INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

FIBRONECTINS (FNs) ARE A FAMILY of large (~440 kDa) dimeric glycoproteins implicated in a variety of biological processes, including tissue remodeling during wound healing and embryonic development (for review, see Refs. 29, 40). FNs influence these processes by affecting cell adhesion and extracellular matrix assembly. They exist in two forms, plasma and cellular. Plasma FN is made by hepatocytes and circulates in the blood; cellular FN is produced by epithelial and connective tissue cells and is deposited into the extracellular matrix. Both forms of FN are encoded by a single gene, whose transcript can be alternatively spliced at three intron-exon boundaries, yielding FN protein subunits containing, in the rat, the extra type III A (EIIIA) repeat, the extra type III B (EIIIB) repeat, and/or the variable (V) region (Fig. 1; for review, see 19, 51-52). Equivalent regions in humans are designated EDA, EDB, and IIICS. During splicing of FN mRNA, these exons are either included or excluded depending on the tissue and cell type. In plasma FN mRNA, exons encoding the EIIIA and EIIIB segments are excluded, whereas the exon encoding the V region is either partially or completely included depending on the species. In cellular FN mRNA, EIIIA and/or EIIIB exons are wholly retained and the V region appears to be at least partially included (43, 45, 56). Up to 12 different combinations of the three alternatively spliced segments are possible, resulting in a heterogeneous mixture of FN isoforms (32, 53).

View larger version (15K): [in this window] [in a new window]   Fig. 1.   Schematic illustration of 1 subunit of a fibronectin (FN) dimer in the rat. FN consists of a series of type I, II, and III repeats, each of which is specialized into binding domains with affinities for fibrin, collagen, cells, and heparin. The role of the first seven type III repeats (III, 1-7) is incompletely understood. The EIIIB and EIIIA (extra type III, A and B) repeats that are produced by alternative splicing of mRNA are darkened. The V120 region that is also alternatively spliced is illustrated by a rectangle with shaded (V25) and hatched (V95) areas. An arginine-glycine-aspartic acid (RGD) sequence present in the cell binding domain and a PEILDV sequence in V25 interact with the 51 and 41 integrins, respectively. Arginine-arginine (RR) in the heparin domain mediates binding to heparin. The synergistic site (brackets) in cooperation with RGD promotes full cell spreading [redrawn from Schwarzbauer (51)].

The function of the various isoforms is incompletely understood. FNs retaining the EIIIA and/or EIIIB regions are widely expressed in embryonic tissues and cultured cell lines compared with normal adult tissues (12, 20-21, 43-44, 46-47, 53, 64). Upregulation of EIIIA- and/or EIIIB-containing FN has also been noted in the adult at sites of tissue injury, such as in the skin (14, 22), eye (41, 67), and vasculature (17, 25, 28, 60). Isoforms containing a subsegment of the V region known as V25 are also preferentially expressed in embryos and injured peripheral nerves (37, 68).

In the kidney, FN isoforms are differentially expressed during development. In human fetal kidney, FN containing the EDA segment is widely found in tubular and glomerular basement membranes, the interstitium, and glomerular mesangium (34). The isoform containing the EDB region follows a similar but more restricted distribution, localizing to the mesangium, the interstitium, and basement membranes of medullary tubules and collecting ducts, albeit at lower levels (34). In the adult human nephron, FN containing EDA is found only in the mesangium; neither EDA- nor EDB-containing FNs are detected in tubular basement membranes or the interstitial compartment (34).

We previously demonstrated in an in vivo model of unilateral renal ischemia that FN strongly increases in expression and changes its distribution in response to injury (76). In particular, high concentrations of FN were observed in tubular lumens, suggesting a possible role in tubular obstruction. This study was initiated to understand whether tubular FN originated from plasma or from epithelial and/or connective tissue cells after ischemic injury. We demonstrate that luminal FN originates from plasma and is voided in the urine. We also observe that in response to ischemia-reperfusion, FN containing the EIIIA and/or V segments dramatically increases and V-FN, in particular, changes its distribution. These observations have important implications for understanding the mechanism(s) involved in the pathophysiology of acute renal failure (26, 62).

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Surgical procedure. A unilateral rat model of renal ischemia in which the renal artery and vein are clamped for 40 min whereas the contralateral kidney is left intact was used in this study.

Urine collection. Urine was collected at 24-h intervals in the presence of Complete protease inhibitors (Boehringer-Mannheim, Indianapolis, IN) from animals housed in metabolic cages (100 µl of 50× protease inhibitor/collection tube). The volume collected was measured, and the urine was filtered through two layers of cheesecloth before being centrifuged at 2,000 rpm at 0°C for 30 min. Urine was then filtered through a 0.45-µm Millex-HA filter (Millipore, Bedford, MA), aliquoted, and frozen in liquid nitrogen. Samples were stored at 80°C.

