Tenascin Is Synthesized and Secreted by Rat Mesangial Cells in Culture and Is Present in in Human Glomerular Diseases1

Luan D. Truong,2 Mark W. Majesky, and Jana Pindur

angial staining for TN by immunohistochemical tech- L.D. Truong, MW. Majesky, J. Pindur, Department of niques and that, in glomerular diseases character- Pathology, Baylor College of Medicine. Houston, TX ized by the expansion of the mesangial matrix such L.D. Truong, Department of Pathology, The Methodist as diabetic glomerulosclerosis, the expanded mes- Hospital, Houston, TX angial matrix was stained positive for TN. These find- ings suggest that mesangial cells in culture synthe- (J. Am. Soc. Nephrol. 1994; 4:1771-1777) size TN and that TN is a component of the mesangial matrix; moreover, increased synthesis of TN may play a significant role in the pathogenesis of mesangial ABSTRACT sclerosis and glomerulosclerosis. Tenascin (TN) is a large oligomeric recently Key Words: Tenascin. mesangial cell. extracellular matrix described as a component of the extracellular ma- trix. The distribution of TN in adult kidney tissue has not been adequately evaluated, but preliminary T enascin (TN), a recently described component of data have suggested that TN is variably seen in rare the extracellular matrix, is a large oligomeric mesangial areas and in stroma surrounding some protein composed of six similar subunits joined to- tubules. The enlargement of the mesangial matrix gether at the amino terminus by disulfide bonds (1). (mesangial sclerosis) is a common feature of many Although there are minor differences among species, renal diseases and is thought to be partially related each subunit with molecular weight ranging from to oversynthesis of the normal components of the 1 90 to 280 kd is composed of a cysteine-rich amino mesangial matrix ( type IV, , fibro- acid terminal domain followed by epidermab growth nectin, and heparan sulfate proteoglycans) by mes- factor-bike homologous repeats, Type III homologous repeats, and a fibrinogen (3 and #{244}chain angial cells. However, the possibility that mesangial homologous domain at the carboxyl terminal end ( 1 - cells are also the source of other extracellular matrix 5). There are only a few original studies in which the that participate in the process of mesangial presence of TN in kidney tissue is mentioned (1-9). sclerosis has not been explored. In this study, the Some of those studies deal with the developmental synthesis of TN by cultured rat mesangial cells was aspect of TN expression and have suggested that the documented by the following observations: (1) epithelium/mesenchyme interface of the S-shaped Northern hybridization of total PNA extracted from tubules in embryonic kidneys contains a barge mesangial cells showed two distinct species of TN amount of TN, which is known to play an important mRNA; (2) immunoblotting of the protein extracted robe in nephrogenesis (6,7). In contrast, the expres- k from the conditioned medium demonstrated four TN sion of TN in mature kidney tissue in normal or protein bands; (3) immunoblotting of the protein ex- pathologic conditions has not been adequately evab- uated (5). tracted from the mesangial cell lysate demonstrated The mesangium is a highly specialized precapibbary at least four TN protein bands; and (4) immunohisto- tissue of the renal gbomerulus, which is composed of chemical techniques identified TN within the cyto- intrinsic mesangial cells, surrounded by mesangial plasm of mesangial cells and in the surrounding ex- matrix (8). The normal mesangiab matrix is known to tracellular matrix. It was also found that normal rat be composed mainly of collagen type IV, baminin, and human glomeruli showed global, diffuse mes- fibronectin, and heparan sulfate proteoglycans, all of which have been known to be synthesized by mesangial cells in culture (8- 1 1 ). Enlargement of the ‘ Received January 22, 1993. Accepted september io, 1993. 2correspondence to Dr. L.D. Truong, Department of Pathology. MS205, The mesangial matrix, also called mesangial sclerosis, is Methodist Hospital, Houston, TX 77030. a common feature of several gbomerular diseases, 1046-6673/0410-177 1$03.00/0 regardless of etiology (8). The expanded mesangial Journal of the American society of Nephroiogy Copyright © 1994 by the American 5ociety of Nephrology matrix is traditionally thought to be in part rebated to

