Glomerulonephritis - Europe PMC

American Journal of Pathology, Vol. 147, No. 5, November 1995 Copynight © American Societyfor Investigative Pathology

Expression of Alternatively Spliced Fibronectin Variants During Remodeling in Proliferative

Glomerulonephritis

Jeffrey L. Barnes,*t Ernestine S. Torres,* Ronda J. Mitchell,* and John H. Peterst From the Division of Nephrology,* Department of Medicine, The University of Texas Health Science Center and The Audie Murphy Memorial Veterans Medical Hospitalt San Antonio, Texas; and Division of Pulmonary Medicine,t Cedars-Sinai Medical Center, Los Angeles, California

Fibronectin (Fn) plays an important role in tissue remodeling during embryogenesis, wound repair, and vascular disease, and is thought to regulate celular processes such as ceUl adhesion, migration, proliferation, and differentiation through specialized domains within the molecule. In addition, Fn can be alternatively spliced at three regions: extradomains EIIIA, EIIIB, and a variable segment V, potentially giving rise to functionally distinct variants of the molecule. We have previously shown a sequential expression of cellular Fn first by platelets, folowed by macrophages, then mesangial ceUs in habu snake venom-induced proliferative glomerulonephritis (Am JPathol 145: 585-597, 1994). These studies examined the ceUular sources and glomerular localization of Fn in general but did not distinguish between the various alternatively spliced isoforms. In this study, we examine by in situ hybridization and immunohistochemistry the temporal expression and celular sources of EIIAI, EIIIB, and Vin a model ofproliferativeglonerulonephritis that has cell migration, proliferation, and extracelular matrix synthesis as features of tissue remodeling. Macrophages were theftrst cells to express Fn mRNA showing an EIIJA +, EIIIB-, and v95+ pattern beginning at 8 hours after habu snake venom injection. Migrating mesangial cells at the margins of early lesions (8 and 24 hours) did not overexpress mRNA encoding these Fn variants, but immunofluorescence microscopy revealed V95 and EIIlAprotein at the margins of lesions. EIIIB was absent in lesions at this time. At 48 hours and peaking at 72 hours after habu

snake venom injection, mesangial ceUs in central aspects ofglomear lesions expressed abundant mRNA and protein for V95 and EIIL4 EIIIB mRNA and protein was slight in the mesangium at these times. Parietal epithelial cells, particularly adjacent to glomnerwlar lesions, also expressed abundant mRNA and protein for aUl three variants throughout the course of the disease, beginning at 24 hours after habu snake venom injection. Expression of mRNA and protein for all three isoforms declined by 2 weeks after habu snake venom

injection. These studies show that migrating mesangial celUs do not require their own synthesis of Fn and suggest that they might rely on exogenous sources ofFn, particularly V95+ and EII+'forms. Commencement of enhanced expression of EIIJA and EIIIB mRNA and protein by resident gkonerular cels coincided with the temporal course of ceUl proliferation, acquisition of a-smooth muscle cell actin phenotype, and matrix synthesis, suggesting that Fn isoforms have specijfcfunctions during the course of glomerular remodeling. (Am J Pathol 1995, 147:1361-1371)

Fibronectins (Fn) are a-family of large glycoproteins that have important roles in tissue remodeling during embryogenesis and wound healing, and in various diseases including atherosclerosis, pulmonary fibrosis, and glomerulosclerosis.1-3 Fn influences a variety of cellular behaviors including cell adhesion, migration, proliferation, and differentiation, which are believed to be mediated through several domains located at specific sites within the molecule. These Supported by grants from the National Institutes of Health (National Institute of Diabetes and Digestive and Kidney Diseases, DK38758) and the Office of Research and Development, Medical Research Service, Department of Veterans Affairs. Accepted for publication August 2, 1995. Address reprint requests to Jeffrey L Barnes, PhD, Department of Medicine, Division of Nephrology, The University of Texas Health Science Center, 7703 Floyd Curl Drive, San Antonio, TX 78284

1361

1362

Barnes et al

AJP November 1995, Vol. 14 7, No. 5

sites include collagen and heparin binding domains and a cell binding domain (arginine-glycine-asparagine-serine, RGDS), which works through an integrin receptor (a5fli).1'3 In addition, Fn has three sites of alternative splicing: extra domains EIIIA and EIIIB, and a variable region V.4`6 EIIIA and EIIIB can be either included or excluded, and the V region can be partially or completely excluded. In the rat, up to 12 different combinations of alternatively spliced variants are possible.5'6 The patterns of splicing can be cell type-specific, but their functions remain unknown. However, the close proximity of the EIIIA and EIIIB regions to the cell binding domain and the non-RGDS cell adhesion characteristics of the V region suggest functionally distinct forms of translated protein.5 Indeed, EIIIA and EIIIB variants are included in mRNA that is overly expressed in embryonic development7 8 and at the margins of healing wounds,9'10 whereas they are often spliced out of tissue-specific areas in adult tissues, suggesting that they may have important roles in tissue remodeling 3,4,7,8 Previous studies in our laboratory have characterized a model of proliferative glomerulonephritis, induced by habu snake venom, by a distinct temporal course involving mesangial cell migration,11'12 proliferation,' 113 and extracellular matrix synthesis 12.13 We have also shown a sequential expression of cellular Fn in the habu snake venom model first by platelets and macrophages followed by mesangial cells,12 which closely paralleled the temporal expression of Fn by macrophages and resident fibroblasts in cutaneous wound healing.9'10 Our studies examined the cellular sources and glomerular localization of Fn in general, but did not distinguish between the various alternatively spliced subsets of Fn. Because alternatively spliced patterns of Fn might be associated with specific cell functions during the course of the disease, experiments were performed in the habu snake venom model to examine the temporal and spatial expression and cellular sources of mRNA-encoding extradomains EIIIA and EIIIB, and the variable domain V95 of Fn and their translated proteins.