Tissue harvesting. Animals were anesthetized before tissue harvesting. For immunocytochemistry, rats were perfused through the left ventricle with warm Hanks' balanced salt solution and then with 2% paraformaldehyde-75 mM L-lysine-10 mM sodium periodate (PLP; 38). Kidneys were removed and stored intact in fixative overnight at 4°C. They were then rinsed in PBS and stored in PBS containing 0.02% sodium azide at 4°C.

Antibodies. A variety of monoclonal and polyclonal antibodies against FN were used in this study. Rabbit anti-rat FN polyclonal antibody (GIBCO-BRL, Grand Island, NY) was generated against rat plasma FN. This antibody recognizes regions of the FN molecule that are common to both the plasma and cellular forms. It was used to detect all forms of FN in tissue sections (1:750) and immunoblots (1:7,500). It is therefore referred to as anti-total FN antibody. Specificity of the anti-total FN antibody was determined by competition experiments between the antibody and purified antigen (76). Briefly, the anti-total FN antibody was incubated with purified rat plasma fibronectin, and the mixture then applied to tissue sections. After rinsing, sections were processed as described below for immunocytochemistry. Under these conditions, immunostaining was obliterated (76).

Cryosectioning and immunofluorescence. For frozen sections, fixed intact kidneys were sliced and equilibrated overnight at 4°C in 0.6 M sucrose/PBS. Kidney slices were then embedded in OCT (optimal cutting temperature) compound (Miles Laboratories, Naperville IL) and frozen in liquid nitrogen. Sections (5 µm) were cut on either a Reichert Frigocut cryostat (Leica, Deerfield, IL) or a Leica CM1850 cryostat. Sections were placed on SuperFrost Plus glass slides (Fisher Scientific, Boston, MA) and stored at 20°C until use.

Western Blotting

Tissue samples. Frozen kidney tissue was quickly weighed and transferred to a Dounce homogenizer on ice. Nine parts 2× sample buffer (400 mM Tris, pH 8.8, 10 mM EDTA, 2% SDS) at 4°C containing protease inhibitors (10 µg/ml aprotinin, 1 µg/ml antipain, 1 µg/ml pepstatin, 17.5 µg/ml benzamidine) but lacking sucrose, bromophenol blue and dithiothreitol (DTT) were then added. Tissue was extensively homogenized and boiled for a total of 8 min at 95°C to completely solubilize cell proteins. Samples were then cooled to room temperature, DTT was added to 40 mM, and samples were further homogenized and boiled (4 min). After cooling to room temperature, samples were sheared in succession with 22-, 23-, and 26-gauge needles, transferred to microfuge tubes, and spun twice at 16,000 g for 5 min at room temperature. Any remaining pellet was reextracted (2× sample buffer containing DTT, SDS, and protease inhibitors), sheared, and centrifuged as before. Supernatants were pooled, and protein concentration was then measured using the Standard Bio-Rad Protein Assay (Bio-Rad, Hercules, CA) with BSA as standard.

Urine samples. Urine samples were thawed on ice for 15-20 min and centrifuged for 5 min at 4°C. Typically, no pellet was visible. Volumes of urine containing equivalent amounts of protein were mixed with an equal volume of 2× sample buffer containing 24% sucrose, 0.2% bromophenol blue, 40 mM DTT and protease inhibitors. Samples were boiled at 95°C for 4 min and quickly cooled on ice.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

We and others have previously demonstrated in in vivo rat models of renal ischemia that cell extracellular matrix interactions dynamically change in response to injury (70, 76). In particular, we demonstrated that FN, a major glycoprotein of the plasma and renal interstitium, significantly altered its distribution and expression. Most surprisingly, it was intensely immunostained in tubular lumens. In this location, FN might contribute to tubular obstruction, a major outcome of renal artery occlusion and reperfusion. Because our earlier work did not distinguish among the various forms of FN, we were unable to determine whether lumenal FN originated from leakage of plasma proteins or from the synthesis and deposition of the cellular form by tubular epithelial or connective tissue cells. Resolution of this issue is important for our understanding of how FN is involved in the pathogenesis of ischemic renal injury. To investigate the source of luminal FN, polyclonal and monoclonal antibodies against specific regions of this molecule that are found in either plasma and/or cellular FNs (EIIIA, EIIIB, and V95) were used to determine the distribution of plasma and cellular FNs in response to ischemia reperfusion.

Distribution and Expression of FN in the Outer and Inner Stripes

Total FN. Using a polyclonal antibody that recognizes both plasma and cellular FNs, referred to as anti-total FN antibody, FN was immunolocalized in kidney sections of rats killed at various times after reperfusion after a 40-min unilateral clamp of the renal artery and vein. In kidneys made ischemic but not reperfused (0 h postischemia, Fig. 2, A and B), anti-total FN antibody outlines tubules in the outer stripe (Fig. 2A) and inner stripe (Fig. 2B); it also stains the interstitial space (Figs. 2, A and B). One to three hours after reperfusion, localization is maintained (outer stripe, Fig. 2C; inner stripe, Fig. 2D). However, anti total-FN antibody also stains the luminal space of some tubules of the thick ascending limb in the inner stripe (asterisks, Fig. 2D).