Journal of the American Society of Nephrology 1771 Tenascin and Mesangial Cells 1 ‘

oversynthesis of the normal components of the mes- film (Eastman Kodak Co. , Rochester, NY) at -70#{176}C. angial matrix (8-12); however, the possibility that The following probes were used for the Northern mesangial cells are also the source of other mesangial hybridization: for TN, a 2.3-kb EcoRI, mouse cDNA matrix proteins that participate in the process of (1 5) subcloned into pBluescript II; for fibronectin, a mesangial sclerosis has not been explored. 1 .2-kb EcoRI, rat cDNA fragment released from In this study, we present data to suggest that rat pRLF-l (16); for smooth muscle actin, an obigonucle- mesangial cells in culture express various species of otide probe, 5’-AGT GCT GTC CTC TTC TTC ACA mRNA for TN, that TN is secreted by mesangial cells CATA-3’, specific for smooth muscle a-actin (17,18). into the culture medium, and that TN is incorporated into the extracellular matrix. These observations may in turn imply a significant robe for TN in the Antibodies process of mesangiab sclerosis and gbomerular scle- rosis. The antibody used for the immunoblotting study was a gift from Dr. Eleanor J. Mackie (Basel Univer- sity, Basel, Switzerland). This rabbit pobycbonal an- MATERIALS AND METHODS tibody is directed against TN extracted from cultured rat embryo fibroblasts (19). The antibody used for Culture of Rat Mesangial Cells immunohistochemistry is a mouse monocbonal anti- Rat mesangiab cells were cultured according to es- body against TN extracted from a glioma cell line tablished techniques ( 1 2). Briefly, fresh kidneys were commercially available from Dako (Dakopatts, Car- obtained from anesthetized male Sprague-Dawley pinteria, CA 120]). Although these two antibodies rats (Harlan Sprague-Dawley, Houston, TX) weighing were previously characterized ( 1 9,20) and found to between 1 50 and 200 g. Individual gbomeruli (with be specific for TN, we have proceeded to confirm less than 2% contaminating tubular fragments) were their specificity by both Western blotting and absorp- isolated by compression of the renal cortical tissue tion studies. For the monocbonal antibody, the West- through nylon sieves of graded pore sizes (1 3). Gb- em blotting study included protein extracted from merubar explants from three rats were cultured in whole human kidney and human aortic wall, baminin RPMI medium (Sigma Chemical Company, St. Louis, (Gibco BRL, Gaithersburg, MD), and human plasma MO) supplemented with fetal calf serum (20%), dex- fibronectin (Sigma Chemical Company). The proce- amethasone (0.002 zg/mL), transferrin (5 zg/mL), dures for Western blotting study and protein extrac- penicillin-streptomycin (0.01 tL/mL), and HEPES tion are described below. For the polycbonal antibody, buffer (1 0 oL/mL) (all of these reagents were obtained the Western blotting study included protein extracted from Sigma Chemical Company). Initially, only epi- from isolated rat gbomeruli ( 1 3), laminin, and human thelloid cells grew out of the gbomerular explants, but plasma fibronectin. For the absorption study, the between 20 and 30 days, these cells were completely antibodies at optimal dilutions for immunohisto- replaced by spindle cells. These spindle cells were chemistry were mixed with human plasma fibronec- subcultured every 3 to 5 days and maintained their tin (15 and 50 g/mL) for 2 h at room temperature characteristics up to 80 passages; cells between pas- and were then used for immunostaining of tissue sages 4 to 1 0 were used for all studies. sections.