Materials and Methods Induction of Glomerulonephritis Glomerular lesions were induced in male SpragueDawley rats (Charles River, Wilmington, MA) weighing 200 to 250 g as previously reported.12 Briefly, the rats were unilaterally nephrectomized, to increase the incidence of subsequent glomerular lesions, and

24 hours later were injected intravenously with habu snake venom (Trimeresurus flavoviridis, Sigma Chemical Co., St. Louis, MO) at a dose of 3 mg/kg. At 8 (n = 4), 24 (n = 4), 48 (n = 4), and 72 (n = 4) hours, and 2 weeks (n = 4) after injection of habu snake venom, the rats were sacrificed and slices of renal cortex obtained and fixed in 10% buffered formalin and processed for routine light microscopical evaluation of hematoxylin and eosin (H&E)-stained tissue sections. Additional slices of cortex were snap frozen in liquid nitrogen for subsequent immunohistochemistry and for in situ hybridization (see below).

Characterization of Cell Types Within Glomerular Lesions Cell types within glomerular lesions were identified by immunofluorescence histochemistry using phenotypic markers as previously described1 1-4: Mesangial cells were identified by the expression of desmin (clone D33), Thy-1, (clone OX7), and a-smooth muscle cell actin (clone 1A4) using mouse monoclonal antibodies obtained from Dako Corp. (Santa Barbara, CA), Accurate Chemical & Scientific Corp. (Westbury, NY), and Sigma Chemical Co., respectively. Monocytes and macrophages were identified by a mouse monoclonal antibody to rat ED-1 (Bioproducts for Science, Inc., Indianapolis, IN). Primary antibodies were detected by indirect techniques using fluorescein isothiocyanate (FITC)labeled monoclonal rat anti-mouse or anti-rabbit immunoglobulin G (IgG; Zymed Laboratories, Inc., San Francisco, CA) as a second antibody. Controls consisted of nonimmune IgG or diluent (0.02 mol/L phosphate-buffered saline, pH 7.4, 0.1% bovine serum albumin, PBS-BSA) without primary antibody. Macrophages were also identified and verified in glomerular lesions by in situ hybridization for the expression of lysozyme mRNA, a marker for macrophages during wound healing10 and previously used to identify macrophages in glomerular lesions in the habu snake venom model (see below12).

Immunohistochemical Localization of Fn Isoform Protein Glomerular localization of Fn isoforms was assessed using three segment-specific antibodies. Anti-Fn EIIIA A mouse monoclonal antibody (IgM) against cellular Fn (clone FN-3E2, Sigma Chemical Co.). This anti-

Fibronectin Variants in Glomerular Disease

1363


body recognizes cellular Fn but not plasma (p) Fn as indicated by the lack of ability of antibody adsorbed with purified rat pFn to eliminate or reduce staining within glomerular lesions. Also, the antibody specifically recognizes only full length recombinant rat Fns15 that contain the EIIIA segment by Western blot analysis (JH Peters and RO Hynes, manuscript in preparation). Anti-EIIIB Rabbit antiserum was raised against an EIIIB-glutathione S-transferase fusion protein, and antibody was immunopurified by passage of the antiserum over an EIIIB-maltose binding protein-Sepharose 4B affinity column as described.16

In Situ Expression of Fn Isoforms Plasmid cDNA cDNA probes containing a region common to all known forms of Fn mRNA (Fn-C) or recognizing the alternatively spliced regions EIIIA, EIIIB, or V95 (generously provided by Dr. Richard 0. Hynes, Massachusetts Institute of Technology)9 were used for the generation of labeled riboprobes to detect cellular localization of mRNA encoding Fn variants within glomerular lesions. In situ phenotypic identification and verification of macrophages within lesions was performed using a 438-bp cDNA fragment of murine lysozyme cDNA in pGEM-3 (generously provided by Dr. Livingston Van De Water, Harvard University).10 All experiments were performed simultaneously with labeled sense riboprobe as negative controls.

Anti-V95

Immunopurified antibody to the V95 region of Fn (generously provided by Dr. Richard 0. Hynes, Massachusetts Institute of Technology, Cambridge, MA) was purified from rabbit anti-V95 fusion protein antiserum4 by passage over a rat plasma Fn-Sepharose 4B affinity column and elution at low pH as described (JH Peters and RO Hynes, manuscript in preparation). Frozen sections were incubated for 1 hour with each isoform-specific antibody followed by three 10minute washes with PBS-BSA and a 1-hour incubation with the appropriate FITC-conjugated second antibody (rat monoclonal anti-mouse or rabbit IgG, Zymed Laboratories, Inc., South San Francisco, CA). Before incubation with EIIIB antibody, sections were acetone-fixed, washed with PBS-BSA, and incubated at 37°C overnight with 50,000 units/ml N-glycanase (peptide: N-glycosidase F, New England Biolabs, Beverly, MA) to deglycosylate the EIIIB protein and unmask antigenic epitopes.16 Sections receiving EIIIA or V95 antibodies were pretreated with buffer without glycanase under the same conditions as described for EIIIB antibody above. Controls consisted of nonimmune mouse or rabbit IgG or PBS in place of primary antibody. Specificity of the antibodies was determined by using primary antibody adsorbed with their respective fusion protein before immunostaining. Specificity of EIIIB antibody was also determined by comparing sections incubated overnight in buffer with or without glycanase. After a final wash with PBS-BSA, the sections were coverslipped, examined, and photographed utilizing an Olympus AH-2 research microscope (Olympus Instruments, Dallas, TX) equipped with epifluorescence.