View larger version (162K): [in this window] [in a new window]   Fig. 2.   Immunofluorescent staining of cryostat sections of postischemic kidneys reacted with the polyclonal total FN antibody recognizing both the plasma and cellular forms. A, C, E, and G represent the outer stripe of the outer medulla; B, D, F, and H represent the inner stripe of the outer medulla. At 0 h postischemia (A-B), total FN localizes to the interstitium surrounding basement membranes of S3 proximal tubules in the outer stripe (A) and thick ascending limbs in the inner stripe (B). At 3 h postischemia (C-D), total FN continues to localize to these regions; however, in the outer stripe (C), it also highlights broken basement membrane (arrow). In the inner stripe (D), it is detected in a few tubular lumens (asterisks). At 24 h postischemia (E-F), interstitial staining for total FN increases. The frequency of tubular FN is also greater within lumens of the outer (E) and inner (F) stripes (asterisks). By 120 h postischemia (G-H), luminal FN is not detected in the outer (G) or inner (H) stripes; however, interstitial FN staining remains strong. The faint apical staining in G is nonspecific. Bar = 40 µm.

View larger version (131K): [in this window] [in a new window]   Fig. 3.   Immunofluorescent staining of cryostat sections of contralateral (A-D) and mock (E-H) kidneys reacted with polyclonal total FN antibody. A, C, E, and G represent the outer stripe of the outer medulla; B, D, F, and H represent the inner stripe of the outer medulla. The 24-h time point is demonstrated in A-B and E-F whereas the 120-h time point is shown in C-D and G-H. Total FN localizes to the interstitium and surrounds basement membranes of S3 proximal tubules in the outer stripe and thick limbs in the inner stripe. It is not detected in tubular lumens at 24 h (A-B and E-F), nor does interstitial staining increase at 120 h (C-D and G-H). Bar = 28 µm.

View larger version (77K): [in this window] [in a new window]   Fig. 4.   Immunoblot of kidney tissue (10 µg/lane) after 6% SDS-PAGE probed with polyclonal total FN antibody and donkey anti-rabbit horseradish peroxidase (HRP); HRP was visualized with diaminobenzidine. Total FN (220 kDa under reducing conditions) significantly increases from 24 to 120 h in postischemic kidneys (A) compared with contralateral (B) and mock (C) kidney tissue. Purified plasma FN (10 ng/lane) served as positive control (+C).

EIIIA-FN. At 0 h postischemia, FN containing the EIIIA segment is not detected in the outer (Fig. 5A) or inner stripes (Fig. 5B) of the outer medulla. Rarely, some patchy, faint immunostaining associated with tubular basement membranes is observed (arrow, Fig. 5A); red blood cells also autofluoresce at this time (asterisks; Fig. 5, A and B). Production of EIIIA-FN significantly increases 1-3 h after reperfusion, associating primarily with basement membranes of S3 proximal tubules in the outer stripe (Fig. 5C) and thick limbs of the inner stripe (Fig. 5D). Interstitial localization in the inner stripe is also strong. FN containing the EIIIA segment is not detected in tubular lumens of the inner stripe (Fig. 5D), in contrast to the observation with anti-total FN antibody (Fig. 2D).

View larger version (141K): [in this window] [in a new window]   Fig. 5.   Immunofluorescent staining of cryostat sections of postischemic kidneys reacted with monoclonal antibody recognizing the EIIIA region of cellular FN. A, C, E, and G represent the outer stripe of the outer medulla; B, D, F, and H represent the inner stripe of the outer medulla. At 0 h reperfusion (A-B), the EIIIA region of cellular FN is not detected in the outer (A) or inner (B) stripes. A few tubules, however, demonstrate a faint patchy distribution (arrows, A). Note the autofluorescence of red blood cells (asterisks, A-B). At 3 h reperfusion (C-D), FN containing EIIIA is detected at high levels in the interstitial space between S3 proximal tubules in the outer stripe (C) and thick limbs in the inner stripe (D), where it outlines tubular basement membranes. At 24 h reperfusion (E-F), staining for EIIIA-FN further increases, especially in the outer stripe (E), where the interstitial space appears enlarged. At 120 h after reperfusion (G-H), EIIIA-FN persists at high levels in an expanded interstitium in both the outer (G) and inner (H) stripes, suggesting an incipient fibrosis. Note that at 3 and 24 h after reperfusion, FN containing EIIIA is not detected in tubular lumens. Bar = 40 µm.