Northern Hybridization Study Immunoblotting Study Total RNA was isolated from cultured mesangial The immunobbotting studies were done for protein cells grown to confluence by acid guanidine extrac- extracted from cultured rat mesangial cells and from tion. The details of the method were previously de- conditioned medium. Rat mesangial cells grown to scribed (14). A similar technique was used to extract confluence were washed twice with phosphate-buff- total RNA from snap-frozen renal cortical and med- ered saline and maintained in serum-free medium ublary tissue from normal Sprague-Dawbey rats. Re- for 36 h. The conditioned medium was collected and peated efforts to obtain total RNA from gbomerular subjected to protein precipitation with 1 0% trichlo- isolates failed to produce intact RNA suitable for roacetic acid (TCA) at 4#{176}Cfor 1 h, followed by cen- Northern blotting study. trifugation at 1 2000g for 1 5 mm and repeated wash- The RNA was subjected to agarose gel electropho- ing with ethanol/ether (1 : 1 vol/vol) to remove TCA. resis and transferred to a nylon membrane (Zeta After being washed, the TCA-precipitated protein Probe; Bio-Rad Laboratories, Richmond, CA). Blots was dissolved directly in sodium dodecyl sulfate (SDS) were hybridized with a cDNA probe labeled with 132P] sample buffer (0.0625 M Tris-HC1 [pH 6.8], 2% SDS, dCTP by random primer extension (Amersham Corp., 5% 2-mercaptoethanol, 10% glycerol, 0.002% bro- Arlington Heights, IL) and exposed to Kodak X-AR5 mophenol blue). The cell monolayer was directly dis-

I 772 Volume 4 ‘ Number 10 ‘ 1994 Truong et al

solved in SDS sample buffer. Protein samples ob- RESULTS tamed from the conditioned medium and the cell Expression of the Gene for TN by Cultured Rat monolayer, respectively, were boiled for 5 mm and Mesangial Cells, Normal Rat Renal Cortex, and subjected to ebectrophoresis, with polyacrylamide gel. Medulla The acrybamide concentrations of the stacking and running gels were 4 and 6% respectively. The proce- The cultured cells displayed several features of dure of Towbin et at. was used with slight modifica- mesangial cells including absence of epithebial and tion to transfer proteins from the gel to the Zeta Probe endothebial differentiation by electron microscopy, membrane by electrobbotting (2 1). The membrane positive immunohistochemicab staining for a-smooth was immunobbotted for TN with rabbit-polycbonal muscle actin, negative staining for and Fac- antibody at a 1 :4,000 dilution. Biotinybated goat anti- tor VIII-related antigen, and the ability of these cells rabbit immunogbobubin G (IgG) was used as a second- to grow normally when the amino nucleoside of pu- ary antibody by the avidin-biotin-peroxidase complex romycmn (1 tg/mL; 1CM Pharmaceutical, Cleveland, technique (22). OH), which is known to be toxic to gbomerular epithe- hal cells, was added. The mesangiab nature of these cells was also supported by the expression of mRNA for a-smooth muscle actin and fibronectin. Immunohistochemical Study The Northern hybridization study showed that rat Immunohistochemical staining for TN was per- mesangiab cells in culture expressed two distinct TN formed for cultured rat mesangial cells and for kid- mRNA species (approximately 6.8 and 8.0 kb [Figure ney tissue from five normal Sprague-Dawbey rats. To 1 J). The renal cortex and medulla, respectively, assess whether there is human relevance of the showed two bands similar to that of mesangial cells. staining reaction, a similar study of 1 8 human renal Additionally, a third band of bow density was also biopsies showing expansion of mesangial matrix was noted for the renal medulla (Figure 1 ). The Northern also performed (three IgA nephropathy, three focal blotting findings for renal cortex and medulla seemed segmental sclerosis, three mesangial proliferative lu- to correspond with the immunohistochemical stain- pus nephritis, three focal proliferative lupus nephri- ing of the normal rat kidney tissue, where TN stain- tis, three membranoprobiferative gbomerubonephritis. ing in the cortex was limited to mesangium and ar- and three diabetic gbomerubosclerosis). Three normal teriobar wall, but there was diffuse interstitial stain- kidney samples obtained from nephrectomy speci- ing in the deep medulla (see also the results of men for renal cell carcinoma were also included in immunohistochemical staining below). this study. Mesangial cells grown to confluence were gently trypsinized and subjected to cytocentrifuge Monospecificity of the Polyclonal and Mono- preparation. Mesangial cells were also grown to clonal Antibodies Against TN monolayer on pobybysine-coated histology glass The Western blotting study using the monocbonal slides. Both the monolayer and the cytocentrifuge antibody showed two distinct bands of TN for the preparation were fixed with acetone at 4#{176}Cfor6 mm and air dried. Kidney tissues from rats and humans were fixed and paraffin embedded or snap frozen and 123 cut at 4 m. The immunostamning was done by a standard avidin-biotin-peroxidase complex tech- 8.OkB_ 6.8kB. nique (22) with the mouse monocbonab antibody de- scribed above (20). Briefly, the following steps were fr Involved: blocking with normal horse serum ( 1 / 100 dilution) to prevent nonspecific binding; primary an- tibody (1 /50 dilution); secondary antibody (polycbonal horse anti-mouse immunogbobulins, 1 / 1 00 dilution); Figure 1. Expression of the gene for TN by cultured rat avidin-biotin-peroxidase complex ( 1 / 1 00 dilution); mesangial cells (Lane 1), normal rat renal cortex (Lane 2), color development with diaminobenzidine (0.5 mg%, and normal rat renal medulla (Lane 3). Total cellular RNA is Sigma Chemical Company); counterstain with he- extracted from cultured rat mesangial cells, normal rat matoxybin or methyl green. All of the reagents for renal cortex, and normal rat renal medulla. A total of 10 g of RNA from each specimen is electrophoresed, transferred this procedure, unless otherwise indicated, were to a nylon membrane, and hybridized with a 32P-labeled from Vector Laboratory (Burlingame, CA). The nega- cDNA probe for TN. Two distinct species of mRNA for TN tive controls included replacement of the primary (approximately 6.8 and 8.0 kb) are seen for the cultured antibody, the secondary antibody, or the avidin-blo- rat mesangial cells. The renal cortex and medulla, respec- tin peroxidase complex, in different combinations, tively, show two bands similar to those of mesangial cells. with buffer. For positive control, tissue from a decu- Additionally, a third band of low density is noted for the bitus ulcer was used. renal medulla.

