Nonmuscle and Smooth Muscle Myosin Heavy Chain Expression in Rejected Cardiac Allografts
A Study in Rat and Monkey Models
Background Diagnosis of acute rejection and graft arteriosclerosis (chronic rejection) is critical to the success of cardiac transplantation, but accurate diagnosis is often difficult. We have reported that there are three types of vascular myosin heavy chain (MHC) isoforms: SM1, SM2, and SMemb. SM2 is specifically expressed in differentiated smooth muscle cells (SMCs). SMemb is a nonmuscle-type MHC abundantly expressed in SMCs of fetal aorta.
Methods and Results To evaluate the usefulness of MHC expression for diagnosis and analysis of acute and chronic rejection, heterotopic cardiac transplantation was performed in rats and monkeys. Immunohistochemistry, electron microscopy, and Northern blot assay were performed to evaluate MHC expression. SMemb was expressed in spindle-shaped cells located in acutely rejected myocardium in the rats and monkeys. These cells were also observed in areas lacking cellular infiltration. These SMemb-positive cells were activated fibroblasts or myofibroblasts. SMemb mRNA was enhanced parallel to the progression of acute rejection. In the coronary arteries of chronically rejected allografts, enhanced SMemb and reduced SM2 expression was observed in both thickened intima and media. The reduced medial SM2 expression was observed before the intimal thickening occurred. These cells were phenotypically modulated SMCs.
Conclusions Altered expression of MHC isoforms is a sensitive indicator in the diagnosis of acute and chronic cardiac rejection. The pathophysiology of this alteration in MHC isoform expression should be studied further to elucidate the pathogenesis of cardiac rejection.
Cardiac transplantation in humans as treatment for end-stage cardiac diseases has gained popularity. However, the problems of acute and chronic rejection still remain. Both types of rejection are related to the duration of survival of the grafts; few recipients are completely free of these complications. Therefore, precise diagnosis of rejection is very important in the management of transplant recipients.1 2 Pathological examination of endomyocardial biopsied tissue is valuable in the clinical diagnosis of rejection phenomena. Acute rejection is characterized pathologically by mononuclear cell infiltration, myocyte damage, and fibrotic change. Chronic rejection is recognized as intimal thickening of coronary arteries.3 The pathogenesis of these cardiac rejections remains unclear.
We recently reported that there are three types of smooth muscle MHC isoforms expressed in vascular SMCs: SM1, SM2, and SMemb. SM1 is expressed throughout early developmental stages to mature stages, and SM2 is expressed only after birth. These two isoforms are specific to SMCs, whereas SMemb is a nonmuscle-type MHC isoform that is expressed in embryonic SMCs of fetal aorta and proliferating SMCs.4 5 6 7 8 The expression of these MHC isoforms can be a useful molecular marker for detection of pathological conditions in acute or chronic rejection in which fibrotic changes and SMC proliferation, respectively, are the predominant features. However, expression of MHC isoforms had not been examined previously in transplanted heart grafts.
As an experimental model of graft rejection, transplantation of heterotopic abdominal cardiac allografts has been developed in rats9 10 11 12 and in Japanese monkeys.13 These models have demonstrable usefulness in studies of the pathophysiology of rejection: the histological findings are similar to the acute and chronic rejection seen in clinical heart transplantation.
MHC expression in animal models was examined to determine its usefulness in the evaluation of clinical rejection and to elucidate the pathophysiology of acute and chronic cardiac rejection.
Adult male Lewis (RT1l) and F344 (RT1lvl) strain rats (age, 6 weeks; weight, 200 to 250 g) and adult male Japanese monkeys of the species Macaca fuscata (5.5 to 13.5 kg) were used in this study. They were housed under conventional conditions and fed a standard diet and water. They were obtained from Japan Charles River Laboratories.