View larger version (106K): [in this window] [in a new window]   Fig. 6.   Immunofluorescent staining of cryostat sections of contralateral (A-D) and mock (E-H) kidneys reacted with monoclonal antibody recognizing the EIIIA region of cellular FN. A, C, E, and G represent the outer stripe of the outer medulla; B, D, F, and H represent the inner stripe of the outer medulla. The 3-h time point is demonstrated in A-B, the 24-h time point in C-F, and the 120-h time point in G-H. In the contralateral kidney (A-D), FN containing EIIIA is detected in the outer and inner stripe outlining tubular basement membranes. Interstitial staining does not increase at 24 h (C-D) nor is EIIIA-FN detected in tubular lumens (C-D). In sham-operated kidneys (E-H), FN containing EIIIA is absent in the outer stripe (E, G) and is barely detected in the inner stripe (arrows, F, H). Bar = 28 µm.

View larger version (20K): [in this window] [in a new window]   Fig. 7.   Immunoblot of kidney tissue (25 µg/lane) after 6% SDS-PAGE probed with mouse anti-human EDA FN and donkey anti-mouse HRP; HRP was visualized with diaminobenzidine. In this series of postischemic kidneys (A), FN containing EIIIA (220 kDa under reducing conditions) is detected at significant levels after 24 h of injury and continues to increase thereafter. In contralateral control (B) or mock-operated (C) kidneys, EIIIA-FN is barely detected. Ischemic kidney tissue (120 h) served as positive control (+C).

EIIIB-FN. FN containing the EIIIB region was examined next. EIIIB, like EIIIA, is unique to cellular FN. EIIIB-FN was not detected in kidney sections at any of the time points examined after ischemic injury. It is also not detected in normal adult rat (42, 44), mouse (46), or human (34) kidneys. As a positive control for immunostaining, EIIIB-FN was detected in Descemet's membrane of the adult rat cornea (data not shown; 46).

V95-FN. Localization of the V95 region of FN, which in the rat is found in a fraction of both plasma and cellular FNs (54, 55), was also examined. At 0 h postischemia, little if any FN containing the V95 region is detected in the outer (arrows, Fig. 8A) and inner stripes (Fig. 8B). One to three hours later, localization in the outer stripe (Fig. 8C) does not change; however, in the inner stripe (Fig. 8D), FN containing V95 begins to localize to some lumenal compartments (asterisks, Fig. 8D).

View larger version (130K): [in this window] [in a new window]   Fig. 8.   Immunofluorescent staining of cryostat sections of postischemic kidneys reacted with polyclonal antibody recognizing the V95 region of rat plasma and cellular FN. A, C, E, and G represent the outer stripe of the outer medulla whereas B, D, F, and H represent the inner stripe of the outer medulla. At 0 h reperfusion (A-B), FN containing the V95 region is not detected in the outer medullary region except for infrequent patchy distribution that faintly associates with renal tubules (arrow, A). It is rarely detected in the interstitium 3 h after reperfusion (arrow, C) when it begins to localize to the luminal space of some thick limbs in the inner stripe (asterisks, D). At 24 h after reperfusion (E-F), staining intensity for FN containing V95 increases in the interstitium around S3 proximal tubules (E); in the inner stripe a delicate fibrillar staining is seen between tubules (F). Localization of the V95 isoform to lumens of thick ascending limbs also increases at this time (asterisks, F). At 120 h after reperfusion (G-H), FN containing V95 continues to increase in the interstitial compartment of the outer medulla, associating with basement membranes of renal tubules; V95-FN is absent from the luminal space in the inner stripe. Bar = 40 µm.

Distribution and Expression of FN Isoforms in the Cortex

Total FN. At 0 (Fig. 9A), 1, 3 (Fig. 9B), 24, 48 (Fig. 9C), and 120 h (Fig. 9D) postischemia, anti-total FN antibody highlights Bowman's capsule and the glomerular capillary tuft. Total FN is not observed in the urinary space of renal corpuscles at any time (asterisk, Figs. 9, A-D); however, from 3 to 48 h after reperfusion, it is more frequently detected in tubular lumens of the distal nephron adjacent to glomeruli (double asterisk, Figs. 9, B and C, respectively). The latter observation of FN in tubular lumens of nephron segments distal to glomeruli is consistent with results from the outer (see Fig. 2E) and inner stripes (see Fig. 2, D and F).

View larger version (157K): [in this window] [in a new window]   Fig. 9.   Immunofluorescent staining of cryostat sections of postischemic cortex reacted with anti-total FN polyclonal antibody. At 0 (A), 3 (B), 48 (C), and 120 h (D) postischemia, total FN outlines Bowman's capsule, the glomerular capillary tuft and proximal and distal tubules in the cortex; it also localizes to the interstitium, significantly increasing in this location at 48 (C) and 120 h (D) after reperfusion. Total FN is absent from the urinary space (*) surrounded by Bowman's capsule at all time points even though at 3 (B) and 48 h (C) postischemia it concentrates in lumens of nephron segments distal to glomeruli (**). Bar = 29 µm.

