Proteins in Rat Glomerular Mesangial Cells - Europe PMC

American Journal ofPathology, Vol. 137, No. 6, December 1990 Copyright © American Association of Pathologists

Ultrastructural Organization of Contractile Proteins in Rat Glomerular Mesangial Cells D. Drenckhahn,* H. Schnittler,* R. Nobiling,t and W. Krizt From the Institut fur Anatomie,* Universitat Wiurzburg,

Wiirzburg; Pbysiologisches Institut,j Universitit Heidelberg, Heidelberg; and Institutfur Anatomie und Zellbiologie,j Universitat Heidelberg, Heidelberg; Federal Republic of Germany

Glomerular mesangial cells of the rat kidney contain actin, nonmuscle myosin, tropomyosin, and the muscular Z-line protein, a-actinin. This was shown for actin, myosin, and a-actinin by immunoblotting as well as by immunoelectron microscopy. Tropomyosin was localized in mesangial cells by immunofluorescence. In cultured mesangial cells, actin, myosin, and a-actinin constitute a considerable amount of the total cellular protein contents. In mesangial cells in situ actin, myosin and a-actinin were found to be colocalized within conspicuous microfilament bundles that traverse the cell body or major processes in various directions and project into either the tonguelike pericapillary processes, which run toward mesangial angles, or into the microvilluslike lateral extensions that abut on the perimesangial portion of the glomerular basement membrane (GBM). Thereby, the GBM of opposing mesangial angles as well as of opposing portions of the perimesangial GBMI are regularly interconnected byfilament bundles within mesangial cells that contain actin, myosin, and aactinin. The authors suggest that the majorfunction of actin-, myosin-, and ac-actinin-containing filament bundles in mesangial cells is to create an isometric tension (or minute isotonic contractions) to counteract the distendingforces of the rather high intracapillary hydraulic pressure and its resulting pressure gradients across the capillary wall and across theperimesangial GBM. (AmjPathol 1990,

groups.1"4 However the conclusion that mesangial cells in situ behave as contractile cells is based on indirect evidence. Functional studies on isolated glomeruli in vitro have demonstrated contraction of the glomerular tuft in response to vasoactive substances,56 but it has not been shown definitely that contraction was effected by mesangial cells. In several ultrastructural studies it was shown that mesangial cells in situ contain bundles of microfilaments similar to vascular smooth muscles.7-10 The microfilaments frequently insert into cytoplasmic dense bodies in a manner reminiscent of the association of contractile filaments with dense patches in smooth muscle cells.21112 The microfilaments are most dense in the many cell processes that extend toward the GBM711,1314 The first evidence for the presence of muscular contractile proteins in the glomerular mesangium was obtained with fluorescein-labeled antibodies to a mixture of actin and myosin extracted from human uterus. The antibodies reacted with the mesangial region of human renal glomeruli.15.16 Using fluorescent phallacidin, a probe specific for filamentous actin (F-actin), Andrews and Coffey11 showed a strong actin-specific label associated with the glomerular mesangium of the rat. Preliminary immunoelectron microscopic observations indicate the presence of actin, myosin, and a-actinin in chicken, rat, and human mesangial cells.17 In addition, in mesangial cells of chicken glomeruli, talin and vinculin were demonstrated at the lightmicroscopic level.17 Both proteins are involved in linking actin filaments to the plasma membrane and to the receptor for fibronectin and laminin. The ultrastructural pattern of microfilament bundles and the distribution of contractile proteins within mesangial cells is still unknown. In this study we have applied recently developed electron microscopic techniques to get further insight into the ultrastructural organization and distribution of muscular contractile proteins in mesangial cells of the rat kidney.

137:1343-1351)

Mesangial cells are generally believed to be contractile cells similar to smooth muscle cells. Contraction of mesangial cells in culture has been reported by several

Supported by the Deutsche Forschungsgemeinschaft, grant Dr 91/7-13 and Kr 546/5-3, Forschergruppe 'Niere' Heidelberg. Accepted for publication July 23, 1990. Address reprint requests to Professor Dr. med. D. Drenckhahn, Department of Anatomy, University of WUrzburg, Koellikerstr. 6, D-8700 WOrzburg, FRG.

