrees of neointima for-mation and atrophy of the media (Figure 3). A prominentneointima was present in two grafts where only one weekhad passed since the operation. In one of these grafts, desminlabelling in the entire intima was reduced (results not shown),but in the other graft, desmin labelling was reduced only in theinner portion of the intima (Figure 3B). Smoothelin labellingin the intima and media, however, was identical in these twovessels and was as shown in Figure 3C.

Desmin was not detected in the neointima of the graft thathad beenin situ for 3 months (results not shown) or in threeof the grafts where 3 or more years had passed since theoperation (Figure 3E). In these grafts, smoothelin was eithernot detected in the neointima (Figure 3F) or was present onlyin occasional smooth muscle cells.

One graft in situ for 8 years contained many cells inthe neointima that were detected by the desmin and by thesmoothelin antibodies (Figure 3H and I respectively). In twografts in situ for 8 years, only minor areas in the neoin-tima contained smooth muscle cells labelled with desminand smoothelin antibodies (not shown). When smoothelinand desmin were detected, smoothelin was detected in morecells than desmin.

Smooth muscle cells in the media of most grafts werelabelled with the vimentin, desmin and smoothelin antibod-ies (Figure 3A–F). In two graftsin situ for 8 years, withmarked atrophy of the circular smooth muscle cells layer inthe media, it was difficult to identify the different wall lay-ers. Thus, smooth muscle cells in the media of these graftswere not clearly distinguished from cells in the neointima(Figure 3G–I).

Discussion

In the present study, we demonstrate that smoothelin typi-cally is detected where the contractile smooth muscle pheno-type is expected according to ultrastructural studies (Yamadaet al. 1997). Thus, we conclude that smoothelin is expressedin contractile smooth muscle cells in human aortocoronaryvein grafts. Our results correspond within vitro data wheresmoothelin is reported to be down regulated in culturedsmooth muscle cells that modulate to the synthesising phe-notype (van der Loopet al. 1996). Thus, when investigating

Smoothelin and IF proteins in vein grafts 725

Figure 1. Serial transverse sections of the native great saphenous vein stained with antibodies against vimentin (A), desmin (B) and smoothelin (C).The vimentin, desmin and smoothelin antibodies label most smooth muscle cells in all wall layers. Note that the innermost part of the vessel is linedwith concentrically arranged smooth muscle cells (small arrows). Beneath this layer, a thin layer with longitudinally arranged smooth muscle cellsisseen. The major part of the vessel wall consists of circularly arranged smooth muscle cells. In the adventitia, bundles of longitudinally oriented smoothmuscle cells are seen (large arrow). Lumen is in the upper part of the figure. Bar in C indicates 100µm.

Figure 2. Transverse section of a native great saphenous vein double stained with antibodies against desmin (A) and smoothelin (B). The antibodystaining is visualised with FITC (A) and Cy3 (B) labels. Desmin and smoothelin exhibit similar but not identical distribution in the smooth musclecells. Arrows indicate areas where differences in the distribution are seen. Lumen is to the left. Bar in B indicates 100µm.

smooth muscle cells, smoothelin may be a candidate markerprotein for the contractile smooth muscle cell phenotype.

In the native great saphenous vein, we report a high propor-tion of smooth muscle cells containing desmin in the media(Figure 1B,C), a finding that is uncommon in arteries. Amongthe arteries, only the basal cerebral arteries exhibit such ahigh proportion of desmin-containing smooth muscle cells(Johanssonet al. 1997, 1999). Thus, the great saphenousvein differs from most arteries in the distribution of inter-mediate filament proteins in the media. This high proportionof desmin-containing cells in the media may indicate a con-tractile function of the vessel.

Our investigation shows that desmin and smoothelinmainly have similar, but not identical distribution in the greatsaphenous vein (Figure 2). The slight variation in the locali-sation of desmin and smoothelin probably reflects that theamount of these proteins varies between individual cells.The function of smoothelin is unknown, but it is known that

smoothelin is associated with actin (Wherenset al. 1997). Thepresence of smoothelin in smooth muscle cells of the contrac-tile phenotype and location in the same cells as desmin, theintermediate filament of muscle cells, indicate a physiologi-cal role related to the contractile apparatus.

Vimentin is a major intermediate filament protein in vas-cular smooth muscle cells, but not in smooth muscle cellsin other tissues (Gabbianiet al. 1981). Here we observedvimentin in most cells in the wall of the normal vein, in linewith previous data (Yamadaet al. 1997). Since we find thatboth vimentin and desmin are expressed in the smooth musclecells in the media of the great saphenous vein (Figure 1B,C),we conclude that these smooth muscle cells are also of thevascular type and different from other smooth muscle cells,e.g. those in the gastrointestinal tract, airways and urinarybladder, which mainly contain desmin (Gabbianiet al. 1981).Furthermore, vimentin is often the main intermediate fil-ament protein in neointimal cells, in line with data from