View larger version (116K): [in this window] [in a new window]   Fig. 10.   Immunofluorescent staining of cryostat sections of the cortex from contralateral (A-D) and mock-operated (E-H) kidneys at 24 h (A, C, E, and G) and 120 h (B, D, F, and H), reacted with either polyclonal antibody recognizing total FN (A, B, E, and F) or monoclonal antibody recognizing the EIIIA region of cellular FN (C, D, G, and H). In the cortex of contralateral (A-B) and mock (E-F) kidneys, total FN outlines Bowman's capsule, the glomerular capillary tuft, and proximal and distal tubules; it also localizes to the interstitium. Total FN is absent, however, from the urinary space surrounded by Bowman's capsule (A, B, E, and F) and from tubular lumens of distal nephron segments (A, B, E, and F). Interstitial staining does not increase at 120 h (B, F). FN containing the EIIIA region highlights glomerular elements in contralateral (C-D) and mock kidneys (arrows, G-H), although to a lesser extent in the latter. The dashed lines in G-H demarcate Bowman's capsule. Renal tubules stain nonspecifically in G. Bar = 28 µm.

EIIIA-FN. Immunostaining of both glomeruli and the interstitium of the cortex was examined with antibody recognizing the EIIIA region of FN. In glomerular regions at 0 h postischemia, anti-EIIIA antibody faintly demarcates the connective tissue of the capillary tuft (Fig. 11A; 34). This staining pattern does not change 1-3 h after reperfusion (data not shown). However, 24-48 h after injury (Fig. 11B), components of the renal corpuscle, including Bowman's capsule, stain more intensely, persisting at 120 h postischemia (Fig. 11C). FN containing EIIIA does not localize to Bowman's space at any time after reperfusion (asterisk, Figs. 11, A-C) and is absent from tubular lumens of the distal nephron.

View larger version (176K): [in this window] [in a new window]   Fig. 11.   Immunofluorescent staining of cryostat sections of postischemic cortex at 0 (A and D), 48 (B and E), and 120 h (C and F) after reperfusion reacted with monoclonal antibody recognizing the EIIIA region of cellular FN (A-C) or polyclonal antibody against the V95 region (D-F). At 0 h after reperfusion (A, D), EIIIA (A)-, and V95 (D)-FNs faintly localize to glomerular connective tissue; no reactivity is seen in the renal interstitium. At 48 h after reperfusion (B and E), the renal corpuscle is highlighted with both anti-EIIIA (B) and anti-V95 (E) antibodies; Bowman's capsule and the capillary tuft are outlined, although the localization for V95-FN appears to be less distinct. Some tubular basement membranes are also immunoreactive with both antibodies. In particular, EIIIA-FN shows a gradient of staining with greater intensity of localization in regions proximal to the outer medulla (B); cortical areas distal to this site are more weakly stained (B). A gradient of staining is not detected to the same extent for V95-FN (E). At 120 h postischemia (C, F), immunostaining of the renal corpuscle and an expanded interstitium remains intense for EIIIA-FN (C); however, a gradient of staining (C) is not as prevalent as at 48 h. In addition, V95-FN (F) is not present to the same extent as EIIIA-FN at 120 h postischemia. Note that at none of the timepoints after reperfusion is EIIIA (A, B, C)- or V95 (D, E, F)-FNs found in Bowman's space (*). Bar = 29 µm.

EIIIB-FN. The FN isoform containing the EIIIB region was not detected in the cortex either before (42) or after ischemic injury (data not shown).

V95-FN. Distribution and expression of FN containing the V95 region was examined in the glomerular and interstitial regions of the cortex. In the glomerular region at 0 (Fig. 11D), 3 (data not shown), 48 (Fig. 11E), and 120 h (Fig. 11F) postischemia, V95-FN weakly localizes to the connective tissue surrounding the capillary tuft and steadily increases in Bowman's capsule. FN containing V95 is absent from the urinary space at all times after reperfusion (asterisk, Fig. 11, D-F) but is sometimes seen in the lumen of the distal nephron between 3 and 48 h after reperfusion (data not shown).

Plasma But Not Cellular FN is Voided in the Urine

Immunoblotting with the anti-total FN antibody demonstrates that FN is voided in the urine in response to ischemia-reperfusion injury. Two patterns emerge. In some animals (n = 3), total FN is voided at increasing levels (Fig. 12A), beginning as early as 24 h after reperfusion, peaking 72-96 h after reperfusion and in some instances declining thereafter. In other animals (n = 4), a biphasic excretion pattern is seen (Fig. 12A). Total FN is voided at high levels 24 h after reperfusion, declines 48-72 h after reperfusion, and again increases from 72-120 h, reaching levels equivalent to that at the 24-h time point. In urine of mock operated or normal animals, total FN is excreted at very low levels (Fig. 12A).