Preparation of Riboprobes Linearized cDNA was transcribed in vitro using a Riboprobe system 11 kit (Promega, Madison, WI) according to the manufacturer's instructions. Either SP6 or T7 RNA polymerase and [35S]uridine-5'-(athio)-triphosphate (1300 Ci/mMol, New England Nuclear, Boston, MA) were included in the reaction mixture to generate 35S-labeled anti-sense and sense riboprobes. The reaction mixture was incubated for 60 minutes at 400C, the DNA template removed by digestion with 0.5 U RNAse free DNAse, and unincorporated nucleotides removed by phenol/ chloroform extraction and ethanol precipitation. RNA probes (activity -4 x 106 cpm/li) were stored at -700C and used within 3 days.

Tissue Preparation Frozen sections (6 gm) were cut and collected onto aminosilane-glutaraldehyde treated slides, then fixed for 20 minutes in 4% paraformaldehyde in 0.01 mol/L PBS, pH 7.4. The sections were washed twice in PBS, dehydrated through a graded series of ethanols, air-dried, and immediately used for in situ hybridization.

Tissue Hybridization In situ hybridization procedures were performed as previously described, involving prehybridization, hybridization, and removal of nonspecifically bound probe.12 Prehybridization steps included treatment with 0.2 N HCI, proteinase K (1 ,ug/ml), and acetic anhydride to block background and enhance probe

1364

Barnes et al


penetration. Twenty-five ,uL of hybridization mixture containing 50% formamide, 10% dextran sulfate, 10 mmol/L dithiothreitol, 0.1 mol/L Tris-HCI pH 7.5, 0.1 mol/L NaPO4, 0.3 mol/L NaCI, 50 mmol/L EDTA, 1X Denhardt's solution, 0.2 mg/ml yeast tRNA, and 3 x 105 cpm of 35S-labeled riboprobe was applied to each section and covered with a siliconized coverslip. Hybridizations were performed in a sealed humid chamber for 18 hours at 500C. Excess probe was removed by washing in TE buffer, treated with RNAse A to decrease nonspecific background activity and rinsed in 2X SSC. Sections were dehydrated in graded ethanols, air-dried, and immersed in the dark in Kodak NTB-2 photographic emulsion (Eastman-Kodak, Rochester, NY). After air-drying, the sections were exposed for a 2 to 3 weeks at 40C. The emulsion was developed and sections stained with H&E for subsequent bright- and darkfield light microscopical analysis.

Results Morphology Habu snake venom induced a proliferative glomerulonephritis as described previously.11-14 The lesions were characterized by mesangiolysis and development of microaneurysms filled with platelet aggregates, leukocytes, erythrocytes, and plasma proteins at 8 and 24 hours. By 48 hours after habu snake venom injection, most lesions contained numerous ovoid and round cells in addition to an inflammatory infiltrate. At 72 hours after habu snake venom injection, lesions were characterized by a confluent cellular mass forming micronodules. Micronodules became smaller in size with an expanded space between cells at 2 weeks after habu snake venom

injection.

Cell Types within Glomerular Lesions The progression of habu snake venom-induced glomerular lesions followed the same pattern as previously described.11-14 Platelets and platelet aggregates within glomerular lesions, identified by light microscopy, were prominent at 8 and 24 hours after habu snake venom injection. Platelets were less conspicuous at 48 hours and not observed at 72 hours and 2 weeks after habu snake venom injection. Macrophages, identified by ED-1 staining by immunohistochemistry and expression of lysozyme mRNA by in situ hybridization (Figure 1), were observed within central aspects of glomerular lesions, first becoming apparent at 8 hours, then increasing in number until

Il~ Figure 1. Verification of macrophages within lesions by in situ hybridization using 35S-labeled anti-sense riboprobe to detect lysozyme mRNA as aphenotypic marker. Glomerularlesions 24 hours after habu snake venom injection (a) show numerous macrophages (silvergrains, arrows) within central aspects of microaneurysms. The number of macrophages declined in glomerular lesions at 72 hours (b) and 2 weeks after habu snake venom injection. Dark field reflected light image counterstained with H&E. x 170.

peaking at 48 hours after habu snake venom injection. The macrophages were observed almost exclusively within the central aspects of microaneurysms (Figure la) and occasionally at the periphery of lesions including the glomerular tuft-lesion interface. In advanced lesions at 72 hours after habu snake venom injection, macrophages declined (Figure 1 b). Only isolated macrophages were detected at 2 weeks after habu snake venom injection. As reported previously,12 macrophage numbers based on ED-1 positive cell counts averaged 0.68 + 0.1 (SEM), 2.5 ± 0.3, 3.7 ± 0.3, 1.9 ± 0.1, and 0.7 ± 0.05 at 8, 24, 48, and 72 hours and 2 weeks after habu snake venom injection, respectively. Mesangial cells, identified by phenotypic staining with desmin, Thy 1.1, and a-smooth muscle actin antibody, showed the same pattern of distribution as already described. Mesangial cells were absent in early 8-hour lesions, except occasional focal cells at the interface between lesions and intact capillary tufts.1 1-13 Glomeruli showed desmin and Thy-1 positive mesangial cells at the interface and/or along the entire periphery of nearly all lesions by 24 hours after habu snake venom injection (Figure 2a). Desmin, Thy-1, or a-smooth muscle actin-positive mesangial cells were not observed in the central aspects of microaneurysms at 8 and 24 hours after habu snake venom injection. More advanced lesions at 48 hours after injection showed numerous mesangial cells within the central aspects of microaneurysms that stained for desmin or Thy-1. By 72 hours after habu snake venom injection, the lesions were made up primarily of desmin- and Thy-i-staining cells. Marginating mesangial cells in early lesions were negative for a-smooth muscle cell actin; however, they acquired staining for this marker at 48 and 72 hours after habu snake venom injection (Figure 2, b and c) as previously described.11-13 Epithelial cells lining Bowman's capsule also acquired