1343

1344

Drenckhahn et al

AJP December 1990, Vol. 13 7, No. 6

Materials and Methods

Immunocytochemistry

Transmission Electron Microscopy

For immunoelectron microscopy, kidneys of adult rats (Hannover Wistar strain) of both sexes were fixed by vascular perfusion through the aorta using a fixative containing 2% paraformaldehyde and 0.1% glutaraldehyde in PBS (ph 7.4). Small pieces of cortical tissue were dehydrated in a graded series of ethanol, embedded in LR-White (London Resin Co., Woking, UK) and processed for immunogold-electron microscopy exactly as described in a previous study.-026 Ultrathin sections of kidney tissue embedded in LR-white were collected on water and retrieved on gold grids. Antibody labeling of the sections was done at room temperature and involved the following steps: 1) Incubation of the grids in a drop (50 AI) of 0.2% gelatin in PBS; 2) incubation for 60 minutes in affinitypurified antibodies against myosin, a-actinin, and actin (10 ,g/ml) containing 0.1% gelatin in PBS; 3) rinsing in 20 drops (50 Al each) of PBS containing 0.2% Triton X100 (15 seconds per drop); 4) transfer to goat anti-rabbit IgG coupled to 10 nm collodial gold (Janssen, Beerse, Belgium) diluted 1: 50 in PBS containing 0.1% gelatin; 5) rinsing and washing as after the first antibody; 6) repeat of the washing procedure using distilled H20 instead of PBS and Triton; 7) contrasting in a drop (50 ,ul) of 2% uranylacetate in H20 for 30 minutes; 8) final rinsing in 5 drops of distilled H20 (1 minute per drop). Afterwards the grids were blotted on filter paper and allowed to air dry. The sections then were examined with a Zeiss EM 9 electron microscope (Carl Zeiss, Oberkochen, FRG). Immunofluorescence was applied to semithin sections (0.5 A) of quick-frozen, freeze-dried, and Epon-embedded tissue pieces exactly as described elsewhere.' Briefly, semithin sections (0.5 A) of the tissue embedded in Epon were mounted on glass slides. Removal of the resin was obtained by placing the slides in 10% sodium methoxide dissolved in a 1:1 mixture of methanol and toluene. After rinsing with acetone (2 X 5 minutes), distilled water (2 x 5 minutes), and PBS containing 2% bovine serum albumin (15 minutes), sections were incubated with primary antibodies (dissolved 1:30 in PBS) for 30 minutes. The bound antibodies were visualized by rhodamine-labeled goat anti-rabbit IgG (Behring-Werke, Marburg, Federal Republic of Germany) diluted with PBS (1:40). Controls were performed using the primary antibodies absorbed with an excess of purified actin, a-actinin, and tropomyosin from chicken gizzard and myosin from human platelets, respectively.

Renal tissue used in this study was taken from previous experiments12,14'18 and handled as described there. Briefly, female Munich-Wistar rats (body weight, 140 to 180 g) were fixed by vascular perfusion with a solution of 1.5% glutaraldehyde/1.5% formaldehyde in 0.1 mol/I (molar) cacodylate buffer (7.4) complemented with 0.5 g/l picric acid, 5 mI/I 16% CaCI2, and 30 g/l hydroxyethylcellulose with or without prior flushing. Blocks of cortical tissue were postfixed in the same fixative for 12 hours, washed, and cut into 500-,u-thick slices using a tissue sectioner. The slices were pretreated with DMSO (dimethylsulfoxide; 5% in 0.1 mol/l cacodylate buffer for 1 hour) or briefly with a weak solution of OsO4 (0.1% OSO4 in 0.1 mol/l cacodylate buffer for 30 minutes), and afterwards stained en bloc with tannic acid (1% in 0.05 mol/l maleate buffer for 3 hours) and uranyl acetate (1% in 0.05 mol/l maleate buffer for 2 hours). Dehydration was performed in a graded series of acetone at temperatures gradually decreasing to -30 C.14,19 Immersion in a 1/1 mixture of propylene oxide and epoxy resin (Epon-812; E. Fullam, Inc., Latham, NY) was carried out at 30°C. The temperature then was allowed to rise to room temperature and the final embedding in Epon-812 was performed by standard procedures at 650C. Ultrathin sections (gray to silver) were stained with uranyl acetate and lead citrate and observed in a Philips 301 electron microscope (Philips Electronics, Eindhoven, The Netherlands).