Heterotopic Cardiac Allografts and Immunosuppression
Rats were anesthetized with 3.6% chloral hydrate IP at a dose of 1 mL/100 g. Heterotopic cardiac transplantation was performed with a standard microsurgical technique with some modification.14 F344 hearts were transplanted into F344 strain rats as isografts and into Lewis strain rats as allografts. Ischemic time averaged ≈40 minutes, and the overall success rate was >90%. In 15 of 26 allograft recipients, FK506 (Fujisawa Pharmaceutical Co) was administered after surgery at a dose of 0.1 mg/kg via daily intramuscular injection. Isografts were harvested for examination on day 7 (n=1) and day 28 (n=1). Allografts from untreated rats were harvested on days 3 (n=3), 5 (n=3), 7 (n=3), and 14 (n=2) and those from FK506-treated rats on days 7 (n=3), 14 (n=4), 21(n=4), 28 (n=2), and 35 (n=2). Native F344 hearts were used as controls. Randomly paired Japanese monkeys were anesthetized by injection of ketamine hydrochloride 25 mg/kg IM; then during surgery, the monkeys were maintained under controlled ventilation with 0.1 mg/kg panucronium bromide. Infrarenal allografting was performed by the standard technique.14 The monkeys were treated with daily intravenous administration of chimeric anti–LFA-1 mAb (kindly provided by Centocor) (3 mg·kg−1·d−1). Heart grafts were harvested for examination on day 44 (n=5). Native monkey hearts were used as controls.
Sample Preparation and Histological Examination
The cardiac allografts were cut into two parts: One half was fixed in a mixed solution of 95% ethanol/1% acetic acid, and the other half was snap-frozen. The specimens fixed in the mixed solution were embedded in paraffin and sectioned into 3-μm slices. The sections were rehydrated in PBS and stained by the standard methods of hematoxylin-eosin and van Gieson's elastic stain.15 16
The sections fixed in a mixed solution of 95% ethanol/1% acetic acid were used for immunostaining of SMemb, SM2, and α-SMA using anti-rabbit antibodies in rat samples with a Dako LSAB kit (Dako Corp). Immunoperoxidase staining for 1A29 (a mouse mAb to rat ICAM-1, CD54) was performed on frozen sections as described before.17 Anti-human antibody was used for SMemb staining in monkey samples fixed in the mixed solution. Because some antibodies did not cross-react with rat or monkey myosin, rat and monkey SM1 and monkey SM2 expression were not examined in this study. To reduce nonspecific reactions, sections were preincubated with 0.3% hydrogen peroxide and normal rabbit serum. Antibodies (1A4, Dako A/S)18 were applied and incubated for 20 minutes at room temperature. Sections were then incubated with biotinylated immunoglobulins (E464, Dako) for 10 minutes followed by horseradish peroxidase–labeled streptavidin solution for an additional 10 minutes. The slides were rinsed after each incubation step in Tris-buffered saline with Tween 20 (Wako Pure Chemical Industries). Peroxidase activity was detected by 3,3′-diaminobenzidine tetrahydrochloride (0.2 mg/mL, Sigma Chemical Co) with hydrogen peroxide (0.014%).
Transmission Electron Microscopy
Rat allografts (harvested on day 14 without and day 35 with FK506 therapy) were fixed with 2.5% glutaraldehyde in sodium cacodylate buffer (pH 7.3) and processed by the cold dehydration technique.19 Sections 1 μm thick were cut and stained with toluidine blue to isolate selected areas for ultrathin sectioning. The sections were mounted on copper grids and stained with uranyl acetate and lead citrate. Photomicrographs were taken with a JEOL 1200EX transmission electron microscope at 80 keV.