1365 Fibronectin Variants in Glomerular Disease AJP November 1995, Vol, 147, No. 5

Figure 2. Immunofluorescence micrographs of mesangial cell expression of Thy-1 (a) and a-smooth muscle cell actin (b and C) over the course of habu snake venom-induced glomerular injury. Tby-1 antigen was expressed in the mesangium of the intact glomerular capillary tuft (G) and at the margins of lesions (arrows) 94 noun; Anpirc ajter 17ftor nam h,,7/w snaw iwnsm. injewturt iniprtionn tcty. (aq) infflkP venumt ---i

Mesangial cells did not express a-smooth muscle cell actin in early lesions, but acquired expression within lesions at 48 (b) and 72 (c) hours after habu snake venom injection (arrows). Proliferation of mesangial cells within glomerular lesions led to the formation of micronodules filled with a confluent mass of cells that stained for a-smooth muscle cell actin (c) and Thy-1 antigen. Parietal epithelial cells in Bowman's capsule also acquired expression of a-smooth muscle cell actin (small arrows). Controls incubated with nonimmune IgG in place ofprimary antibody were negative in a section 24 hours after habu snake venom injection (d). Sections stained for indirect immunofluorescence using primary antibody followed by FITC-conjugated secondary antibody and viewed by epifluores-- cence microscopy. X200.

I

staining for a-smooth muscle cell actin at 48 and 72 hours (Figure 2, b and c) and 2 weeks after habu snake venom injection. Microaneurysms and micronodules were determined to be devoid of glomerular epithelial and endothelial cells on the basis of the lack of staining for phenotypic markers of these cell types.'1-13 At 2 weeks after habu snake venom injection, mesangial cells within glomerular lesions continued to stain for desmin, Thy-1, and a-smooth muscle cell actin, although at a lower intensity than at previous time points. Controls incubated with nonimmune IgG (Figure 2d) or PBS in place of primary antibody were negative.

Expression of Fn Isoforms by Immunohistochemistry Detection of Fn was similar to that previously described for immunoperoxidase microscopy12 except the resolution and intensity of staining was less diffuse and showed a more defined localization of the isoforms. At 8 and 24 hours after habu snake venom injection, protein for EIIIA and V95 was detected at the margins of lesions (Figure 3, a and b) consistent with the position of Thy-1+ cells. Staining for EIIIA and V95 was minimal in the central aspects of glomerular microaneurysms associated with platelets and macrophages. These

observations are in agreement with the localization of EIIIA by immunoperoxidase as previously reported, except that immunofluorescence was less sensitive with weaker staining. EIIIB was not detected in microaneurysms at these times (Figure 3c). At 48 and 72 hours after habu snake venom injection, staining for EIIIA and V95 within the central aspects of lesions became more intense, peaking at 72 hours (Figure 3, d and e), and was associated with mesangial cells as the primary cell population in lesions at this time. Staining for EIIIB was faint within the micronodules, but more intense in the parietal epithelium (Figure 3f). As with in situ hybridization for mRNA (see below), staining of all three isoform proteins was much less intense at 2 weeks after habu snake venom injection, when EIIIA and V95 faintly stained the mesangium in glomerular lesions and EIIIB was not detected.

Expression of Fn Variants by in Situ Hybridization Fn-C

Expression of Fn mRNA in general using a probe to Fn-C, which recognizes all variants of Fn, was as described previously12 and not illustrated here. Early (8-hour) lesions were generally devoid of cells ex-

1366

Barnes et al

AJP November 1995, Vol. 147, No. 5

i

I

I

Figure 3. Immunofluorescence localization of Fn variants V95 (a and d), EHIA (b and e), and EIIIB (c and f) within glomerular lesions at 24 (a-c) and 72 (-f) hours after injection of habu snake venom. Fn vanIants V95 and EIJA (a and b) localized in microaneurysms at the margins of lesions associated with mesangial cells andfaintly in central aspects of lesions associated with platelets and macrophages 24 hours after habu snake venom injection. EIIIB (C) was absent from lesions at this time. Staining for Fn variants V95 and EIIIA (d and e) was abundant, and EIIIB (f ) was faint in glomerular micronodules 72 hours after habu snake venom injection u'hen the population of cells within lesions were primarily mcesangial (arrows). All three variants of Fn localized to parietal epithelial cells lininig Bowman's capsule (wide arrows). X 200.