Antibodies, Immunoblotting All antibodies used in this study were raised in rabbits. Specificity of the antibodies to actin, a-actinin, and tropomyosin from chicken gizzard and myosin from human platelets has been described.'2," All antisera were affinity purified using the respective antigens immobilized by transfer to nitrocellulose paper. The bound immunoglobulins (IgG) were eluted either with phosphate-buffered (25 mmol/I [millimolar]) saline (120 mmol/l) (PBS, ph 7.4) warmed to 560C, or with 0.2 mol/ glycine-HCL (ph 2.8) at 40C, as described elsewhere.22 For immunoblot analysis, proteins of isolated rat glomeruli23 or cultured mesangial cells24 were separated by sodium dodecyl sulfate (SDS) polyacrylamide electrophoresis (PAGE) and electrophoretically transferred to mitrocellulose paper (Schleicher & SchOll, Darmstadt, Federal Republic of Germany). Strips of the nitrocellulose paper were then processed for antibody labeling using the peroxiodase-anti-peroxidase method as described.25

Results Ultrastructure The mesangial cell is a highly branched cell type (Figure 1). Mesangial cell processes were seen to extend in all

Contractile Proteins in Mesangial Cells

1345


directions. Large processes could be distinguished from abundant microvilluslike microprojections arising directly from the cell body or a major cell process. The microprojections generally run toward the perimesangial part of the glomerular basement membrane (GBM). As already shown in a previous study,14 typical were sectional profiles of tonguelike processes (Figures 1 and 2) arising from the terminal, often attenuated portion of a major cell process (less frequently from the cell body directly) and extending for variable distances into the space between the endothelium and the GBM. Two such processes, arising from a common trunk, often were observed to bridge the "Af _ * : j1:entire distance between two mesangial angles. All these cell processes contained abundant microfilaments, usually arranged in bundles. The highest density of microfilaments was found within the tonguelike cell processes and the microvilluslike microprojections. Within major cell processes, bundles of microfilament bundles occasionally were found to extend longitudinally in the axis of the cell process. More frequently, however, the bundles were seen to traverse these processes running in various angles diagonally from one side to the opposite NvGBM (Figure 3). Bundles of microfilaments crossing each other also were observed. Within the cell body, microfilament bundles were less frequently encountered. The bundles found in the cell periphery immediately beneath the | ; ; | w e *were So plasma membrane. The perinuclear region generally was free of microfilaments. In addition to microfilaments, the cell body and the major cell processes contained nuF i .........merousmicrotubules and intermediate filaments running in all directions.

.

t

._ M

Immunoblotting _

2

t '

%

.'.,i' ''

Figure 1. Glomerular mesangium bordering two capillaries (C) and Bowman 's space (S). The GBM from opposing mesangial

Isolated rat glomeruli as well as cultured rabbit mesangial cells contained polypeptide bands at Mr 200,000, 100,000, and 42,000 that were specifically labeled with antibodies to myosin (200,000), a-actinin (100,000), and actin (42,000). In cultured cells, the a-actinin antibody also labeled a minor polypeptide band at Mr 80,000, which is probably a proteolytic fragment of a-actinin. Thus, both the glomerular and mesangial isoforms of myosin, a-actinin, and actin have the same molecular weights as their muscular counterparts in smooth muscle. In corresponding ¢',,,,' lanes, .r':';''S these. bands turned out to correspond precisely to the three predominant glomerular peptide bands that migrate at Mr 42,000; 100,000; and 200,000 (actin, myosin, a-actinin; Figure 4).

angles (arrows) at both capillaries is connected to tonguelike Immunofluorescence mesangial cellprocesses that contain bundles of microfilaments (MF). Within the axial region ofthe mesangium smallfingerlike Incubation of semithin tissue sections with polyclonal anmicroprojections of mesangial cells run toward the perimensangial GBM. TEM; X-- 1900. ~tibodies raised against actin and a-actinin from chicken

1346

Drenckhahn et al

AJP December 1990, Vol 13 7, No. 6

-/t-G

B

M~GB

Figure 3. Axial region of the mesangium showing many fingerlike microprojections (some are marked by arrows) arising from a major mesangial cellprocess and abutting on the perimesangial GBM. Bundles of microfilaments (MF) traverse the cytoplasm of the mesangial cell process continuing into the microprojections. S, Bowman's space. TEM; X -24, 000.

a

A

gizzard and myosin from human platelets resulted in a staining pattern of glomeruli that was virtually identical for all three antibodies (Figure 5). As documented already in a recently published paper, all antibodies reacted with the foot processes of podocytes, the squamous epithelium of Bowman's capsule, and with numerous irregularly shaped, branching cells interposed between glomerular capillaries, which most likely represent sectional profiles of mesangial cells. In the present study, an antibody raised against chicken gizzard tropomyosin was also applied. Staining was found to be most intensive over mesangial regions (Figure 5). All antibodies also reacted with the squamous epithelium of Bowman's capsule (not shown).