Journal of the American Society of Nephrology 1773 Tenascin and Mesangial Cells -. ...,, , .., ...-..,. ‘.1.,

human whole kidney and human aortic wall, respec- A MW B c tiveby. No band was seen in the baminin or fibronectin kD 320 lane (Figure 2). The Western blotting study using the 280 pobycbonab antibody showed at beast three distinct >TN 200 - 220 bands of TN for the gbomerubar isolates but no bands kD 200 for the baminin or fibronectin bane (Figure 3). The absorption by fibronectin did not induce any changes in the immunohistochemicab staining pattern by either antibody. These results indicate that both an- tibodies are specific for TN and that neither of them 116 - cross-react with fibronectin or baminin. Figure 4. Western blotting study shows TN isoforms in both cultured mesangial cell lysates and the corresponding con- TN Isoforms Was Identified in the Proteins Ex- ditioned medium. Lane A represents the control in which tracted From the Condition Medium and the the polyclonal antibody against TN is replaced by buffer. Lysates of the Cultured Mesangial Cells MW represents molecular weight markers. Proteins ex- Four distinct bands of TN isoforms were identified tracted from mesangial cell lysates and from conditioned medium are in Lanes B and C, respectively. Both Lanes B in the Western blotting study of the total proteins and C show four distinct bands of similar molecular weight extracted for the conditioned medium (Figure 4); their but of different density. Lane B also shows a faint band of molecular weights were approximately 320, 280, low molecular weight. This band most probably represents 220, and 200 kd, respectively. The blotting study of the proteolytic product related to cell protease and indeed the corresponding mesangial cell bysates also re- is not seen in the Lane C for the conditioned medium. veabed four distinct bands of TN isoforms. Although

1 2 3456 the molecular weights of these four bands were sim- kD ilar to those of the four bands seen in the conditioned medium, the density of each of the corresponding pair of bands was markedly different. Additionally,