pressing Fn mRNA; however, occasional cells within the central aspects of microaneurysms expressed Fn-C mRNA. These cells were identified as mononuclear cells or macrophages (as above) based on the temporal location of macrophage markers (ED-1 antigen and lysozyme mRNA) and the absence of resident glomerular cells in this region at this time. Expression of Fn-C transcript increased at 24 hours after habu snake venom injection and was associated with the increase in the number of macrophages within lesions. Few cells at the margins of lesions expressed Fn message above that observed in the mesangium in intact glomerular capillary tufts. At 48 hours after habu snake venom injection, the number of macrophages and mesangial cells within lesions increased as did the expression of Fn-C mRNA, correlating with the increase in translated protein. Cells at the interface between the intact capillary tuft and glomerular lesions at this time frequently expressed Fn mRNA. The most abundant Fn-C message was detected 72 hours after habu snake venom injection, at a time when lesions were filled with mesangial cells and when expression of

cellular Fn protein was maximal. At 2 weeks after habu snake venom injection, expression of mRNA encoding Fn-C diminished, but remained slightly elevated above controls. Parietal epithelial cells adjacent to glomerular lesions expressed abundant Fn-C message beginning at 24 hours after habu snake venom injection.

Fn Isoforms (EIIIA, EIIIB, and V95) The three cell types in glomerular lesions that expressed Fn variants were macrophages, mesangial cells, and parietal epithelial cells. Expression of mRNA encoding alternatively spliced Fn isoforms in glomerular lesions showed a differential expression depending on the time of study and cell type (Figure 4). EIIIA and V95 mRNAs were expressed by all three cell types with patterns similar to the localization of Fn-C mRNA as described above. The first cell type in glomerular lesions to express EIIIA and V95 mRNA were macrophages based on the spatial expression of phenotypic markers of macrophages, ED-1 antigen, and lysozyme mRNA by immunofluo-


1367 AJP November 1995, Vol. 14 7, No. 5

,f:

I

I s:

i

II !

i

Figure 4. Cellular expression of mRNA encoding Fn splicing variants V95 (a, d, and g), ERIMA (b, e, and h), and EIIIB (c, f, and i) 24 (a-), 48 (df), and 72 (g-i) hours after habu snake venom-induced glomerulonephritis. Cells (short arrows) determined to be macrophages ( by ED-1 surface antigen and lsozyme mRNA) uwithin the cenztral aspects ofglomerular lesions at 24 hours after habu snake venom injection expressed mRNA encoding V95 (a) and EIIIA (b). mRNA enicoding EIIIB was nzot expressed above background in cells in central aspects ofglomerular lesions at this time C. Also, cells at the margins of ear/y lesions, determined to be mesangial cells, did not usuallv overexpress mRNA for any of the Fn variants. However, cells in the central aspects of microaneurysms and in micronodules (long arrows) at 48 (df-) and 72 (gi) hours after habu snake venom injection expressed all three splicing variants. Macrophages uerepromninent uwithin lesions at 24 and 48 hours after habu snake venom injection, but declined in number b1 72 hours after habu snake venom injection at a time when mesangial cells were the predominant cell tjpe (identified by Thy-i and a-smooth muscle actin; see Figure 2). Panietal epithelial cells lininzg Bouman 's capsule (arrowheads) also expressed V-95, EIIIA, and FIIIB mRATA at 24, 48, and 72 hou.rs after habu snake venom injection. Expressioni of FIIB mRNA u'as more intense in parietal epithelial cells than in mcsangial cells (f,i). G, intact glomenrlar capillary tuft. Darkfield images of autoradiographs counterstained uith H&E. X 200.

rescence and in situ hybridization, respectively. mRNA for these two variants (Figure 4, a and b) was first observed in a small population of cells located in central aspects of microaneurysms at 8 hours (not shown), increasing in numbers at 24 and 48 hours after habu snake venom injection. On the other hand, the majority of macrophages did not express EIIIB (Figure 4c), and when detected, the message was

weak. Mesangial cells in glomerular capillary tufts of normal control glomeruli, and in unaffected regions of diseased glomeruli in habu snake venom injected rats expressed small amounts of Fn-C. However, mesangial expression of EIIIA, EIIIB, and V95 was difficult to detect over background. Expression of all three isoforms was generally not detected in marginating mesangial cells at the interface and periphery

1368

Barnes et al


of glomerular lesions at 8 and 24 hours; however, occasional focal cells did express V95 and EIIIA above that observed in cells in the intact glomerular capillary tuft. mRNA encoding V95 and EIIIA was highly expressed in cells in the central aspects of glomerular lesions 48 and 72 hours after habu snake venom injection (Figure 4, d, e, g, and h) with a peak in expression of mRNA for both variants at 72 hours after habu snake venom injection, corresponding to the shift in cell population to mesangial cells. Parietal epithelial cells in Bowman's capsule, particularly at locations adjacent to glomerular lesions, also expressed mRNA for EIIIA, and for V95 beginning at 24 hours after habu snake venom injection (Figure 4 a, b, d, e, g, and h). Cells in the central aspects of glomerular lesions at 48 hours and in micronodules at 72 hours after habu snake venom injection also expressed EIIIB mRNA (Figure 4, f and i), with most abundant message occurring at 72 hours after habu snake venom injection; however, the level of staining was less than that observed for V95 and EIIIA. On the other hand, parietal cells lining Bowman's capsule expressed abundant EIIIB mRNA at 48, and 72 hours after habu snake venom injection (Figure 4 c, f, and i). Parietal epithelial cell expression of EIIIB mRNA and protein was higher than in mesangial cells in micronodules. Expression of all three isoforms declined by 2 weeks after habu snake venom injection.