Immunoelectron Microscopy Figure 2. Juxtacapillary region of the mesangium showing tonguelike mesangial cell processes that contain bundles of microfilaments (MF) and are connected to the GBM at both mesangial angles (arrows) either directly or by interposition of extracellular microfibrils (F). C, capillary lumen; S, Bowman's space. TEM; X 31,500. -

Incubation of ultrathin sections of the rat kidney with antibodies specific for actin, myosin, and a-actinin resulted in a more or less identical IgG-gold labeling pattern (Figures 6 and 7).

Contractile Proteins in Mesangial Cells 1347 AJP December 1990, Vol. 13 7, No. 6

1 2 2k200oD o3 4

kD i . 200

100

100 ||

42

1

~~42

|g E

a 311 1 i10b

2 34

:

with the uftrastructural distribution of microfilament bundles within mesangial cells. With respect to actin, these results confirm and extend previously published studies dealing with the localization of actin in mesangial cells at the lightmicroscope level.11'17 The role of a-actinin within the microfilament bundles would be to cross-link actin filaments and to produce arrays of actin filaments of opposing polarity, as does a-actinin in the Z-line of striated muscle or

|#

gg:;

b,l

Figure 4. Western blot of isolated glomeruli (a) and cultured mesangial cells (b) of the rat kidney. Proteins were separated by SDS-PAGE transferred to nitrocellulose filters and stained either for proteins by Amido Black (lane 1) or labeled with antibodies to human platelet myosin (lane 2), chicken gizzard a-actinin (lane 3), or chicken gizzard actin (lane 4). The immunolabeled bands correspond to majorpolypeptide bands at 200 kd (myosin), 100 kd (a-actinin), and 42 kd (actin).

The antibody to tropomyosin did not bind under the fixation and embedding conditions used in this study. The other three antibodies directed against actin, myosin, and a-actinin showed a high density of label in the various mesangial cell processes. The relative density of immunogold particles (particles per unit area of cross section) overlying processes versus the cell body was 77% for actin (1352 particles counted), 84% for a-actinin (850 particles), and 68% for myosin (1224 particles), respectively. Cell processes running toward mesangial angles and extending into the space between the GBM and the capillary endothelium exhibited the highest density of labeling for all three antibodies. In longitudinally sectioned profiles of these processes, the pattern of the antibody label clearly correlated with the microfilament bundles generally observed at this site. In addition, high density of label was found also overlying the many small microvilluslike processes. The density of antibody label overlying the large cell processes was quite heterogeneous: labeling was found in some sites, whereas in others little or no label was encountered. Labeling also was found close to the plasma membrane of the cell body. Absorption of the primary antibodies with an excess of the purified antigens completely abolished immunostaining and immunogold labeling of mesangial cells.

Discussion This study clearly shows that the distribution of three contractile proteins (actin, myosin, and a-actinin) correlates

Figure 5. Localization ofactin (a), myosin (b), and tropomyosin (c) by immunofluorescence in semitbin (as5 m) sections of a glomeruli of the rat kidney. a and b are serial sections of the same glomerulus. Note brightly stained sectional profiles of mesangial cells at the level of their cell body (n, nucleus) and their processes (arrowheads). Small arrows in (a) and (b) indicate immunolabel associated withpodocytepedicles abutting on the wall of capillaries (C). Anti-tropomyosin (c) does not

stain pedicles. X- 1,380.

1348

Drenckhahn et al

AJP December 1990, Vol. 137, No. 6

lF

"

C

Figure 6. Ultrathin sections of rat glomerulU incubated with antibodies to actin (a and b), to myosin (c) and to a-actinin (d) and 10nm anti-IgG gold particles. Note the dense labeling with all three antibodies overlying the juxtacapillary, often tonguelike processes of mesangial cell (stars). In addition antibody label is also found in more central parts of mesangial cells. C, capillary lumen; S, Bowman's space. X (a) -40,000, (b) -37,000, (c) 48,000, and(d) 35,000. -


1349


'01t.

C .e ..+ ..

.l.