200 another faint band of low molecular weight was also observed for the mesangial cell lysates. Although the remote possibility of this band representing a TN isoform cannot be completely ruled out, most prob- ably this band represents a proteolytic product re- bated to cell protease; the repeated presence of this Figure 2. Western blotting studies with the monoclonal an- band in the cell-associated protein extract and its tibody against TN. Lane I represents the control in which consistent absence in the protein extracted from the the primary antibody is replaced by buffer. Lane 2 repre- supernatant are consistent with this hypothesis. sents the molecular weight markers. Protein samples ex- These observations indicate that isoforms of TN exist tracted from whole human kidney, human aortic wall, lam- and that the nature of the soluble and mesangial mm, and fibronectin are in Lanes 3, 4, 5 and 6, respectively. Distinct bands of TN isoforms are seen for the kidney and cell-associated TN is probably different., aorta tissue, whereas no bands are present in the laminin and fibronectin lanes. Immunohistochemistry The mesangial cells in the cytocentrifuge prepara- 4 tions displayed strong, diffuse cytoplasmic staining for TN (Figure 5a). In the monolayer obtained at Day 7 after plating in primary culture, there was staining rN of the extracellular matrix and less cytoplasmic 200 kD staining (Figure 5c). There was global, diffuse stain- Ing of the mesangial matrix of gbomeruli from normal rats (Figure 5e). The controls for these staining re- actions, including the ones in which the primary antibody was replaced by buffer, showed negative Figure 3. Western blotting studies with the polyclonal anti- results (Figure 5b, d, and f). All three normal human body against TN. Lane I represents the molecular weight markers. Protein extracted from rat glomerular isolates, lam- kidney samples showed global, diffuse staining of mm, and fibronectin are in Lanes 2, 3, and 4, respectively. the mesangial matrix, whereas the gbomerular capil- Three bands of TN isoforms were noted in Lane 1, whereas laries were negative (Figure 5g); this pattern of stain- Lanes 2 and 3 displayed no bands. ing was similar to that of normal rat gbomeruli (Figure

1774 Volume 4’ Number 10’ 1994 .. ‘ .- .. .. . Truong et al

glomeruboscberosms, in which all of the nodular le-

sions were strongly stained (Figure 5h).