Discussion This study shows that Fn isoforms are differentially expressed by platelets, macrophages, mesangial cells, and parietal epithelial cells over the course of habu snake venom-induced glomerular disease. EIIIA and V95 protein in early glomerular lesions was determined to be derived from platelets and macrophages based on immunohistochemistry12 and the abundant expression of EIIIA and V95, but not EIIIB mRNA in macrophages by in situ hybridization. At later stages of the lesions (48 and 72 hours), mesangial cells replaced macrophages and became the predominant cell type associated with Fn synthesis. Cells within lesions at these times expressed abundant mRNA for EIIIA and V95 and to a lesser extent EIIIB, observations that coincided with staining intensity of their respective proteins. Parietal epithelial cells, particularly at sites adjacent to lesions, expressed mRNA and protein for all three variants at 24, 48, and 72 hours after habu snake venom injection, suggesting that soluble factors released from the vicinity of glomerular lesions induce epithelial cell synthesis of these Fn isoforms.

The functions of the various spliced regions of Fn have not been determined; however, EIIIA and EIIIB have been referred to as "embryonic" because they are expressed in tissue type-specific areas associated with cell migration and proliferation during embryogenesis and lost in adult tissue.7'8 Cell migration and proliferation are also an integral part of wound healing, and the recapitulation of expression of EIIIA and EIIIB variants in localized areas of cutaneous injury suggests that these embryonic forms might play a role in cell behavior during remodeling.9'10 In addition, the terminal segment of V (V25 is also referred to as IIICS for human protein) has been shown to have a role in cell migration and may have a role in tissue remodeling.17 We find that mesangial cells in normal glomeruli and in intact glomerular capillary tufts of affected glomeruli, express only trace levels of Fn-C, EIIIA, EIIIB, and V95. These results are in agreement with immunohistochemical studies indicating that mesangial cells in normal adult human glomeruli synthesize low levels of EIIIA (EDAcFn), and express no detectable EIIIB (EDBcFn) protein.18'19 However, during embryogenesis, the mesangium expresses high levels of EDAcFn and EDBcFn protein18; thus, in glomerular disease, as in wound healing, the sequential expression of these two Fn variants appears to recapitulate embryonic expression. Our study indicated that mesangial cells at the margins of early glomerular lesions (at 8 and 24 hours) did not overexpress any form of Fn mRNA. However, a strong localization of V95 and EIIIA protein was observed at the margins of early glomerular lesions, suggesting that these cells are cued to move toward or across tracts of Fn deposited or secreted from exogenous sources. The general lack of mesangial cell expression of mRNA encoding these isoforms and presence of platelets and macrophages in central aspects of lesions at these times suggest that Fn protein secreted from platelets and macrophages may partition to these sites. Platelets and macrophages contain EIIIA- and EIIIB-spliced Fns, which are typically present in cellular forms of Fn and not associated with pFn derived from hepatocytes.6' 16'20 Thus, soluble Fn released from platelets and macrophages or in plasma may deposit as an insoluble material at the margins of lesions similar to that described in the walls of blood vessels after intravenous injection,21 potentially constituting a target stimulus or concentration gradient for migration. Fn is known to stimulate migration of a variety of cell types including fibroblasts,2223 smooth muscle cells,24 neural crest cells,25'26 keratinocytes,27'28 and several others.3 We have also shown that mesangial cells migrate in response to platelet releas-


1369


ate and to platelet Fn in vitro.29 Our studies showed a potent migratory response of mesangial cells toward gradients of platelet Fn and platelet releasate that could be blocked by inhibiting an integrin receptor for Fn with RGDS tetrapeptide, indicating that the primary component in platelet releasate responsible for cell migration was cellular Fn29 or other RGDScontaining proteins. It should be noted that glomerular lesions throughout the course of habu snake venom-induced glomerulonephritis are filled with platelet secretory products14; and other components of platelet releasate might contribute toward mesangial cell migration in glomerular lesions in vivo. For example, PDGF is present in microaneurysms throughout the course of the disease13 and also has a modest effect on mesangial cell migration in vitro.30 Other factors present in platelets including TGF-a, TGF-,B1, EGF, and PF4 did not stimulate mesangial cell migration29 in vitro, but may have an effect in vivo. Maximal expression of Fn-C mRNA and all three alternatively spliced variants was observed in glomerular lesions at 48 and 72 hours after habu snake venom injection and suggests that these isoforms might also be involved in cell functions related to cell proliferation. We have shown that peak mesangial cell proliferation in this model occurs at 48 hours and persists, although at a lesser rate, at 72 hours after habu snake venom injection.' 1'13 EIIIA and EIIIB mRNA and protein were not expressed by mesangial cells until 48 hours and after, associating these isoforms with proliferation or differentiation. Similarly, parietal epithelial cells showed an enhanced expression of all three spliced Fn variants, particularly at regions adjacent to microaneurysms and micronodules. Parietal epithelial cells do not express EDAcFn or EDBcFn protein in adult glomeruli18 19; but elevated expression of these isoforms has been associated with epithelial cell proliferation in synechiae of diseased glomeruli in chronic renal allograft rejection.19 Indeed, 3H-thymidine labeling studies1'13 indicate that peak parietal epithelial cell proliferation occurs before maximum mesangial cell proliferation in the habu snake venom model, coinciding with the course of cellular expression of EIIIA+ and EIIIB+ Fn isoforms observed in this study. Mesangial and parietal epithelial cells also acquire the expression of a-smooth muscle cell actin during the course of the HSV-induced glomerular disease. Acquisition of the a-smooth muscle cell actin phenotype by mesangial cells has been associated with proliferation both experimental and clinical settings of glomerular disease,31'32 and expression of EIIIA has been shown to correlate with myofibroblast expression of a-smooth muscle cell

