Perivenular Fibrosis in Alcoholic Liver Injury: Ultrastructure and Histologic Progression

Transcription

1 GASTROENTEROLOGY 1982;83: Perivenular Fibrosis in Alcoholic Liver Injury: Ultrastructure and Histologic Progression MASAYUKI NAKANO, THERESA M. WORNER, and CHARLES S. LIEBER Alcohol Research and Treatment Center and Section of Liver Disease and Nutrition; Bronx Veterans Administration Medical Center; Mount Sinai School of Medicine (CUNY), New York, New York Twenty alcoholic patients underwent sequential biopsy of the liver as part of their medical evaluation. Of the 10 patients with simple fatty liver, 9 showed no progression of the histologic lesion after 1-2 yr. By contrast, of 10 patients with perivenular fibrosis, of the 9 who continued to drink all showed progression. Laboratory parameters did not distinguish between these two groups at the time of initial biopsy. The ultrastructure of the peri venular lesion was evaluated in 11 patients. The thickness of the perivenular rim was variable, but there was a good correlation between the thickness of the perivenular rim and the number of mesenchymal cells surrounding the venules (correlation coefficient r = , P < 0.001). Myofibroblasts represented the most common cell type but there was also infiltration with other mononuclear cells. Collagen fibers surrounding myofibroblasts were observed in the peri venular fibrotic area. Thus, in the group of alcoholics who were investigated in this study, myofibroblast proliferation and collagen deposition around the terminal hepatic venule represent the first apparent lesions in the sequence of events leading to alcoholic cirrhosis. Liver cirrhosis develops as the final stage of injury to the liver, with nodular regeneration and accumulation of fibrous tissue (1). In the development of Received November 11,1981. Accepted April 21, Address requests for reprints to: Dr. Charles S. Lieber, Alcohol Research and Treatment Center, Veterans Administration Medical Center, 130 W. Kingsbridge Road, Bronx, New York This study was supported by the Veterans Administration and United States Public Health Service Grants AA and AA The authors thank Mr. S. Martillo for excellent technical help and the secretarial staff for their assistance by the American Gastroenterological Association /82/ $02.50 cirrhosis, alcoholic hepatitis (necrosis of the hepatocyte with polymorphonuclear leukocyte infiltration) has been considered to represent an essential step (2,3) and perivenular fibrosis (also called pericentral sclerosis) has been described in association with alcoholic hepatitis (4,5). Our laboratory more recently reported in baboons fed alcohol chronically that even in fatty livers without alcoholic hepatitis, deposition of collagen fibers around steatotic hepatocytes created a netlike fibrosis which prevailed in the perivenular areas and progressed to septum formation and cirrhosis (6,7). An electron microscopic study of perivenular fibrosis in baboons revealed that various collagen fibers are deposited around the terminal hepatic venule and that myofibroblasts proliferate with the progression of fibrosis (8). The present investigation was undertaken to define the histologic progression of this lesion in male alcoholics. In addition, we have attempted to clarify the cellular component of the perivenular lesion using electronmicroscopy. Materials and Methods Percutaneous biopsy of the liver was performed within 1 wk of admission as part of a medical evaluation in 20 male alcoholics hospitalized for detoxification or medical complications of alcoholism. They had consumed alcohol up to the day of admission. All patients subsequently underwent a second biopsy after 1-2 yr. For routine histologic examination, liver tissue was fixed in formalin and embedded in paraffin. Sections were stained with hematoxylin and eosin, and Masson's trichrome. The ultrastructure of the terminal hepatic venule (THV) was evaluated in 21 THVs of 11 patients, 4 with simple fatty liver and 7 with both fatty liver and perivenular fibrosis. Tissue was fixed in 2.5% glutaraldehyde in icecold cacodylate buffer for over 2 h, postfixed in osmium tetroxide, dehydrated in graded ethanol, processed