Electron Microscopic Immunocytochemistry
Rat allografts (day 14, no treatment) were cut into slices and immersed overnight in paraformaldehyde-lysine-periodate solution at 4°C. After rinsing with PBS, the tissue was processed for electron microscopic immunocytochemistry by a preembedding immunoperoxidase procedure. Sections were washed with 50 mmol/L NH4Cl in PBS and permeabilized for 1 hour by incubation in PBS containing 1% ovalbumin, 0.2% gelatin, and 0.05% saponin (solution A). The graft sections were incubated overnight at 4°C with the primary antibodies against SMemb diluted 1:200 and in PBS with 1% ovalbumin. As a negative control, sections were incubated in PBS with ovalbumin without antibody. After rinsing in solution A, the sections were incubated for 2 hours with biotinylated rabbit antiserum against mouse immunoglobulins diluted 1:200. They were rinsed in solution A and then incubated for 1 hour in horseradish peroxidase–labeled streptavidin solution. The graft sections were rinsed and fixed in 1% glutaraldehyde in PBS. After they were washed in PBS and Tris buffer, peroxidase was detected by incubation in 0.1% diaminobenzidine in Tris buffer for 5 minutes followed by addition of hydrogen peroxide to a final concentration of 0.01% and subsequent incubation for 10 minutes. The sections were postfixed for 1 hour at 4°C in 2% osmium tetroxide in 0.1 mol/L sodium cacodylate buffer containing 7.5% sucrose, dehydrated in ethanols, and embedded in epoxy resin. The ultrathin sections were stained with lead citrate and observed with the transmission electron microscope.20
Assessment of SMemb by Northern Blot Assay
The SMemb transcripts were evaluated by Northern blot assay. Total RNA was prepared from the transplanted heart grafts by the acid guanidinium thiocyanate–phenol-chloroform method.21 Total RNA (20 μg) from heart grafts was size fractionated in a 1.2% denaturing agarose gel (0.37 mol/L formaldehyde) and blotted to nylon membranes (Hybond-N, Amersham Co). The membranes were then baked for 4 minutes under UV light. The rabbit cDNA probe used for SMemb was described.7 We included a cDNA probe for GAPDH in the assay for SMemb mRNA to standardize SMemb mRNA levels in each specimen. The membranes were prehybridized, then hybridized in hybridization solution (50% deionated formamide, 5×SSC, 1×PE [50 mmol/L Tris-HCl, pH 7.5, 0.1% sodium pyrophosphate, 1% SDS, 0.2% polyvinylpyrrolidone, 0.2% Ficoll, and 5 mmol/L EDTA]) and 150 μg/mL denatured salmon sperm DNA with [32P]dCTP overnight at 42°C. Then the membranes were washed twice in 2×SSC/0.1% SDS for 5 minutes at 22°C. For 3 days at −80°C, hybridized filters were exposed to Kodak X-OMAT AR film.
Pathological Findings of Myocardium
Light microscopic study
In rat isografts, pathological changes were not observed in the myocardium. Rat allografts harvested on day 3 from recipients without immunosuppression demonstrated mild rejection with limited perivascular cell infiltration; however, interstitial mononuclear cell infiltration had not occurred. Pathological findings of moderate rejection with interstitial cell infiltration on day 5 and severe rejection with not only markedly increased cell infiltration but also interstitial edema and focal myocyte necrosis on day 7 were observed. Marked changes of acute rejection were detected by day 14. In rats with FK506 treatment, allografts on days 7, 14, and 21 showed only mild rejection with focal areas of cell infiltration and interstitial edema, but severe change was observed in allografts on days 28 and 35, suggesting slow progression of acute rejection under FK506 treatment. Monkey allografts demonstrated cell infiltration, interstitial edema, and myocyte necrosis on day 44.
In the rat native hearts and in isografts, expression of SMemb was not demonstrated. SMemb was expressed in the spindle-shaped cells located in the interstitial area in nontreated rat allografts on days 3, 5, 7, and 14. The expression was also observed in FK506-treated rat allografts on days 28 and 35. SMemb-positive cells were also observed in the interstitial area with or without cell infiltration in monkey allografts on day 44 (Fig 1⇓).
SMemb was expressed in the myocardial interstitium without interstitial cell infiltration or ICAM-1 expression in nontreated rat allografts on day 3. ICAM-1 expression was observed in the interstitial area with interstitial cell infiltration in nontreated rat allografts on day 5. SMemb-positive cells were present in the interstitial area before the advent of mononuclear cell infiltration, ICAM-1 expression, or myocardial damage (Fig 2⇓). The expression of SMemb was enhanced in correlation with the progression of acute rejection in cell infiltration and myocyte damage. However, SMemb expression was downregulated in fibrotic lesions in the rejected allografts.
Electron microscopic study
The ultrastructure of the SMemb-positive spindle-shaped cells was examined in the allografts harvested on day 14 from nontreated rats. More than 90% of the spindle-shaped cells showed SMemb-positive expression determined by immunohistochemical observation. These cells showed features similar to fibroblasts, although the Golgi complex and endoplasmic reticulum were more abundant than in quiescent fibroblasts. Some were characterized as myofibroblasts containing abundant actin filaments in the cytoplasm (Fig 3⇓).
Immunoelectron microscopic examination of allografts harvested from nontreated rats on day 14 demonstrated that SMemb was present both in the cytoplasm of fibroblasts and in myofibroblasts. Immunoreactivity for SMemb was located in the cytoplasm close to the plasma membrane and along the actin filaments (Fig 4⇓). The negative control samples without primary antibodies showed no immunoreactivity in these cells.