actin during idiopathic pulmonary fibrosis33 and during lipocyte conversion to myofibroblasts in hepatic fibrogenesis.34 Thus, cellular expression of EIIIA and EIIIB at these times appears to be related to a phenotypic switch favoring cell proliferation or differentiation, in agreement with the proposed functions of these isoforms during embryogenesis and dis-

ease.3,9,10,18,19 Fn might also serve as a foundation for subsequent matrix assembly because of its different binding domains for heparin (heparan sulfate), collagen and other Fn molecules.1' 335 Extracellular matrix accumulation is regulated by TGF-p,36-38 and TGF-j preferentially stimulates El IA.39 Expression of Fn40-43 and specifically EIIIA protein44-47 has been associated with TGF-,13 mRNA in several glomerular diseases involving matrix expansion. Our study also associates maximum expression of all three Fn variants with TGF-,B1 mRNA by Northern analysis, described in a previous report.13 These studies associate Fn variants with various cellular events during the course of habu snake venom-induced glomerular injury. It is obvious that further investigation will be necessary to define a role for Fn and elucidate specialized functions for Fn isoforms in the control of cell behavior in general and in cell remodeling during glomerular disease.

Acknowledgments The authors thank Dr. Richard Hynes for providing the cDNA probes and antibodies for detection of Fn isoforms and Livingston van de Water for the lysozyme cDNA probe.

References 1. Proctor RA: Fibronectin: a brief overview of its structure, function, and physiology. Rev Infect Dis 1987,

9:S31 7-S321 2. Dufour S, Duband J-L, Kornblihtt AR, Thiery JP: The role of fibronectins in embryonic cell migrations. Trends Genet 1988, 4:198-203 3. Hynes RO: Fibronectins. New York, Springer-Verlag New York Inc., 1990, pp 1-546 4. Schwarzbauer JE, Paul JI, Hynes RO: On the origin of species of fibronectin. Proc Natl Acad Sci USA 1985, 82:1424-1428 5. Schwarzbauer JE: Alternative splicing of fibronectin: three variants, three functions. BioEssays 1991, 13: 527-533 6. Paul JI, Schwarzbauer JE, Tamkun JW, Hynes RO: Cell-type-specific fibronectin subunits generated by alternative splicing. J Biol Chem 1986, 261:12258-12265

1370

Barnes et al


7. Ffrench-Constant C, Hynes RO: Patterns of fibronectin gene expression and splicing during cell migration in chicken embryos. Development 1988, 104:369-382 8. Ffrench-Constant C, Hynes RO: Alternative splicing of fibronectin is temporally and spatially regulated in the chicken embryo. Development 1989, 106:375-388 9. Ffrench-Constant C, Van De Water L, Dvorak HF, Hynes RO: Reappearance of an embryonic pattern of fibronectin splicing during wound healing in the adult rat. J Cell Biol 1989, 109:903-914 10. Brown LF, Dubin D, Lavigne L, Logan B, Dvorak HF, Van De Water L: Macrophages and fibroblasts express embryonic fibronectins during cutaneous wound healing. Am J Pathol 1993, 142:793-801 11. Barnes JL, Hevey KA, Hastings RR, Bocanegra RA: Mesangial cell migration precedes proliferation in habu snake venom-induced glomerular injury. Lab Invest 1994, 70:460-467 12. Barnes JL, Hastings RR, De La Garza MA: Sequential expression of fibronectin by platelets, macrophages, and mesangial cells in proliferative glomerulonephritis. Am J Pathol 1994, 145:585-597 13. Barnes JL, Abboud HE: Temporal expression of autocrine growth factors corresponds to morphological features of mesangial proliferation in habu snake venominduced glomerulonephritis. Am J Pathol 1993, 143: 1366-1376 14. Barnes JL: Glomerular localization of platelet secretory proteins in mesangial proliferative lesions induced by habu snake venom. J Histochem Cytochem 1989, 37: 1075-1082 15. Guan J, Trevithick JE, Hynes RO: Retroviral expression of alternatively spliced forms of rat fibronectin. J Cell Biol 1990, 110:833-847 16. Peters JH, Trevithick JE, Johnson P, Hynes RO: Expression of the alternatively spliced EIIIB segment of fibronectin. Cell Adhesion Comm 1995, 3:67-89 17. Kocher 0, Kennedy SP, Madri JA: Alternative splicing of endothelial cell fibronectin mRNA in the IIICS region. Functional significance. Am J Pathol 1990, 137:15091524 18. Laitinen L, Vartio T, Virtanen I: Cellular fibronectins are differentially expressed in human fetal and adult kidney. Lab Invest 1991, 64:492-498 19. Gould VE, Martinez-Lacabe V, Virtanen I, Sahlin KM, Schwartz MM: Differential distribution of tenascin and cellular fibronectins in acute and chronic renal allograft rejection. Lab Invest 1992, 67:71-79 20. Schwarzbauer JE, Tamkun JW, Lemischka IR, Hynes RO: Three different fibronectin mRNAs arise by alternative splicing within the coding region. Cell 1983, 35:421-431 21. Oh E, Pierschbacher M, Ruoslahti E: Deposition of plasma fibronectin in tissues. Proc Natl Acad Sci USA 1981, 78:3218-3221 22. Postlethwaite AE, Keski-oja J, Balian G, Kang AH: Induction of fibroblast chemotaxis by fibronectin. Localization of the chemotactic region to a 140,000-molec-

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35. 36.