DISCUSSION Studies addressing the expression of TN in mature a kidney tissue in either normal or pathologic condi- tions are scanty and preliminary (5,6,20,23-26). These studies have alternatively suggested that, in normal mature kidney tissue, TN is either absent (6) or is focally present in rare vascular walls, some mesangial areas, stroma immediately outside Bow- man’s capsule of rare gbomerubi, and stroma outside some tubules, especially the ones in the medulla (5,20,24). In pathologic conditions, increased TN expression in the mesangium and interstitium, as- sociated with mesangiab sclerosis and interstitial fi- brosis, has been briefly mentioned (5,24). Neverthe- less, the cell types responsible for TN synthesis, fac- tors controlling this synthesis, and the significance of TN expression in kidney tissue have not been addressed. In this study, we have demonstrated that TN is synthesized by rat mesangial cells in culture and is secreted into culture medium. Supporting evidence includes the observations that ( 1 ) rat mesangial cells express mRNA for TN; (2) TN is detected in the conditioned medium by immunoblotting; (3) TN is detected in mesangial cell lysates (which contain Figure 5. Immunohistochemical staining for TN. (a) Cultured both mesangial cells and the surrounding matrix); mesangial cells are mildly trypsinized and cytocentrifuged; and (4) TN is immunohistochemicably detected in the staining for TN is noted in the cytoplasm of these cells. (b) extraceblular matrix surrounding the cultured mes- In the control study, in which the primary antibody is re- angial cells and in the cytoplasm of these cells as placed by nonimmune serum, staining is not noted in these well. Moreover, TN may be a component of normal cells. (c) Staining of mesangial cells cultured to monolayer mesangial matrix, as evidenced by the global, diffuse on glass slides shows a positive reaction for TN in the extra- mesangial staining for TN. Although repeated efforts cellular matrix. (d) Extracellular matrix is not stained in the to extract total RNA from gbomerular isolates yield control. (e) Kidney tissue from normal rat displays global degraded products unsuitable for Northern blotting and diffuse staining of the mesangial matrix for TN; the wall study, TN mRNA were clearly identified from both of an arteriole also shows positive staining. (f) The mesan- gium shows a negative reaction in the control. (g) A normal the cortex and the medulla of normal rat kidney. glomerulus from a human specimen shows global meson- suggesting that TN is indeed synthesized in vivo. It gial staining. (h) A human glomerulus with diabetic glomer- is interesting to note that different species of TN ulosclerosis shows strong staining of a mesangial nodule mRNA were found both for cultured rat mesangial (arrowheads). Other mesangial areas with milder mesan- cells and for rat kidney tissue, indicating that iso- gial enlargement also display a weak staining. Avidin-bio- forms of TN are synthesized. Isoforms of TN, identi- tin-peroxidase method with slight counterstain, x I 600 for fied in several species including chicken, human, all pictures. mouse, and rat (1 ,2, 1 5,23,27), are due to alternative splicing, whereby different insertional fibronectin- 5e). Staining of the human renal biopsies showing like homologous repeats are added or deleted from IgA nephropathy, focal segmental sclerosis, mesan- the subunit of the TN molecule (23,25-27). The func- gial proliferative lupus nephritis, focal proliferative tional consequences of TN variants containing one lupus nephritis, membranoproliferative gbomerubo- or more alternate fibronectin Type III domains are nephritis, and diabetic gbomerubosclerosis, respec- not well understood. Chiquet-Ehrismann et at. (28) tively, showed that, in each of these lesions, when- showed that the smallest form of TN bound more ever there was an expansion of the mesangial matrix, strongly to fibronectin and was preferentially incor- this matrix was stained strongly for TN. This obser- porated into a fibronectin-rich extracellular matrix vation was most pronounced in the cases of diabetic than larger TN forms that contained extra fibronec-

Journal of the American Society of Nephrology 1775 Tenascin and Mesangial Cells

tin Type III repeats. Murphy-Ullrich et at. showed In conclusion, our study documents that rat mes- that the ability of TN to reduce the number of focal angial cells in culture synthesize TN and that TN contacts in cultured endotheliab cells was dependent may be a component of the mesangial matrix in both on the inclusion of an extra fibronectin Type III re- normal and pathologic conditions. Further studies peat defined by a monocbonal antibody (29). addressing the mechanisms controlling the expres- It is also noted that the immunoblotting study of sion of TN by mesangial cells, and possibly by other proteins extracted from the conditioned medium, the cell types in the gbomerulus, may provide additional mesangial cell lysates, and the glomerular isolates, insight into the process of gbomerubosclerosis. respectively, has revealed bands of TN protein with different molecular weights. Several studies, using Western blotting, have reported up to two TN iso- ACKNOWLEDGMENTS forms in whole-kidney tissue in several species in- cluding human, mouse, and chicken (23,26,30). The This study was supported in part by funds from the Moran Foun- differential expression of TN isoforms in the cortex dation (Project #2900047) and by generous start-up funds from the Department of Pathology. Baylor College of Medicine. We thank and medulla is not addressed in any of these studies. Drs. Juan Lechago. and Mary Ostrowski for their review of the TN isoforms are also reported in other organs includ- manuscript. Ms. Melinda Sanchez’s secretarial help is appreciated. ing rat vascular wall, chicken gizzard, chicken lung, We also thank Drs. E. Mackie and P. Ekblom for providing tenascin and mouse brain ( 1 ,23,25,26). It is interesting to note polyclonal antibody and cDNA clone. respectively. from our study that, although only two species of TN mRNA were found in both mesangiab cells and kidney tissue, up to four corresponding TN isoforms were REFERENCES observed. The explanation for this discrepancy is not clear but may be rebated to several observations. 1 . Erickson HP, Bourdon MA: Tenascin: An extra- cellular matrix protein prominent in specialized First, although the two apparently distinct TN mRNA embryonic tissue and tumors. Annu Rev Cell Biol bands were identified in the Northern blot, each of 1 989;5:7 1-92. them may be in fact composed of multiple mRNA, 2. Sakakura T, Kusano I: Tenascin in tissue per- which because of their similar molecular weight, are turbation repair. Acta Pathol Jpn 1991:41: 247-258. detected as one band. Second, significant glycosyba- 3. Chiquet-Ehrismann R: What distinguishes ten- tion of TN has been demonstrated in viva and in vitro ascin from fibronectin? FASEB J 1990;4:

(1 , 1 5) and may alter the molecular weight of the TN 2598-2604. subunits, accounting for additional bands in the im- 4. Pearson CA, Pearson D, Shebahara 5, Hof- steenge J, Chiquet-Ehrismann R: Tenascin: munobbotting study (7). Third, the observation that cDNA cloning and induction by TGFB. EMBO J TN may bind to chondroitin sulfate proteoglycans, 1988;7:2677-2981. also a synthetic product of mesangial cells in culture. 5. Koukoulis GK, Gould VE, Bhattacharyya A, may have produced the results seen in our immuno- Gould JE, Howeedy AA, Virtanen I: Tenascin blotting study ( 1 ); however, the batter possibility in normal, reactive, hyperplastic and neoplastic tissues: Biologic and pathologic implications. seems remote because the reducing condition used in Hum Pathol 1991:22:636-643. the ebectrophoresis tends to dissociate such binding 6. Aufderheide E, Chiquet-Ehrismann R, Ekblom interactions. P: Epithelial-mesenchymab interactions in the Mesanglab cell proliferation and mesangial scle- developing kidney lead to expression of tenascin of the mesenchyme. J Cell Biol 1987;105: rosis are seen in several pathologic processes invobv- 599-608. ing the kidney and may play an important robe in 7. Crossin KL, Hoffman S, Grumet M, Thiery J-P, gbomeruboscberosis (8). It has been documented that, Edelman GM: Site-restricted expression of cyto- despite minor quantitative differences seen in var- tactin during development of the chicken em- bus types of gbomerubar diseases, all components of bryo. J Cell Blob 1986;102:1917-1930. 8. Kashgarian M: Mesangium and gbomerular dis- normal mesangiab matrix participate in the process ease. Lab Invest 1985;52:569-571. of mesangial sclerosis (9- 1 1 ). Furthermore, mesan- 9. Oomura A, Nakamura T, Arakawa M, Ooosh- gial cells in culture are known to synthesize these ima A, Isemura M: Alterations in the extraceb- components (9- 1 1 ). Our study, documenting the syn- bular matrix components in human gbomerular diseases. Virchows Arch A Pathob Anat 1989; thesis of TN by rat mesangial cells in culture, sug- 415:151-159. gests that TN may also play a significant role in the 1 0. Morel-Maroger Striker M, Killen PD, Chi E, Stri- process of mesangial sclerosis and gbomerubosclero- ker GE: The composition of gbomeruboscberosis. sis. The human relevance of this observation is also I. Studies in focal sclerosis, crescentic gbomeru- supported by our finding that human renal biopsies bonephritis and membranoproliferative gbomer- ubonephritis. Lab Invest 1 984;5 1 : 1 8 1 -192. show strong staining for TN in areas with expansion 1 1 . Bergijk EC, Munaut C, Baelde JJ, et at.: A of the mesangiab matrix, especially in diabetic gb- histologic study of the extraceblubar matrix dur- merubosclerosis. ing the development of gbomerubosclerosis in mu-