View larger version (30K): [in this window] [in a new window]   Fig. 12.   Immunoblots of urine representing equal amounts of protein resolved by 6% SDS-PAGE, reacted with chemiluminescence reagents. A and B: 5 and 10 µg protein/lane, respectively. Urine was collected at 24-h intervals from animals that had undergone unilateral renal ischemia or from mock-operated animals; animals were housed in metabolic cages. Each immunoblot is from 1 representative animal. In the postischemic animal, antibody recognizing total FN (A) shows that FN is excreted in the urine in 2 patterns. In the first, it is voided at increasing levels from 24-120 h after reperfusion with high levels predominating between 72-120 h. In the second, a biphasic pattern is observed. High levels of FN are initially excreted at 24 h after reperfusion, followed by a precipitous decrease and subsequent increase that persists at 120 h. In the mock animal, fibronectin is excreted at very low levels, comparable to that detected in the urine of normal animals. Note the lower molecular weight breakdown fragments of FN excreted in the urine. Purified plasma FN (2.5 ng) is used as the positive control (+C) in A. Antibody recognizing the EDA region of human cellular FN (B) shows that cellular FN containing EIIIA is not excreted in the urine in either the ischemic or mock animal. Ischemic kidney tissue (5 µg, 48 h postischemic) is used as positive control (+C) in B.

In contrast to the data with the anti-total FN antibody, immunoblotting with anti-cellular EDA antibody confirms that cellular FN synthesized by renal epithelia or connective tissue cells is not detected in the urine at any time after ischemia-reperfusion (Fig. 12B). It is also not detected in urine collected from mock-operated animals (Fig. 12B). Thus cellular FN is not excreted, suggesting that all the FN that is excreted is from the plasma and that it is plasma not cellular FN that is found in tubular lumens.

Distribution and Expression of FN Isoforms 6 Wk After Ischemic Injury

At 6 wk postischemia, FN continues to localize at high levels in previously injured regions of the outer medulla, identified by the heterogeneous shapes and sizes of the regenerated tubules. Antibody against total FN strongly stains the interstitial compartment, sometimes concentrating at higher levels adjacent to basement membranes of renal tubules (Fig. 13A). In contrast, antibodies to EIIIA and V95-FNs mainly identify tubular basement membranes of varying thicknesses, whereas other basement membranes are not outlined (Fig. 13, B-C, respectively). The interstitial compartment contains some EIIIA and V95-FN, however, at significantly lower levels than that observed for total FN.

View larger version (90K): [in this window] [in a new window]   Fig. 13.   Immunofluorescent staining of cryostat sections of 6-wk postischemic kidney reacted with polyclonal and monoclonal antibodies to FN. Total FN (A) strongly localizes to the interstitial space of the outer medulla and concentrates at high levels in the connective tissue adjacent to some renal tubules. FN containing the EIIIA (B) or V95 (C) regions primarily identifies tubular basement membranes of varying thicknesses in this region; overall, EIIIA and V95-FNs are less conspicuous in the interstitium than total FN. Bar = 23 µm.

In the cortex, FN containing the EIIIA or V95 regions is not detected in association with basement membranes or the interstitium (data not shown); sclerosis of glomeruli also is not evident (data not shown).

FN containing EIIIB is not detected in either the outer medulla or cortex of the 6-wk postischemic kidney (data not shown).

Thus EIIIA and V95-FNs are found in the 6-wk postischemic kidney, localizing to the basement membrane region of previously injured tubules. They are also found to a lesser extent in the surrounding interstitium.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

We demonstrate using an in vivo model of unilateral renal ischemia that production of FN splice variants dramatically changes as a result of renal injury and repair (Table 1). In ischemic kidneys that are not reperfused (0 h), the alternatively spliced forms of FN, which contain the EIIIA, EIIIB, and/or V95 regions, are absent from the renal interstitium. However, at early times after reperfusion (1-3 h), FN containing EIIIA increases in this location in the outer medulla whereas FN containing V95 does not. The isoform containing V95, but not the EIIIA or EIIIB region, is found in a few lumens of the distal nephron. At 24-48 h postischemia, as further injury and regeneration coordinately occur, localization of both EIIIA- and V95-containing FN isoforms increases in the interstitial matrix in both the outer medulla and cortex, with the increase in EIIIA-FN being more dramatic. The V95-containing FN isoform, but not that containing EIIIA, also frequently fills the luminal space. As regeneration continues at 120 h after reperfusion, EIIIA- and V95-containing FN strongly localize to the interstitial compartment, and V95-FN is no longer detected in tubular lumens. Luminal FN is ultimately excreted in the urine. Six weeks after ischemic injury, FN containing EIIIA and/or V95 persists around some renal tubules.

Our data using polyclonal and monoclonal antibodies specific for unique regions of the various isoforms of FN indicate that luminal FN originates from plasma. FN in tubular spaces was detected by immunohistochemistry with antibodies recognizing both total and V95-FNs but not with antibodies against the EIIIA- or EIIIB-containing cellular isoforms. In addition, FN excreted in the urine was not detected by antibody specific for cellular FN.