37.

ular weight non-gelatin-binding fragment. J Exp Med 1981, 153:494-499 Shoji S, Rickard KA, Ertl RF, Robbins RA, Linder J, Rennard SI: Bronchial epithelial cells produce lung fibroblast chemotactic factor: Fibronectin. Am J Respir Cell Mol Biol 1989, 1:13-20 Koyama N, Hart CE, Clowes AW: Different functions of the platelet-derived growth factor-a and -p receptors for the migration and proliferation of cultured baboon smooth muscle cells. Circ Res 1994, 75:682-691 Dufour S, Duband J-L, Humphries MJ, Obara M, Yamada KM, Thiery JP: Attachment, spreading, and locomotion of avian neural crest cells are mediated by multiple adhesion sites on fibronectin molecules. EMBO J 1988, 7:2661-2671 Greenberg JH, Seppa S, Seppa H, Hewitt AT: Role of collagen and fibronectin in neural crest cell adhesion and migration. Dev Biol 1981, 87:259-266 Nickoloff BJ, Mitra RS, Riser BL, Dixit VM, Varani J: Modulation of keratinocyte motility. Correlation with production of extracellular matrix molecules in response to growth promoting and antiproliferative factors. Am J Pathol 1988, 132:543-551 Sarret Y, Stamm C, Jullien D, Schmitt D: Keratinocyte migration is partially supported by the cell-binding domain of fibronectin and is RGDS-dependent. J Invest Dermatol 1992, 99:656-659 Barnes JL, Hevey KA: Glomerular mesangial cell migration: response to platelet secretory products. Am J Pathol 1991, 138:859-866 Barnes JL, Hevey KA: Glomerular mesangial cell migration in response to platelet-derived growth factor. Lab Invest 1990, 62:379-382 Johnson RJ, lida H, Alpers CE, Majesky MW, Schwartz SM, Pritzl P, Gordon K, Gown AM: Expression of smooth muscle cell phenotype by rat mesangial cells in immune complex nephritis. a-smooth muscle actin is a marker of mesangial cell proliferation. J Clin Invest 1991, 87:847-858 Alpers CE, Hudkins KL, Gown AM, Johnson RJ: Enhanced expression of "muscle-specific" actin in glomerulonephritis. Kidney Int 1992, 41:1134-1142 Kuhn C, McDonald JA: The roles of the myofibroblasts in idiopathic pulmonary fibrosis. Ultrastructural and immunohistochemical features of sites of active extracellular matrix synthesis. Am J Pathol 1991, 138:1257-1265 Jarnagin WR, Rockey DC, Koteliansky VE, Wang S-S, Bissell DM: Expression of variant fibronectins in wound healing: cellular source and biological activity of the EIIIA segment in rat hepatic fibrogenesis. J Cell Biol 1994, 127:2037-2048 McDonald JA: Extracellular matrix assembly. Annu Rev Cell Biol 1988, 4:183-207 Roberts AB, Heine Ul, Flanders KC, Sporn MB: Transforming growth factor-,8. Major role in regulation of extracellular matrix. Ann NY Acad Sci 1990, 580:225-232 Ignotz RA, Massague J: Transforming growth factor-a stimulates the expression of fibronectin and collagen


1371


38. 39.

40.

41.

42.

and their incorporation into the extracellular matrix. J Biol Chem 1986, 261:4337-4345 Ignotz RA, Endo T, Massagu6 J: Regulation of fibronectin and type collagen mRNA levels by transforming growth factor-P. J Biol Chem 1987, 262:6443-6446 Balza E, Borsi L, Allemanni G, Zardi L: Transforming growth factor P3 regulates the levels of different fibronectin isoforms in normal human cultured fibroblasts. FEBS Letters 1988, 228:42-44 Goyal M, Wiggins R: Fibronectin mRNA and protein accumulation, distribution, and breakdown in rabbit anti-glomerular basement membrane disease. J Am Soc Nephrol 1991, 1:1334-1342 Ding G, Pesek-Diamond I, Diamond JR: Cholesterol, macrophages, and gene expression of TGF-B1, and fibronectin during nephrosis. Am J Physiol 1993, 264: F577-F584 Jones CL, Buch S, Post M, McCulloch L, Liu E, Eddy AA: Pathogenesis of interstitial fibrosis in chronic purine aminonucleoside nephrosis. Kidney Int 1991, 40: 1020-1031

43. Tamaki K, Okuda S, Ando T, Iwamoto T, Nakayama M, Fujushima M: TGF-31 in glomerulosclerosis, and interstitial fibrosis of adriamycin nephropathy. Kidney Int 1994, 45:525-536 44. Kagami S, Border WA, Ruoslahti E, Noble NA: Coordinated expression of f31 integrins and transforming growth factor-,B-induced matrix proteins in glomerulonephritis. Lab Invest 1993, 69:68-76 45. Yamamoto T, Noble NA, Miller DE, Border WA: Sustained expression of TGF-,1l underlies development of progressive kidney fibrosis. Kidney Int 1994, 45:916927 46. Yamamoto T, Nakamura T, Noble NA, Ruoslahti E, Border WA: Expression of transforming growth factor p is elevated in human and experimental diabetic nephropathy. Proc Natl Acad Sci USA 1993, 90:1814-1818 47. Border WA, Noble NA, Yamamoto T, Harper JR, Yamaguchi Y, Pierschbacher MD, Ruoslahti E: Natural inhibitor of transforming growth factor-a protects against scarring in experimental kidney disease. Nature 1992, 360:361-364