Pathological Findings of Coronary Arteries
Light microscopic study
In rat isografts, no pathological changes were observed in the coronary arteries. In rat allografts taken on day 35 from FK506-treated recipients, coronary arteries demonstrated development of intimal thickening, as indicated in samples with hematoxylin-eosin and van Gieson's elastic stain. None of the coronary arteries showed intimal thickening by day 14. Monkey allografts harvested on day 44 treated with anti–LFA-1 mAb showed coronary arterial intimal thickening with perivascular cell infiltration.
In native rat hearts and isografts, SM2 was positive in the media of coronary arteries. In FK506-treated rat allografts harvested on day 35, SM2 expression was eliminated in both the media and thickened intima. However, a reduced SM2 expression was also observed in the media, even in the area lacking the intimal thickening in the FK506-treated rat allografts on day 14. The expression of α-SMA was observed both in the media and in the proliferative intima of all the coronary arteries (Fig 5⇓). In native monkey hearts, SMemb was negative in the media of coronary arteries. However, SMemb was expressed in thickened intima and media in monkey coronary arteries harvested on day 44 (Fig 6⇓).
Electron microscopic study
Transmission electron microscopy revealed intimal thickening of graft coronary arteries in the allograft from FK506-treated rats on day 35. Thickened intima consisted of phenotypically modulated SMCs: some had abundant actin filaments, caveolas, and dense bodies similar to the normal phenotype. Others showed abundant mitochondria and endoplasmic reticula. Reduced actin filaments demonstrated phenotypic modulation (Fig 7⇓).
SMemb mRNA Expression by Northern Blot Assay
SMemb cDNA was hybridized to ≈7.0-kb mRNA obtained from heart grafts. The SMemb transcripts were detected in the nontreated grafts on days 7 and 14 and in the FK506-treated grafts on day 28, but the transcript was not detected in the FK506-treated grafts on day 7 (Fig 8⇓).
Clinical Diagnosis of Cardiac Rejection Remains to Be Elucidated
In human cardiac transplantation, acute rejection remains the most serious complication. Endomyocardial biopsy is the most reliable method for pathological grading and diagnosis of acute rejection. However, because of sampling error, cell infiltration may not be present in small endomyocardial biopsy specimens from rejected myocardium. Acute rejection develops through many immunological stages.22 23 Because pathologically observed cell infiltration may not be the initial event in the development of acute rejection, earlier changes should be monitored to prevent complications.1 2 3 ICAM-1 expression is found in many types of cells, including endothelial cells, monocytes, lymphocytes, fibroblasts, and epithelial cells, and thus may function in a variety of pathophysiological settings. Several studies have demonstrated that expression of ICAM-1 increased in parallel with the severity of cellular rejection and in response to therapy.24 ICAM-1 is known to be a marker for acute rejection; however, its expression may not be sufficient to allow detection of the early changes of acute rejection. Therefore, markers other than cell infiltration or expression of ICAM-1 are necessary to detect acute rejection in the early stage.
As demonstrated by concentric intimal thickening, chronic rejection is characterized by accelerated graft coronary arteriosclerosis. Vasculopathy is currently the primary cause of late death after cardiac transplantation. To explore effective treatment, it is critical to diagnose the changes before coronary intimal thickening develops, but the early stage of vasculopathy cannot be detected by coronary angiography.1 2 3 Moreover, the pathogenesis remains unclear: the mechanisms of the arteriopathy have not been examined in detail. Therefore, development of a methodology for monitoring and analyzing chronic rejection is necessary.
Acute Rejection Is Indicated by Activated Fibroblasts and Myofibroblasts Expressing SMemb
In our present study, we demonstrated that SMemb was expressed in the cytoplasm of spindle-shaped cells present in the interstitium of acutely rejected allografts. These SMemb-positive cells were activated fibroblasts rich in rough endoplasmic reticulum and the Golgi complex.25 SMemb expression was reduced when fiber production in fibroblasts became relatively inactive. Some cells expressing SMemb in the acutely rejected myocardium were characterized as myofibroblasts as verified by abundant actin filaments. Myofibroblasts are present in granulation tissue and share many morphological and biochemical features with fibroblasts and SMCs.26 They appear transiently during wound healing and play vital roles in the process of wound contraction. Several studies have revealed that myofibroblasts are observed in the fibrotic lesions of many organs.27 28 However, nothing has been reported about myofibroblasts in the myocardium of rejected heart grafts. Although the precise pathophysiological role of myofibroblasts is unclear at present, it may contribute to both the formation of fibrosis in the myocardium and the remodeling process after graft rejection.