I

I I

I Figure 7. Ultrathin sections of rat glomeruli. Distribution of the antibody label specific for actin (a) and myosin (b) in mesangial cells. Note that the labeling pattern with both antibodies is most dense in the juxtacapillary processes (stars) and in the fingerlike microprojections (arrows) found in more centralparts of the mesangium. C, capillary lumen; S, Bowman 's space. (a) X 75,000; (b) X 37, 000. -

-

in the dense bodies of smooth muscle cells.27-29 To create contraction, an interaction with myosin filaments is necessary. In nonmuscle cells, myosin filaments are usually too thin and short to be shown between the bundles of microfilaments by conventional transmission electron microscopy. Under certain in vitro conditions, however, nonmuscle myosin filaments may associate to form thicker filaments. That is probably the reason why Pease' observed numerous myosinlike filaments in mesangial cells of the rat kidney when fixation of the tissue was omitted and dehydration performed by ethyleneglycol or freezesubstitution. In the present study, evidence is provided that myosin is regularly associated with actin and a-actinin

at places where microfilament bundles are encountered. In addition, tropomyosin-as shown at the LM-level-is also abundant in mesangial cells; however its subcellular location still remains to be determined. The functional relevance of the contractile apparatus in mesangial cells is in debate. Based mainly on studies of cultured mesangial cells, it has been postulated that

mesangial cells in vivo also have contractile properties and that mesangial cell contraction may cause the changes in the ultrafiltration coefficient that occur in vivo in response to the application of various vasoactive substances such as arginine-vasopressin or angiotensin 11. In vitro, both substances initiate contractions of cultured

1350

Drenckhahn et al


mesangial cells.31 32However morphometric studies33-36 on the glomerular tuft after in vivo fixation have so far failed to show adequate changes in the filtration area to explain the large changes in the ultrafiltration coefficient found in comparable functional studies.373 As shown recently,14 mesangial cells have tonguelike processes extending underneath the capillary endothelium toward mesangial angles, where they are attached to the GBM either directly or indirectly by the interposition of extracellular microfibrils. Typically, actin filament bundles within these processes display a transverse orientation forming continuous actin-, myosin-, and a-actinin-containing arrays of filaments that interconnect the plasmalemmal attachment sites of the GBM from two opposing mesangial angles. At other places, bundles of actin filaments were seen to project from processes at mesangial angles to major cell processes or to the cell body of mesangial cells, where they intermingle with bundles running toward a mesangial angle of an opposite capillary. Most frequently, microfilament bundles within major cell processes were seen to traverse the cytoplasm of these processes and to extend into one of the many microvilluslike microprojections. These latter processes are often found to abut on the perimesangial GBM or to be connected to the GBM by microfibrils. Interaction of myosin and actin may generate isometric tension or a minute isotonic contraction by which opposite parts of the GBM might be held together. The contractile apparatus of mesangial cells obviously provides stable cellular cross-connections between opposite parts of the GBM, including cross-connections between the GBM of opposing mesangial angles. This system can be interpreted as to counteract the distending forces acting on the mesangial angles and on the perimesangial GBM. Such forces probably are created by the high transcapillary pressure gradient and the intramesangial pressure. Thus to a certain degree the mesangial actomyosin system may balance wall tension of glomerular capillaries and maintain a 'slim' mesangium. As a whole, isometric centripetal contractions of the mesangium would prevent the tuft architecture from distension and distortion. If mesangial cells are capable of musclelike isotonic contraction, contraction of the tonguelike subendothelial mesangial cell processes should bring the GBM from opposing mesangial angles closer together. Even assuming an isotonic shortening of the processes of 50%, the effect on capillary diameter would necessarily be small (less than 10%), inasmuch as the length of the spanning mesangial process constitutes only a small fraction of the entire capillary circumference-' Those small effects cannot explain the significant changes of the filtration coefficient of up to 50% that have been observed by in vivo studies.373 In this context it is of interest to know that we were unable to demonstrate the smooth muscle alpha-isoform of actin