It is unclear how plasma FN enters the luminal space. There is no evidence that it is made by other cell types in addition to hepatocytes (61). Furthermore, our data suggest that luminal plasma FN does not originate from damage to the glomerular filtration barrier, which normally is impermeant to plasma proteins. Total or V95-FNs, with molecular weights of ~440 kDa, were not detected in Bowman's space at any time after ischemia. We believe it likely, therefore, that plasma proteins, including FN, extravasate from interstitial blood vessels into the interstitial space from where it penetrates damaged tubules and enters the lumen. Permeability changes in interstitial blood vessels would allow the escape of plasma FN into the interstitium. In combination with tubulorrhexis and/or disruption to the epithelial permeability barrier (59, 69), plasma FN would then gain entry into tubular lumens. Once in the lumen of the injured tubule, plasma FN concentrates in more distal nephron segments.

Under these circumstances, luminal plasma FN might play a very significant role in the pathophysiology of acute renal failure. Our data indicate that within the first 48 h of ischemic injury, FN is detected by immunohistochemistry in the luminal space of the injured kidney while simultaneously being excreted in the urine. Although we have no evidence plasma FN is occluding the lumen, the high concentration of plasma FN in this location may suggest that its movement is retarded. This may occur through the formation of obstructive complexes with other molecules such as the Heymann antigen gp330 (39) on microvilli shed in response to injury (16, 65) as well as other cellular debris, contributing to the formation of casts. We do not believe plasma FN adheres to sloughed dead or viable cells in the luminal space, because FN fibrils, which assemble on cell-FN interactions (27, 71, 73-74), are not detected. Ultimately, all plasma FN is excreted, indicating that any obstruction is not permanent (see Figs. 2H and 8H).

We also demonstrate that after ischemic injury the FN isoforms containing the EIIIA and V95 regions but not EIIIB localize to the renal interstitium surrounding tubular basement membranes. Before injury, FN containing EIIIA or V95 is not detected in this location. Within 3 h of reperfusion, however, an increase in EIIIA but not V95-FN is observed in the interstitial matrix of the outer medulla and continues to increase at 24-48 h after injury. Although V95-FN increases somewhat at this time, a parallel rise is not detected. Both persist at high levels in the renal interstitium 120 h after reperfusion.

The upregulation of FN isoforms suggests that ischemic injury may stimulate a signal regulating the production of the alternatively spliced forms. Transforming growth factor- (TGF-), a multifunctional cytokine that operates in either a paracrine or autocrine manner, stimulates fibronectin synthesis (30) and regulates production of the EIIIA-FN isoform by acting on FN pre-mRNA (13, 35). It is derived from a number of cell types including platelets (5), wound macrophages (49), and fibroblasts/myofibroblasts (75), as well as injured proximal tubular epithelial cells (10). It has a known role in the pathogenesis of renal disease (11, 58) and is believed to promote tissue regeneration after renal injury (8, 10). It is unclear which cells at the various time points after reperfusion produce TGF- in the in vivo model of unilateral renal ischemia used in this study. Whether TGF- acts in a paracrine or autocrine fashion on injured epithelial or connective tissue cells (e.g., infiltrating macrophages, fibroblasts/myofibroblasts) to promote the synthesis, secretion, and deposition into the extracellular matrix of FN containing the EIIIA region also needs to be determined. TGF- when administered in vitro increases expression of EIIIA-FN mRNA by proximal tubule cells (35, 66); its immunoneutralization in vivo in a bilateral model of renal injury also attenuates the rise in FN containing EIIIA (9). The production of TGF- and other paracrine growth factors may explain the unusual gradient of staining for EIIIA-FN in the cortex, which is intense in regions adjacent to the outer medulla (see Fig. 11B). Paracrine growth factors may also be responsible for the observation that EIIIA-FN is not detected in the cortex until 24 h after reperfusion (Fig. 11B) even though in the outer and inner stripes it is present at high levels soon after reperfusion (1-3 h postischemia; see Fig. 5, C-D).

Growth factors produced by the injured kidney may also explain the appearance of the alternatively spliced forms in the contralateral kidney detected by immunostaining. The FN isoform containing EIIIA was detected within hours in the contralateral kidney, whereas the FN isoform containing V95 was detected later. Both persisted at the latest time point we examined, although levels did not increase from that seen initially and overall were less than that seen in the ischemic kidney. Perhaps growth factors, such as TGF-1, released by the injured ischemic kidney enter the systemic circulation; in this location, it may act on the contralateral kidney to stimulate the production and accumulation of FN containing EIIIA and/or V95. Alternatively, the experimental procedure of unilaterally clamping the renal artery and vein may have resulted in some generalized reduction in blood pressure, resulting in reduced perfusion of the contralateral kidney.