2 778 NAKANO ET AL. GASTROENTEROLOGY Vol. 83, No.4 Figure 1. Irregularly thickened terminal hepatic venule. In the connective tissue around the venule, there are a few mesenchymal cells. In the lower right of the venule, the connective tissue spreads into the parenchyma. (Toluidine blue, x 250). through propylene oxide. and embedded in Epon 812 (Ladd Research Industries. Burlington. Vt.). Sections of 1- /-tm thickness were cut with an LKB IV ultramicrotome (LKB Instruments, Rockville, Md.), stained with toluidine blue, and studied by light microscopy. From selected blocks, the ultrathin sections of terminal hepatic venules were cut, stained with uranyl acetate and lead citrate, and examined with a Hitachi HS-8 electron microscope (Nissei Sangyo, Mountain View, Calif.). For the morphometric study, the thickness of the perivenular connective tissue rim, the diameter of the venule, and the number of mesenchymal cell nuclei were measured on photomicrographs of the Epon thick sections at a 375 x magnification. Blood was obtained on the day before biopsy for determination by routine laboratory methods of the prothrombin time, albumin, serum glutamic oxaloacetic transaminase (SGOT). serum glutamic pyruvic transaminase (SGPT), bilirubin, and blood urea nitrogen (BUN). Results Histologic and Ultrastructural Changes Associated With Perivenular Fibrosis All liver biopsy specimens revealed steatosis to various degrees. None had features of alcoholic hepatitis such as extensive polymorphonuclear inflammation and necrosis or alcoholic hyaline. Among the 20 alcoholics who underwent an initial biopsy of the liver, 10 had perivenular fibrosis, defined by the presence of a fibrotic rim visible by light microscopy (Masson's trichrome stain) over at least two-thirds of the perimeter of the venule with a thickness >4 JLm. The rim of the venules without fibrosis in most instances was <2 JLm; it was never >4 JLm. The average thickness of the rim of the venule without perivenular fibrosis was 1.83 ± 0.28 JLm, and with perivenular fibrosis it was 7.27 ± 0.67 JLm. The range of the diameters of the terminal hepatic venules varied between 13 and 150 JLm with an average of 89.3±9.1 JLm. The accumulation of connective tissue around the venule was irregular and the fibrous strands spread into lobules along sinusoids (Figure 1). Sometimes they linked up with a netlike fibrosis in the center of the lobule. In the connective tissue a few mesenchymal cells were observed by light microscopy. The number of nuclei of mesenchymal cells in connective tissue surrounding terminal hepatic venules showed good correlation with the thickness of the rim (correlation coefficient r = , P < 0.001, n = 21) (Figure 2). To determine the nature of these cells, their fine structure was examined by electron microscopy. Mesenchymal Cells All mesenchymal cells showing nuclei were counted in the connective tissue surrounding 21 terminal hepatic venules. In fibrotic venules, mesenchymal cells were more abundant than in nonfibrotic venules; expressed as nuclei per 1 mm of inner circumference, the values were 8.1 ± 1.1 and 5.1 ± 0.3, respectively (p < 0.001). Electron microscopically, in both fibrotic and nonfibrotic venules, myofibroblasts (Figure 3A) represented the most common mesenchymal cells. Occasional fibroblasts were also observed beneath the endothelial cells in 3 ::E c: 15 ~ 10 --l ::J Z W > ~ C/) t3 5 z ~ u ~ NUMBER OF CELLS/mm of INNER CIRCUMFERENCE OF VENULE Figure 2. Relationship between thickness of venular rim and number of mesenchymal cells expressed per 1 mm of inner circumference of venule. (y = 0.6x + 1.7, r = , p < 0.001).

3 Figure 3. A. Myofibroblast (mf) around thickened terminal hepatic venule with indented nucleus, microfilaments with dense area (arrow), and many pinocytotic vesicles (v). The myofibroblast is surrounded by numerous collagen fibrils of different diameters and a basal lamina like structure (arrowhead). Bleb formation of myofibroblasts beneath endothelial cells is prominent. Lysosomes (L) are observed in endothelial cells and myofibroblasts. (Uranyl acetate and lead citrate, x 8000). B. Higher magnification of a part of Figure 3A. Collagen fibrils of three different sizes are seen around the cell process of a myofibroblast. Intermediate-size (2) and large (3) fibrils are prominent. (Uranyl acetate and lead citrate, x 22,OOO).