in mesangial cells of the rat in situ.41 This smooth muscle isoform of actin, however, is a major isoform expressed in mesangial cells in vitro, indicating that mesangial cells acquire smooth-musclelike contractile proteins-obviously followed by smooth-musclelike contractility-under the artificial conditions of tissue culture. In contrast to the countless connections of mesangial cell processes to the GBM, mechanical connections between cell processes of different mesangial cells are rare (in the sense that microfilament bundles of one process are functionally connected by junctions of the cell membrane to a continuing bundle within an adjacent cell). Consequently the contractile apparatus of mesangial cells appears to be predominantly connected to the cell surface contacting the GBM, which is probably the main effector structure of mesangial cell contraction. In conclusion, the contractile apparatus of mesangial cells consists of microfilament bundles that are found predominantly within the various cell processes. These processes run in various patterns toward the GBM, in which they insert either directly or by the interposition of extracellular microfibrils. Typical stress fibers and the smooth muscle a-isoform actin, which are the predominant cytoskeletal structure found in mesangial cells in culture,' are absent from mesangial cells in situ. The distribution of actin-, myosin-, and a-actinin-containing microfilament bundles within mesangial cells suggest that a main function of the mesangial contractile apparatus is to maintain the structural integrity of the glomerular tuft against the distending forces of the intracapillary and intramesangial pressure (static function). Whether true isotonic contractions of mesangial cells may occur in situ that are sufficient to cause changes of the glomerular tuft geometry and thereby regulate glomerular microcirculation (dynamic function) remains to be established.

References 1. Holdsworth SR, Glasgow EF, Atkins RC, Thomson NH: Cell characteristics of cultured glomeruli from different animals species. Nephron 1978, 22:454-459 2. Ausiello DA, Kreisberg JI, Roy C, Karnovsky MJ: Contraction of cultured rat glomerular mesangial cells after stimulation with angiotensin 11 and arginine vasopressin. J Clin Invest 1980, 65:754-760 3. Singhal PC, Scharschmidt LA, Gibbons N, Hays RM: Contraction and relaxation of cultured mesangial cells on a silicone rubber surface. Kidney Int 1986, 30:862-873 4. Foidart JB, Mahieu P: Glomerular mesangial cell contractility in vitro is controlled by an angiotensin-prostaglandin balance. Mol Cell Endocrinol 1986, 47:163-173 5. Bernik MB: Contractile activity of human glomeruli in culture. Nephron 1969, 6:1-10 6. Savin VJ: In vitro effects of angiotensin 11 on glomerular function. Am J Physiol 1986, 251:F627-F634


1351


7. Farquhar MG, Palade GE: Functional evidence for the existence of a third cell type in the renal glomerulus. Phagocytosis of filtration residues by a distinctive "third" cell. J Cell Biol 1962,13:55-87 8. Huhn D, Steiner JW, Movat HZ: Die Feinstruktur des Mesangiums in Nierenglomerulum von Hund und Maus. Z Zellforsch Mikrosk Anat 1962, 56:213-230 9. Latta H, Maunsbach AB, Madden SC: The centrolobular region of the glomerulus studied by electron microscopy. J Ultrastruct Mol Struct Res 1960, 4:455-472 10. Michielsen P: Contribution a l'etude du tissu intercapillaire, Proceedings of the First International Congress on Nephrology, Geneve-Evian. Basel, Karger, 1960, pp 657-661 11. Andrews PM, Coffey AK: Cytoplasmic contractile elements in glomerular cells. Fed Proc 1983, 42:3046-3052 12. Mundel P, Elger M, Sakai T, Kriz W: Microfibrils are a major component of the mesangial matrix in the glomerulus of the rat kidney. Cell Tissue Res 1988, 254:183-187 13. Krz W, Sakai T: Morphological aspects of glomerular function, Nephrology, Vol 1. Proceedings of the Tenth International Congress on Nephrology, London, 1987. Edited by AM Davison. London, Bailliere Tindall, 1988, pp 3-23 14. Sakai T, Kriz W: The structural relationship between mesangial cells and basement membrane of the renal glomerulus. Anat Embryol 1987, 176:373-386 15. Becker CG: Demonstration of actomyosin in mesangial cells of the renal glomerulus. Am J Pathol 1972, 66:97-110 16. Scheinman JI, Fish AJ, Michael AF: The immunohistopathology of glomerular antigens. The glomerular basement membrane, collagen, and actomyosin antigens in normal and diseased kidneys. J Clin Invest 1974, 54:1144-1154 17. Drenckhahn D, Franke RP: Ultrastructural organization of contractile and cytoskeletal proteins in glomerular podocytes of chicken, rat, and man. Lab Invest 1988, 59(5):673-682 18. Mbassa G, Elger M, Kriz W: The ultrastructural organization of the basement membrane of Bowman's capsule in the rat renal corpuscle. Cell Tissue Res 1988, 253:151-163 19. Carlemalm E, Garavito RM, Villinger W: Resin development for electron microscopy and an analysis of embedding at low temperature. J Microsc 1982, 126:123-143 20. Drenckhahn D, Derrnietzel R: Organization of the actin-filament cytoskeleton in the intestinal brush border: A quantitative and qualitative immunoelectron microscope study. J Cell Biol 1988,107:1037-1048 21. Drenckhahn D, Wagner J: Stress fibres in the splenic sinus endothelium in situ: Molecular structure, relationship to the extracellular matrix, and contractility. Eur J Cell Biol 1986,