In the postischemic kidney, the role of FN containing EIIIA is unknown; however, our data indicate that it may be necessary for regeneration. EIIIA-FN is widely expressed during organogenesis and tissue remodelling in the embryo (20-21, 43-44, 47, 64) and during wound healing of various tissue types (2-3, 7, 14, 17, 20, 25, 28, 41-42, 60, 67). In each of these cases, FN containing EIIIA may be required for cell migration, proliferation, and differentiation. The association of EIIIA-FN with tubular basement membranes may promote migration of remaining epithelial cells after damaged cells have exfoliated, as well as the spreading of adherent epithelia to fill in denuded areas of basement membrane. Similarly, it may stimulate epithelial proliferation, which occurs during this time and is necessary for repopulation of the damaged tubule (72). These cellular responses are key elements of epithelial redifferentiation after renal ischemic injury (62, 76). It would be interesting to examine at the ultrastructural level whether EIIIA-FN is indeed an integral component of regenerating basement membranes and whether in this location it mediates these cell-extracellular matrix interactions, perhaps by inducing the expression of novel extracellular matrix receptors, such as the 51 integrin, the major fibronectin receptor, and/or contributing to the altered distribution of other integrins (76).

FN containing EIIIA may also play a role in the fibrotic response after renal ischemic injury. In this regard, FN is believed to be the first extracellular matrix molecule deposited in renal interstitial fibrosis (18). In the liver, FN containing EIIIA activates the conversion of normal hepatic lipocytes into myofibroblast-like cells, a critical event in the initiation of fibrogenesis in the liver (31). In the formation of granulation tissue in wounded skin, EDA FN is necessary for induction of the myofibroblastic phenotype by TGF-1 (57). Whether deposition of FN containing EIIIA in the postischemic kidney is necessary for the conversion of stromal fibroblasts and/or perivascular cells into myofibroblast-like cells (18) is unknown. However, TGF- mediates this activation (15) and also stimulates production of FN containing EIIIA (6). We see that in the injured kidney FN containing EIIIA is expressed soon after reperfusion, continues to increase in an expanded interstitium, and persists weeks later in association with thickened basement membranes. Perhaps deposition of EIIIA-FN is the key extracellular matrix component that initiates renal fibrosis after ischemic injury.

More is known about the role of FN containing the V95 region, which in the rat is a component of both plasma and cellular FNs. FN exists as a dimeric glycoprotein, and the V95 region must be present in at least one subunit for its secretion to take place (for review, see Refs. 51-52). Our immunostaining data clearly show that although the EIIIA isoform is present in high amounts in the renal interstitium after ischemic injury, levels of V95 are not necessarily comparable. Similar results have been obtained by immunohistochemistry in skeletal and heart muscle, lung parenchyma, small intestine, pancreas, and peripheral nerve (46). This apparent discrepancy could arise from the presence of the V95 region in only one subunit of the FN dimer and/or from differences in antibody sensitivity or epitope masking.

These same reasons could explain why the FN isoform containing the V95 region is not detected in tubular lumens at the same frequency as total FN, which recognizes both the plasma and cellular forms. The V95 region in the rat is one of three variants produced by alternative splicing of the V region exon (for review, see Refs. 19, 51-52). These variants include V120, which contains all V120 region amino acids, V95, which lacks the first 25 amino acids, and V0, which lacks the entire V region. In rat plasma FN, ~50% of the FN subunits are V0 and 50% are a combination of V120 and V95 (54). Considering that less than 50% of the FN subunits circulating in plasma are V95, it is not surprising that the frequency of the V95 isoform in tubular lumens is not equivalent to that observed with the total FN antibody.

At 6 wk postischemic injury, FN isoforms containing EIIIA and V95 strongly localize to thickened basement membranes of some tubules that initially were damaged and appear incompletely regenerated; localization to the surrounding interstitium is weak. Staining with anti-total FN antibody indicates that total FN is found in tubular basement membranes and is strongly expressed in the interstitial compartment. The data suggest incomplete recovery of some regions of the postischemic kidney. In addition, plasma FN, possibly escaping from interstitial blood vessels, appears to be trapped in the connective tissue compartment, having been deposited into the extracellular matrix. It is possible that plasma FN is prevented from entering the luminal space either because the thickened basement membranes, containing EIIIA and V95 FNs, inhibit its passage and/or the epithelial permeability barrier is reestablished.

Our observation of incomplete recovery 6 wk after unilateral ischemic injury may be clinically relevant. In the patient population, the postischemic kidney has a tremendous capacity to completely regenerate, restoring normal renal function. However, it is possible that some nephrons do not completely recover. We believe that our observation in the unilateral model showing incomplete repair of some nephron segments could indicate heterogeneity of nephron function postischemia (4, 23). Long-term effects could result in chronic renal failure, glomerulosclerosis, and tubulo-interstitial damage (63). Similar processes may ultimately occur in patients suffering from ischemic injury.

In conclusion, we show that in response to ischemic injury, the FN isoforms containing the EIIIA and V95 regions that are not normally expressed in the renal interstitium are significantly upregulated, suggesting a possible dual role in regeneration of the