4 780 NAKANO ET AL. GASTROENTEROLOGY Vol. 83, No.4 Figure 4. Nonthickened terminal hepatic venule. Elongated myofibroblast cell processes without surrounding collagen fibrils are seen as well as a group of collagen fibers located between an endothelial cell (E) and a hepatocyte (H). A few microvilli of the hepatocyte are seen. (Uranyl acetate and lead citrate, x 16,000). the connective tissue. In five terminal hepatic venules with fibrosis, some mononuclear cells (lymphocytes and monocytes) were observed in the connective tissue or immediately under the endothelial cells. When the material of 21 veins was pooled, the cellular population was as follows: myofibroblasts 44%; fibroblasts 13%; and mononuclear cells 43%. Typical Ito cells (fat-storing cells) were not recognized in the connective tissue surrounding the terminal hepatic venule. The Ito cells that were seen were usually located in the midzonal perisinusoidal spaces. Fibrosis Nonthickened terminal hepatic venules. In the adventitia of nonthickened terminal hepatic venules, there were a small number of collagen fibrils or fibers and between these fibers. cell processes of myofibroblasts were observed showing characteristic fine structures of microfilaments with dense areas and pinocytic vesicles close to the cell membrane. In relatively large terminal hepatic venules. there were a few collagen fibers between endothelial cells and hepatocytes; the myofibroblasts between these fibers had branching long-cell processes surrounding these collagen fibers. Usually in these nonthickened terminal hepatic venules the collagen fibrils were-mainly of one size with a diameter of about 50 nm (Figure 4). Perisinusoidal spaces were also examined. In these spaces elongated cell processes of the myofibroblasts were seen (Figure 5) but it was very unusual to observe the nuclei of the myofibroblasts. Thickened terminal hepatic venules. In the connective tissue of the thickened terminal hepatic venules, many elongated cell processes of myofibroblasts were seen surrounded by numerous fibrils and intermediate-sized collagen fibrils with sometimes an amorphous appearance (Figures 3B and 6). In addition to the characteristic ultrastructural features of myofibroblasts, these cells often formed cytoplasmic blebs (Figure 3A) and possessed dilated endoplasmic reticulum and many glycogen granules. Some myofibroblasts between collagen fibers showed fat droplets in the cytoplasm. They looked like fat-storing cells or transitional cells between Ito cells and fibroblasts. However, these cells had microfilaments with dense areas and pinocytic vesicles in treelike branched cell processes. Some myofibroblasts had phagosomes containing small lipid droplets. In the area of the terminal hepatic venules infiltrated by mononuclear cells, the perivenular space was widened with numerous microfibrils and intermediate-sized fibrils surrounding myofibroblasts. The presence of myofibroblasts surrounded by - deposition of collagen fibrils was not limited to the peri venular area. Cell processes were seen in the

5 October 1982 PERIVENULAR FIBROSIS IN ALCOHOLICS 781 Figure 5. Perisinusoidal space in the liver of a subject with perivenular fibrosis. Cell processes of myofibroblast (MF) containing numerous microfilaments with dense areas are shown. In one cell process, a fat droplet (FJ, many pinocytotic vesicles, and coated vesicles (arrows) are seen. (Uranyl acetate and lead citrate, x 15,000). perisinusoidal spaces associated with a netlike fibrosis, sometimes linking up with the perivenular lesion. At the early stages, however, the lesions were more striking and sometimes present exclusively in the perivenular areas. Laboratory Parameters Blood urea nitrogen, albumin, total bilirubin, SGOT, and SGPT at the time of initial biopsy are reported in Table 1 for each