102:1738-1747 22. Drenckhahn D, Franz H: Identification of actin, a-actinin, and vinculin-containing plaques at the lateral membrane of epithelial cells. J Cell Biol 1986, 102:1843-1852 23. Nogaard JOR: Rat glomerular epithelial cells in culture. Parietal or visceral epithelial origin? Lab Invest 1987, 57:277-290 24. Lovett DH, Sterzel RB, Kashgarian M, Ryan JL: Neutral proteinase produced in vitro by cells of the glomerular mesangium. Kidney Int 1983, 23:342-349

25. Drenckhahn D, Hofmann HD, Mannherz HG: Evidence for the association of villin with core filaments and rootlets of intestinal epithelial microvilli. Cell Tissue Res 1983, 228:409414 26. Drenckhahn D, Merte C: Restriction of the human kidney band 3-like anion exchanger to specialized subdomains of the basolateral plasma membrane of intercalated cells. Eur J Cell Biol 1987, 45:107-115 27. Groschel-Stewart U, Drenckhahn D: Muscular and cytoplasmic contractile proteins-biochemistry, immunology, structural organization. Collagen Rel Res 1982, 2:381-463 28. Small JV: Geometry of actin-membrane attachments in the smooth muscle cell: localizations of vinculin and alpha-actinin. EMBO J 1985, 4:45-49 29. Squire J: The Structural Basis of Muscular Contraction. New York, Plenum Press, 1981 30. Pease DC: Myoid features of renal corpuscles and tubules. J Ultrastruct Res 1968, 23:304-320 31. Kreisberg JI: Contractile properties of the glomerular mesangium. Fed Proc 1983, 42:3053-3057 32. Kreisberg JI, Venkatachalam K, Troyer D: Contractile properties of cultured glomerular mesangial cells. Am J Physiol 1985, 249:F457-F463 33. Haley DP, Sarrafian M, Bulger RE, Dobyan DC, Eknoyan G: Structural and functional correlates of effects of angiotensininduced changes in rat glomerulus. Am J Physiol 1987, 253: F111-F119 34. Olivetti G, Giacomelli F, Wiener J: Morphometry of superficial glomeruli in acute hypertension in the rat. Kidney Int 1985, 27:31 -38 35. Racusen LC, Prozialec DH, Solez K: Glomerular epithelial cell changes after ischemia or dehydration: Possible role of angiotensin 11. Am J Pathol 1984, 114:157-163 36. Elger M, Sakai T, Kriz W: Role of mesangial cell contraction in adaptation of the glomerular tuft to changes in extracellular volume. Pflugers Arch 1990, 415:598-605 37. Baylis C, Brenner BM: Modulation by prostaglandin synthesis inhibitors of the action of exogenous angiotensin 11 on glomerular ultrafiltration in the rat. Circ Res 1978, 43:889-898 38. Blantz RC, Konnen KS, Tucker BJ: Angiotensin II effects upon the glomerular microcirculation and ultrafiltration coefficient of the rat. J Clin Invest 1976, 57:419-434 39. Tucker BJ, Blantz RC: Mechanism of altered glomerular hemodynamics during chronic sodium depletion. Am J Physiol 1983, 244:F1 1-F18 40. Kriz W, Elger M, Lemley K, Sakai T: The structure of the glomerular mesangium: a biomechanical interpretation. Kidney Int 1990, 38(Suppl 30) 41. Elger M, Nobiling R, Drenckhahn D, Krz W: Mesangiumzellen (MZ) exprimieren in Kultur eine andere Aktin-lsoform als in situ (abstr). Nieren- und Hochdruckkrankheiten 1990, 19: 378

Acknowledgments The authors thank Hiltraud Hosser and Christa Merte for skillful technical assistance, and Ingrid Ertel and Heidi Schneider for photographic help.