1776 Volume 4’ Number 10’ 1994 Truong et al

rine chronic graft-versus-host disease. Am J Pa- 4350-4354. thob 1992:140:1147-1156. 22. Hsu SM, Raine L, Fanger H: A comparative 1 2. Ayo SH, Radnik RA, Garoni JA, Glass WF, study of the peroxidase-antiperoxidase method Kreisberg JI: High glucose causes an increase and an avidin-biotmn complex method for study- in extracelbular matrix proteins in cultured ing pobypeptide hormones with radloimmuno- mesangial cells. Am J Pathob 1 990; 136: assay antibodies. Am J Clin Pathob 1981;75: 1 339-r348. 734-738. 13. Misra RP: Isolation of gbomerubi from mamma- 23. Prieto AL, Jones FS, Cunningham BA, Crossin han kidneys by graded sieving. Am J Cbin Pathob KL, Edelman GM: Locabization during develop- 1972;58:135-139. ment of alternatively spliced forms of cytotactin 14. Majesky MW, Reidy MA, Bowen-Pope DF, Hart mRNA by in situ hybridization. J Cell Blob 1990; CE, Wilcox JN, Schwartz SM: PDGF bigand and 111:685-698. receptor gene expression during repair of arterial 24. Gould VE, Martinez-Lacabe V. Virtanen I, Sah- injury. J Cell Biol 1990;! 11:2149-2158. un KM, Schwartz MM: Differential distribution 1 5. Weller A, Beck S, Ekblom P: Amino acid se- of tenascin and cellular in acute quence of mouse tenascin and differential and chronic renal abbograft rejection. Lab Invest expression of two tenascin isoforms during em- 1992:67:71-79. bryogenesis. J Cell Blob 1991112:355-362. 25. Saga Y, Tsukamoto T, Jing N, Kusakabe M, 1 6. Schwarzbaur JE, Tamkun JW, Lemishka IR, Sakakura T: Murine tenascin: cDNA cloning, Hynes RO: Three different fibronectin mRNA’s structure and temporal expression of isoforms. arise by alternative splicing within the coding Gene 1991;104:177-l85. region. Cell 1983;35:421-431. 26. Crossin KL, Hoffman S, Grumet M, Thiery J-P, 1 7. Ueyana U, Hamada H, Battula N, Kakunaga T: Edelman GM: Site-restricted expression of cyto- Structure of a human smooth muscle actin gene tactin during development of the chicken em- (aortic type) with a unique intron site. Mob Cell bryo. J Cell Blob 1986;102:1917-1930. Blob 1 984;4: 1073-1078. 27. Aufderheide E, Ekblom P: Tenascin during gut 18. Tso JY, Sun X-H, Kao TH, Reece KS, Wu R: development: Appearance in the mesenchyme, Isolation and characterization of rat and human shift in molecular forms, and dependence on glycerabdehyde-3-phosphate dehydrogenase epithebial-mesenchymab interactions. J Cell Blob cDNAs: Genomic complexity and molecular evo- 1988; 107:2341-2349. lution of the gene. Nucleic Acids Res 1 985; 13: 28. Chiquet-Ehrismann R, Matsuoka Y, Hofer U, 2485-2502. Spring J, Bernasconi C, Chiquet M: Tenascin 1 9. Macide EJ, Halfter W, Liverami D: Induction of variants: Differential binding to fibronectin and tenascin in healing wounds. J Cell Blob 1988; distinct distribution in cell cultures and tissues. 107:2757-2767. Cell Regul 1991;2:927-938. 20. Bourdon MA, Wikstrand CJ, Furthmayr H, 29. Murphy-Ullrich JE, Lightner VA, Aukhil I, Yan Matthews TJ, Bigner DD: Human gbioma-mes- YZ, Erickson HP, Hook M: Focal adhesion mnteg- enchymab extracebbubar matrix antigen defined rity is downregulated by the alternatively spliced by monocbonal antibody. Cancer Res 1983:43: domain of human tenascin. J Cell Blob 1991; 2796-2805. 115:1127-1136. 21. Towbin H, Staehelin T, Gordon J: Ebectropho- 30. Ventiniiglia B, Wikstrand CJ, Ostrowski LE, retic transfer of proteins from pobyacrybamide Bourdon MA, Lightner VA, Bigner DD: Tenascin gels to nitrocebbubose sheets: Procedure and some expression in human glioma cell lines and normal applications. Proc Natb Acad Sd USA 1979:9: tissues. J Neuroimmunob 1992:36:41-55.

Journal of the American Society of Nephrology 1777