group. For BUN, values ranged from 4 to 18 mg/dl in the group with perivenular fibrosis and from 7 to 23 mg/dl in the group without the lesion. Similarly, all patients had albumin levels within the normal range, as well as prothrombin times within 2 s of control values. Results of liver function tests were not helpful in distinguishing between these two groups. Values for bilirllbin ranged from 0.2 to 1.9 mg/dl in the patients with perivenular fibrosis and from 0.3 to 1.8 mg/dl in patients without the lesion. Serum glutamic oxaloacetic transaminase values showed a wide range in both groups. In patients with perivenular fibrosis, values ranged from 20 to 144 U/L, whereas in those without the lesion, values ranged from 29 to 144 U/L. Although there was a trend for the mean SGOT values to be higher in patients with PVF, the difference was not statistically significant. Similarly, for SGPT, there was a large variation in values for both groups: 16,-125 UlL in subjects with the lesion and 1 6 ~ U/L 1 4 in 1 those without. Thus, it appears that routine "liver tests" obtained at the time of initial biopsy are not useful in differentiating the patients with peri venular fibrosis from those with simple fatty liver. Results of laboratory data at the time of follow-up biopsy are shown in Table 2. Values for BUN varied greatly in patients with perivenular fibrosis from a low of 4 mg/dl to a high of 61 mg/dl. All patients with simple fatty liver had normal values. In addition, Table 1. Laboratory Parameters at the Time of Initial Biopsy in Patients With and Without Perivenular Fibrosis Patients with Patients without Test perivenular perivenular (normal range) fibrosiso fibrosiso BUN 10.5 ± ± 2.92 (10-26 mg/dl) Albumin 4.4 ± ± 0.2 ( g/dl) Total bilirubin 0.9 ± ± 0.1 ( mg/dlj SGOT ± ± 8.59 (0-41 U/L) SGPT ± ± 11.1 (0-45 UlL) Mean ± SEM.

6 782 NAKANO ET AL. GASTROENTEROLOGY Vol. 83, No, 4 Figure 6, Fibrotic terminal hepatic venule, with many cell processes containing microfilaments with dense areas, pinocytotic vesicles, and dilated endoplasmic reticulum are seen, They are surrounded by abundant collagen fibrils of three different sizes (1,2,3). One of the cell processes contains a small fat droplet (F). (Uranyl acetate and lead citrate, x 12,000). albumin tended to be low in 2 patients with PVF, but it was normal in all other subjects. There was wide overlap of values for both SGOT and SGPT in both groups. In subjects with PVF, values for SGOT ranged from 11 to 155 UlL; in subjects without PVF, values ranged from 10 to 318 U/L. For SGPT, values in subjects with PVF ranged from 1 to 62 U/L, while in those patients without PVF, values ranged from 11 to 178 U/L. Bilirubin values at the time offollow-up examination tended to be higher in those subjects who had steatosis with PVF on initial examination. Indeed 4 of the 10 subjects had values >3.0 mgldl. Thus, laboratory values at the time of follow-up biopsy were consistent with the histologic progression. Progression of the Peri venular Lesion Of the 20 patients who initially underwent biopsy, 18 continued to drink alcohol in excess of 2 glkg. day until their admission for the second biopsy. Of these, 9 had PVF on initial biopsy and 9 had simple fatty liver. The mean age of the groups were comparable (subjects with PVF, 47.6 ± 4.6 yr; subjects without, 46.3 ± 3.7 yr). Information on alcohol intake was obtained by a person unaware of the