In our experimental model, the expression pattern of SMemb differed from that of α-SMA (data not shown). SMemb was expressed earlier than α-SMA in rejected myocardium and was downregulated in fibrotic lesions. Although α-SMA is a known marker for myofibroblasts,29 30 the difference between α-SMA and SMemb expression indicated that SMemb expression is more sensitive in cell activation. Because activated fibroblasts and myofibroblasts are present in early phases of the lesion with or without cell infiltration, detection of these activated cells could be useful for early diagnosis of acute rejection. SMemb expression was enhanced in the activated fibroblasts and myofibroblasts, showing its usefulness in the early diagnosis of any type of acute rejection.
SMemb mRNA expression represented the severity of acute rejection at the transcription level. Detection of SMemb mRNA is a sensitive indicator for the diagnosis of rejection; mRNA expresses earlier than protein production.
These results show that the appearance of activated fibroblasts and myofibroblasts, as demonstrated by SMemb expression, is a useful marker for diagnosis of acute rejection.
Chronic Rejection Is Indicated by Phenotypically Modulated SMCs Expressing MHC Isoform Alteration
Chronic rejection is characterized by diffuse coronary arteriosclerosis formed by diffuse intimal thickening that consists of proliferative vascular SMCs with phenotypic modulation.31 32 33 Because the proliferation of SMCs is regulated by sustained immune responses,31 the phenotypic modulation of SMCs can be a sensitive indicator of chronic rejection.
We previously reported that altered expression of MHC isoforms was shown in experimental and human atherosclerotic lesions of aortas and coronary arteries.7 8 In the present study, we demonstrate that enhanced SMemb expression and reduced SM2 expression in the media as well as in the thickened intima of coronary arteries were shown in chronically rejected cardiac allografts. These cells are phenotypically modulated SMCs with proliferative states characterized by abundant organelles on electron microscopy, indicating the altered expression of MHC in early chronic rejection.
Altered MHC expression was also observed in the media even in the area without intimal thickening. The usefulness of this approach in sensitive detection of phenotypic modulation, before morphological changes occur, is clearly shown. We also noticed altered MHC expression present simultaneously in both endomyocardial arterioles and epicardial coronary arteries of chronically rejected allografts. This result suggests that chronic rejection could be demonstrated by altered MHC isoform expression in endomyocardial arterioles. Therefore, altered MHC expression in graft coronary arteries is an early sign of chronic rejection.
Dedifferentiation of Mesenchymal Cell Elucidates Mechanism of Rejection
Because both fibroblasts and smooth muscle cells are differentiated from immature mesenchymal cells, we have shown that the mesenchymal cells that exist in the interstitial myocardium and diffuse coronary arterial intimal thickening after acute or chronic rejection are a different phenotype from that expressed in normal heart grafts. Cells of equivalent phenotypes can exist in a culture of fibroblasts, and we have shown that they have characteristics of SMCs in balloon-injured atherosclerosis.4 5 6 7 8 The altered MHC expression can indicate phenotypic modulation of mesenchymal cells before morphological changes occur in rejected grafts. Thus, factors controlling phenotypic modulation of mesenchymal cells are essential to extending our understanding of the mechanisms underlying the pathogenesis of cardiac rejection. In addition, mRNA expression of the MHC isoform can be a sensitive indicator of phenotypic modulation. Further studies are needed to detect mRNA in situ and develop new methods for treatment, including a genetic approach.
We would like to thank Misako Horii, Masako Nakamura, Sanae Ogawa, Tomio Muneishi, Tetsuo Fukasawa, and Koichi Ikarashi for excellent technical assistance. We are grateful to Drs Satoshi Yamazaki, Noboru Watanabe, Koji Maemura, and Takeshi Kasajima for valuable discussions.
Selected Abbreviations and Acronyms
|α-SMA||=||α-smooth muscle actin|
|ICAM-1||=||intercellular adhesion molecule-1|
|LFA-1||=||human lymphocyte function–associated antigen-1|
|MHC||=||myosin heavy chain|
|SMC||=||smooth muscle cell|
- Received December 18, 1995.