726 B. Johanssonet al.

Figure 3. Serial transverse sections of vein graftsin situ for 1 week (A–C) and 8 years (D–I) stained with antibodies against vimentin (A, D, G),desmin (B, E, H) and smoothelin (C, F, I). The sections in A–C and G–I are counterstained with hematoxylin. In the graft that had beenin situ for1 week, smoothelin and vimentin are detected in the neointima. The desmin staining is very weak or absent in the outer part of the neointima (A–C).The graft in D–F illustrates that vimentin but not desmin and smoothelin is detected in the neointima. In H–I the circular smooth muscle cell layerin the media exhibits marked atrophy and it is difficult to separate the media from the neointima. Many cells in the neointima contain desmin andsmoothelin. Lumen is in the upper part of the figures. The bar in I indicates 100µm.

experimental studies (Gabbianiet al. 1982) and from humanvein by-pass grafts (Yamadaet al. 1997).

In one graft specimen that had beenin situ for 8 years,we observed many cells containing smoothelin evenly dis-tributed in the neointima. In two other grafts, minor areasin the neointima contained cells stained with the desminand smoothelin antibodies. Assuming that smoothelin is

expressed in the contractile smooth muscle phenotype as pos-tulated by van der Loopet al. (1996), we also report evi-dence for remodelling to the contractile phenotype in theneointima of human non-stenosed aortocoronary vein by-passgrafts. Our results are consistent with experimental studieson intima thickening where the differentiated smooth mus-cle cell phenotype reappears in the intima after a period

Smoothelin and IF proteins in vein grafts 727

(Kocheret al. 1984). In addition, smooth muscle cells exhibit-ing the contractile smooth muscle cell phenotype reappearin complicated human atherosclerotic lesions (Kocher &Gabbiani 1986) and in human vein grafts with marked lumi-nal stenosis (Yamadaet al. 1997). However, our results showthat the distribution of cytoskeletal proteins in the neointimavaries between graftsin situ for the same time period. It maybe speculated that this modulation to the smoothelin contain-ing smooth muscle phenotype indicates stabilisation of theneointima with no further increase in neointima thickness.

Acknowledgements

We thank Anna-Karin Nordlund, Margareta Enerstedt andMona Lindstrom for skilful technical assistance and GertSchaart for characterisation and production of the smoothelinmonoclonal antibodies.

This work was supported by the Swedish Medical ResearchCouncil (12X-03934), the AFA Health Foundation, theMagnus Bergvall Foundation, ‘Samverkansnamnden’ of theVasterbotten County and the Medical Faculty of UmeaUniversity.

References cited

Davies MG, Hagen P-O (1995) Pathophysiology of vein graft failure:A review.Eur J Endovasc Surg9: 7–18.

Gabbiani G, Schmid E, Winter S, Chaponnier C, de Chastonay C,Vandekerckhove J, Weber K, Franke WW (1981) Vascular smooth

muscle cells differ from other smooth muscle cells: Predominance ofvimentin filaments and a specific alpha-type actin.Proc Natl Acad SciUSA78: 298–302.

Gabbiani G, Rungger-Brandle E, de Chastonay C, Franke WW (1982)Vimentin-containing smooth muscle cells in aortic intimal thickeningafter endothelial injury.Lab Invest47: 265–269.

Johansson B, Eriksson A, Virtanen I, Thornell L-E (1997) Intermediatefilament proteins in adult human arteries.Anat Rec247: 439–448.

Johansson B, Eriksson A, Thornell L-E (1999) Intermediate filamentproteins in developing human arteries.Anat Embryol199: 225–231.

Kocher O, Skalli O, Bloom WS, Gabbiani G (1984) Cytoskeleton ofrat aortic smooth muscle cells. Normal conditions and experimentalintimal thickening.Lab Invest50: 645–652.

Kocher O, Gabbiani G (1986) Cytoskeletal features of normal andatheromatous human arterial smooth muscle cells.Hum Pathol17:875–880.

Sternberger L (1979)Immunocytochemistry, 2nd edn., Wiley Medical,New York.

van der Loop FT, Schaart G, Timmer ED, Ramaekers FC, van Eys GJ(1996) Smoothelin, a novel cytoskeletal protein specific for smoothmuscle cells.J Cell Biol134: 401–411.

van der Loop FT, Gabbiani G, Kohnen G, Ramaekers FC, van Eys GJ(1997) Differentiation of smooth muscle cells in human blood vesselsas defined by smoothelin, a novel marker for the contractile phenotype.Arterioscler Thromb Vasc Biol17: 665–671.

Wehrens XH, Mies B, Gimona M, Ramaekers FC, van Eys GJ, SmallJV (1997) Localization of smoothelin in avian smooth muscle andidentification of a vascular-specific isoform.FEBS Lett405: 315–320.

Wahrborg P (1997) Percutaneous transluminal coronary angioplasty orcoronary artery bypass grafting for coronary artery disease?ScandCardiovasc J31: 201–211.

Yamada T, Shiraishi R, Taki K, Nakano S, Tokunaga O, Itoh T (1997)Immunohistochemical and ultrastructural examination of smooth mus-cle cells in aortocoronary saphenous vein grafts.Angiology 48:381–390.