7 October 1982 PERIVENULAR FIBROSIS IN ALCOHOLICS 783 Table 2. Laboratory Parameters at the Time of Final Biopsy in Patients With and Without Peri venular Fibrosis Patients with Patients without Test perivenular peri venular (normal range) fibrosiso fibrosiso BUN ± ± 1.29 (10-26 mg/dl) Albumin 3.73 ± ± 0.17 ( g/dl) Total bilirubin 3.08 ± ± 0.67 ( mg/dl) SGOT ± ± (0-41 U/L) SGPT ± ± (0-45 U/L) Mean ± SEM. histologic classification and was found to be similar in both groups (subjects with PVF, 267 ± 18 g/day over a 24 ± 1 yr period and subjects without PVF, 271 ± 7 g/day over a 26 ± 1 yr period). Two subjects remained abstinent between the biopsies, 1 with PVF and 1 without. Of the 9 patients with PVF who continued to drink, progression of the liver lesion was noted, as shown in Figure 7. One patient had thickening of the perivenular lesion develop, 2 progressed to septal fibrosis after 1 yr, and 2 had fibrosis develop after 2 yr. In another patient, incomplete cirrhosis developed after 1 yr, while 3 patients had cirrhosis develop after 2 yr. The abstinent patient showed neither histologic progression nor resolution. By contrast, in 9 out of 10 patients without PVF, no such progression was noted. In 1 patient who continued to drink, thickening of the perivenular rim was noted after 2 yr. In none of the subjects studied did classic features of alcoholic hepatitis appear, such as alcoholic hyaline or extensive polymorphonuclear inflammation and necrosis. Discussion It has previously been reported in alcohol-fed baboons that development of PVF has prognostic importance concerning the progression to cirrhosis (7). This report extends those observations to male alcoholic patients and reveals that those subjects with PVF at the fatty liver stage have a rapid progression to more severe histologic damage on repeat biopsy. Whereas histologic features were prognostically helpful, routine laboratory tests performed at the time of initial biopsy were not. Contrary to the prevailing view according to which alcoholic hepatitis is an obligatory precursor lesion of alcoholic cirrhosis (9-11), in the present study progression to cirrhosis was observed in the absence of documentation of a stage of classic alcoholic hepatitis. It is of course possible that such a stage could have been missed, but the absence of intercurrent hospitalization or other evidence of aggravation makes this unlikely. Furthermore, these clinical observations are fully consistent with our previous studies in the baboon (6), and observations made in other populations (12). Thus, although alcoholic hepatitis when present may undoubtedly contribute to the evolution towards cirrhosis, it must not be regarded as a sine qua non of such progression. Perivenular fibrosis is often associated with perisinusoidal fibrosis as described recently by Nasrallah et al. (13) and also illustrated by us previously (see Figure 1 in Reference 7). Nasrallah et al. (13) concluded that for the progression to cirrhosis, perisinusoidal fibrosis rather than PVF was predictive. It must be pointed out, however, that unlike our patients whom we selected to enter the study in the absence of alcoholic hepatitis, most of their subjects had alcoholic hepatitis. Furthermore, these authors did not indicate the size of veins studied (which affects the amount of fibrous tissue that surrounds the vessel) nor did they define THV thickening. Thus, direct comparison with our data is difficult. The present study also describes the ultrastructure of peri venular fibrosis. The increase of connective tissue surrounding the terminal hepatic venules was associated with proliferation of mesenchymal cells in the adventitia of the venule. Our study revealed that one of these cells is the myofibroblast, which has been recognized in granulation and scar tissue of skin and contributes to the retraction of these tissues (14-19). Myofibroblasts have been observed in many CIRRHOSIS INCOMPLETE CIRRHOSIS STEATOSIS INITIAL BIOPSY.. : IYR 2 YRS Figure 7. Prognostic significance of perivenular fibrosis. Progression of alcoholic liver disease was more rapid in the group of patients who had peri venular fibrosis in the initial liver biopsy specimens than in those with mere simple fatty liver.