- Revision received February 27, 1996.
- Accepted March 4, 1996.
- Copyright © 1996 by American Heart Association
Schroeder JS. Cardiac transplantation. In: Braunwald E, Isselbacher KJ, Petersdorf RG, Wilson JD, Martin JB, Fauci AS, eds. Harrison's Internal Medicine. 12th ed. New York, NY: McGraw-Hill; 1992:900-902.
Bruce AR, Reitz A. Heart and heart-lung transplantation. In: Braunwald E, ed. Heart Disease: A Textbook of Cardiovascular Medicine. 4th ed. Philadelphia, Pa: WB Saunders; 1992:520-534.
Billingham ME. Cardiac transplantation. In: Sale GE, ed. The Pathology of Organ Transplantation. Boston, Mass: Butterworths; 1990:133-152.
Nagai R, Kuro-o M, Babij P, Periasamy M. Identification of two types of smooth muscle myosin heavy chain isoforms by cDNA cloning and immunoblot analysis. J Biol Chem. 1989;264:9734-9737.
Kuro-o M, Nagai R, Tsuchimochi H, Katoh H, Yazaki Y, Ohkubo A, Takaku F. Developmentally regulated expression of vascular smooth muscle myosin heavy chain isoforms. J Biol Chem. 1989;264:18272-18275.
Kim HS, Aikawa M, Kimura K, Kuro-o M, Nakahara K, Suzuki T, Katoh H, Okamoto E, Yazaki Y, Nagai R. Ductus arteriosus: advanced differentiation of smooth muscle cells demonstrated by myosin heavy chain isoform expression in rabbits. Circulation. 1993;88:1804-1810.
Kuro-o M, Nagai R, Nakahara K, Katoh H, Tsai RC, Tsuchimochi H, Yazaki Y, Ohkubo A, Takaku F. cDNA cloning of a myosin heavy chain isoform in embryonic smooth muscle and its expression during vascular development and in arteriosclerosis. J Biol Chem. 1991;266:3768-3773.
Aikawa M, Sivam PN, Kuro-o M, Kimura M, Nakahara K, Takewaki S, Ueda M, Yamaguchi H, Yazaki Y, Periasamy M, Nagai R. Human smooth muscle myosin heavy chain isoforms as molecular markers for vascular development and atherosclerosis. Circ Res. 1993;73:1000-1013.
Ono K, Lindsey E. Improved technique of heart transplantation in rats. J Thorac Cardiovasc Surg. 1969;547:225-229.
Skalli O, Ropratz P, Trzeciak A, Benzonana G, Gillessen D, Gabbiani G. A monoclonal antibody against alpha-smooth muscle actin: a new probe for smooth muscle differentiation. J Cell Biol. 1986;103:2787-2796.
Tojo A, Gross SS, Zhang L, Tisher CC, Schmidt HHHW, Wilcox CS, Madsen KM. Immunocytochemical localization of distinct isoforms of nitric oxide synthase in the juxtaglomerular apparatus of normal rat kidney. J Am Soc Nephrol. 1994;4:1438-1447.
The International Society for Heart Transplantation: Billingham ME, Cary RPC, Hammond ME, Kemnitz J, Marboe C, McCallister HA, Snover DC, Winters GL, Zerbe A. A working formulation for the standardization of nomenclature in the diagnosis of heart and lung rejection: heart rejection group. J Heart Transplant. 1990;9:587-593.
Isobe M, Suzuki J, Yagita H, Okumura K, Yamazaki S, Nagai R, Yazaki Y, Sekiguchi M. Immunosuppression to cardiac allografts and soluble antigens by anti-vascular cell adhesion molecule-1 and anti-very late antigen-4 monoclonal antibodies. J Immunol. 1994;153:5810-5818.
Fawcett DW. A Textbook of Histology. 12th ed. New York, NY: Chapman & Hall; 1994:146-150.
Gabbiani G. The biology of the myofibroblast. Kidney Int. 1992;41:531-532.
Kocher O, Skalli O, Cerutti D, Gabbiani F, Gabbiani G. Cytoskeletal features of rat aortic cells during development: an electron microscopic, immunohistochemical, and biochemical study. Circ Res. 1985;56:829-838.