8 784 NAKANO ET AL. GASTROENTEROLOGY Vol. 83, No.4 organs and lesions (20-33), mostly in association with tissue repair. In the liver, the presence of myofibroblasts has been reported in cirrhosis (20-23). Recently, we described early hepatic fibrogenesis in alcoholic liver disease in the baboon model (8). That study revealed that myofibroblasts are present normally around the terminal hepatic venule and in the perisinusoidal space; after alcohol treatment, they increase in number and their proliferation showed good correlation with connective tissue deposition around the terminal hepatic venules (8). The present study reveals that in the human liver also, myofibroblasts are present around the terminal hepatic venule and in the peri sinusoidal space and that after alcohol drinking, there is proliferation of myofibroblasts in association with increased collagen. Myofibroblasts are considered to synthesize the various types of collagen, (type I, III, and IV), and laminin (17,34). Immunofluorescence studies in patients and baboons have shown the diversity of collagens in early hepatic fibrosis (35,36). The various sizes of collagen fibrils observed around myofibroblasts in connective tissue of venules and in perisinusoidal spaces may be considered as its morphologic counterpart. In hepatic fibrosis after hepatic injury, Ito cells (fat-storing cells) or transitional cells between Ito cells and fibroblasts have been considered to represent the main cells contributing to collagen production (37,38). This view originated from the examination of intralobular lesions under experimental conditions with hepatocellular damage more severe than observed in our material. Whereas we did see some Ito cells in the midzonal perisinusoidal spaces, typical Ito cells were absent in the perivenular space. However, some myofibroblasts contain lipid droplets and thus resemble Ito cells. Moreover, some descriptions of fat-storing cells include the presence of microfilaments in the cytoplasm (39). Thus, these fat-storing cells, myofibroblasts, and fibroblasts appear to have some similarities and may belong to the same cell family. Transformation of Ito cells into myofibroblasts has in fact been suggested before (40). Definition of the interrelationships between these cells awaits further studies. Contrary to the heretofore prevailing view that myofibroblasts "have not been observed in noncirrhotic livers" (41), this study shows their presence in nonfibrotic human livers and their proliferation in conjunction with the thickening of connective tissue surrounding terminal hepatic venules. In the latter condition, myofibroblasts were surrounded by deposition of collagen fibrils and their presence was not limited to the perivenular area. At the early stages, however, the lesions were most striking and sometimes present exclusively in the perivenular areas. In conclusion, early stages of perivenular fibrosis in alcoholic liver injury are characterized by myofibroblast proliferation; in the patients we studied, individuals with these lesions who continued to drink heavily had a propensity for rapid evolution toward more severe stages of alcoholic liver disease. References 1. Popper H. The pathogenesis of alcohol cirrhosis. In: Fisher MM, Rankin JG, eds. New York: Plenum Press. 1977: Galambos JT. Hollingsworth MA. Falek A, Warren WD. The rate of synthesis of glycosaminoglycans and collagen by fibroblasts cultured from adult human liver biopsies. I Clin Invest 1977;60: Galambos JT, Shapira R. Natural history of alcoholic hepatitis. IV. Glycosaminoglycans and collagen in the hepatic connective tissue. J Clin Invest 1973;52: Edmondson HA. Peters RL, Reynolds TB. Kuzma OT. Sclerosing hyaline necrosis of the liver in the chronic alcoholic. A recognizable clinical syndrome. Ann Intern Med 1963; 59: Edmondson HA. Peters RL. Frankel HH, Borowsky SB. The early stage of liver injury in the alcoholic. Medicine 1967; 46: Popper H. Lieber CS. Histogenesis of alcoholic fibrosis and cirrhosis in the baboon. Am J Pathol 1980;98: Van Waes L, Lieber CS. Early perivenular sclerosis in alcoholic fatty liver: an index of progressive liver injury. Gastroenterology 1977;73: Nakano M. Lieber CS. Ultrastructure of initial stages of peri venular fibrosis in alcohol fed baboons. Am J Pathol 1982;106: Helman RA. Temko MH. Nye SW, Fallon HJ. Alcoholic hepatitis-natural history and evaluation of prednisolone therapy. Ann Intern Med 1971;74: Boitnott JK. Maddrey WC. Alcoholic liver disease. I. Interrelationships among histologic features and the histologic effects of prednisolone therapy. Hepatology 1981;1: Galambos JT. Natural history of alcoholic hepatitis. III. Histological changes. Gastroenterology 1972;63: Karasawa T. Kushida T. Shikata T. Kaneda H. Morphologic spectrum of liver diseases among chronic alcoholics. Acta Pathol Jpn 1980;30: Nasrallah SM. Nassar VH, Galambos JT. Importance of terminal hepatic venule thickening. Arch Pathol Lab Med 1980;104: Gabbiani G. Ryan GB. Majno G. Presence of modified fibroblasts in granulation tissue and their possible role in wound contraction. Experientia 1971;27: Gabbiani G, Hirschel BI. Ryan GB. et al. Granulation tissue as a contractile organ: a study of structure and function. J Exp Med 1972;135: Baur PS. Barratt G. Linares HA. et al. Wound contractions. scar contractures and myofibroblasts: a classical case study. J Trauma 1978;18: Gabbiani G, Le Lous M. Bailey AJ. Delaunay A. Collagen and myofibroblasts of granulation tissue: a chemical. ultrastructural and immunologic study. Virchows Arch B Cell Pathol 1976;21: Ryan GB, Cliff WJ. Gabbiani G, et al. Myofibroblasts in human granulation tissue. Hum Pathol 1974;5: Rudolph R. Guber S, Suzuki M. Woodward M. The life cycle of the myofibroblast. Surg Gynecol Obstet 1977;145:

9 October 1982 PERIVENULAR FIBROSIS IN ALCOHOLICS Bhathal PS. Presence of modified fibroblasts in cirrhotic livers in man. Pathology 1972;4: Rudolph R, McClure WJ. Woodward M. Contractile fibroblasts in chronic alcoholic cirrhosis. Gastroenterology 1979; 76: Irle C. Ryan GB. Gabbiani G. Majno G. Myo-fibroblasts in the cirrhotic liver. Experientia 1974;30: Irle C. Kocher O. Gabbiani G. Contractility of myofibroblasts during experimental liver cirrhosis. J Submicrosc Cytol 1980;12: Kobayaski K. Fine structure of the mammalian renal capsule: the atypical smooth muscle cell and its functional meaning. Cell Tiss Res 1978;195: Guldner FH. Wolff JR. Keyserlink G. Fibroblasts as a part of the contractile system in duodenal villi of rat. Z Zellforsh 1972;135: Ghadially FN. Lalonde JMA. Yong NK. Myofibroblasts and intracellular collagen in torn semilunar cartilages. J Submicrose Cytol 1980;12: Squier CA. Kremenak CR. Myofibroblasts in healing palatal wounds of the beagle dog. J Anat 1980;130: Thomas WA. Jones R. Scott RF, et al. Production of early atherosclerotic lesions in rats characterized by proliferation of "Modified smooth muscle cells." Exp Mol Pathol 1963; 1(Suppl): Vasudev KS. A sarcoma of myofibroblasts: an ultrastructural study. Arch Pathol Lab Med 1978;102: Goellner JR. Soule EH. Desmoid tumors. An ultrastructural study of eight cases. Hum Pathol 1980;11: Rothouse LS. Majack RA. Fay JT. An ameloblastoma with myofibroblasts and intracellular septate junctions. Cancer 1980;45 : Ohtani H. Sasano N. Myofibroblasts and myoepithelial cells in human breast carcinoma. An ultrastructural study. Virchows Arch Abt A Pathol Anat HistoI1980;385: James WD. Odom RB. The role of the myofibroblast in Dupuytren's contracture. Arch Dermatol 1980;116: Bailey AI. Sims TJ. Gabbiani G. Collagen of Dupuytren's disease. Clin Sci Mol Med 1977;53: Hahn E. Wick G. Pencev D. Timpl R. Distribution of basement membrane proteins in normal and fibrotic human liver: collagen type IV. laminin. and fibronectin. Gut 1980;21: Hahn EG. Timpl R. Nakano M. Lieber CS. Distribution of hepatic collagens. elastin and structural glycoproteins during the development of alcoholic liver injury in baboons. Gastroenterology 1980;79: McGee JO. Patrick RS. The role of perisinusoidal cells in hepatic fibrogenesis: an electron microscopic study of acute carbon tetrachloride liver injury. Lab Invest 1972;26: Kent G. Inouye T. Minick OT. Aahu M. Role of lipocytes (perisinusoidal cells) in fibrogenesis. In: Wisse E. Knook DL. ads. Kupffer cells and other liver sinusoidal cells. Amsterdam: ElsevierlNorth-Holland Biomedical Press. 1977: Hruban Z. Russell RM. Boyer JL. Glagov S. Pagheri SA. Ultrastructural changes in liver of two patients with hypervitaminosis A. Am J Pathol 1974;76: Farrel GC. Bathal PS. Powell LW. Abnormal liver function in chronic hypervitaminosis A. Dig Dis Sci 1977;22: Orrego H. Israel Y. Blendis LM. Alcoholic liver disease: information in search of knowledge? Hepatology 1981;1: