Tight junctions rarely exist in podocytes of the normal renal glomerulus, whereas they are the main intercellular junctions of podocytes in nephrosis and in the early stage of development. Claudins have been identified as tight junction-specific integral membrane proteins. Those of podocytes, however, remain to be elucidated. In the present study, we investigated the expression and localization of claudin-6 in the rat kidney, especially in podocytes. Western blot analysis and RT-PCR revealed that the neonatal kidney expressed much higher levels of claudin-6 than the adult kidney. Immunofluorescence microscopy showed intense claudin-6 staining in most of the tubules and glomeruli in neonates. The staining in tubules declined distinctly in adults, whereas staining in glomeruli was well preserved during development. Claudin-6 in glomeruli was distributed along the glomerular capillary wall and colocalized with zonula occludens-1. The staining became conspicuous after kidney perfusion with protamine sulfate (PS) to increase tight junctions in podocytes. Immunoelectron microscopy showed that immunogold particles for claudin-6 were accumulated at close cell-cell contact sites of podocytes in PS-perfused kidneys, whereas a very limited number of immunogold particles were detected, mainly on the basal cell membrane and occasionally at the slit diaphragm and close cell-cell contact sites in normal control kidneys. In puromycin aminonucleoside nephrosis, immunogold particles were also found mainly at cell-contact sites of podocytes. These findings indicate that claudin-6 is a transmembrane protein of tight junctions in podocytes during development and under pathological conditions.
- protamine sulfate
- puromycin aminonucleoside
the podocyte in the renal glomerulus is a unique epithelium that allows the bulk flow of solute and water to pass through its paracellular space. Morphologically at least two types of intercellular junctional complex are recognized in the paracellular space: slit diaphragms and junctions with close cell-cell contact. The latter have been shown to contain tight junctions and gap junctions (2, 16, 19, 28). The molecular components of slit diaphragms and their significant roles in glomerular filtration have been elucidated through studies of congenital nephrotic syndrome (26). In contrast, the latter, especially tight junctions, remain obscure in their components and functions.
Under physiological conditions, slit diaphragms are the predominant intercellular junction connecting adjacent foot processes of mature podocytes, whereas tight junctions are rarely observed and are restricted to regions of contacts between adjacent podocyte foot processes in or near the plane of the slit diaphragm (18). Under nephrotic conditions, slit diaphragms are extensively dislocated or disappear, and tight junctions are frequently formed in lieu of slit diaphragms in flattened podocyte foot processes (2, 11, 16, 19). During development, tight junctions connect presumptive podocytes at an early stage (17, 22) and disappear along with the opening of the intercellular spaces and the appearance of slit diaphragms. These observations suggest the existence of an intimate interaction or cross-talk between the slit diaphragm and the tight junction in formation of intercellular contacts of podocytes and that the two types of junctional complexes may compensate for each other under physiological and pathological conditions.
The morphological changes in podocytes, reminiscent of those observed in the nephrotic conditions, can be induced by perfusion of rat kidneys with polycationic compounds such as protamine sulfate (PS) or poly-l-lysine (8, 11, 24). On perfusion of polycations, formation of tight junctions and effacement of foot processes are observed within a few minutes, and tight junctions are typically formed below the dislocated slit diaphragms. It has been assumed that the neutralization of negatively charged groups on the surface of podocytes by polycations brings about the morphological changes. The decrease in the negatively charged polyanions on the surface of podocytes was also demonstrated in nephrotic conditions (9, 14), suggesting that a similar sequence of events is supposed to occur in nephrotic podocytes. Furthermore, experiments using perfusion of kidneys with polycations indicated that the formation of tight junctions did not require de novo synthesis of components of the tight junction (8). Together, these findings strongly suggest that components of the tight junction exist in close proximity to slit diaphragms and are redistributed to form tight junctions on perfusion of polycations.
Two distinct types of integral membrane protein have been so far identified as components of tight junctions, namely occludin and claudins (3, 4). Claudins are 20- to 27-kDa proteins with four transmembrane domains and consist of more than 20 members of a gene family (5). Overexpression of claudin genes resulted in formation of continuous networks of intramembranous fibrils (tight-junction strands), suggesting claudins as the major structural components of tight junctions (5). Because it has been reported that podocytes are not stained with anti-occludin antibody (20), we hypothesized that podocytes might constitutively express some claudins and regulate their recruitment to form tight junctions under pathological conditions and in development.
In preliminary experiments to search for claudins expressed in rat glomeruli by immunohistochemistry, we found a significant, definite staining for claudin-6 in the glomerulus. In the present study, we examined the precise localization of claudin-6 protein in the glomerulus of rat kidney under physiological conditions and in development and their changes in localization in nephrotic kidneys as well as in those perfused with PS. We found that claudin-6 was localized in tight junctions of immature podocytes during development as well as in mature podocytes. Interestingly, claudin-6 was predominantly found on the basal membrane of normal podocyte foot processes in addition to tight junctions. The findings suggest that claudin-6 may be redistributed to tight junctions newly formed on perfusion with PS and in nephrotic kidneys.
MATERIALS AND METHODS
Animals and antibodies.
Female Wistar-Kyoto rats were purchased from Charles River Japan (Atsugi, Japan) and were used in these experiments at the ages of 2 days and 8–10 wk. Rabbit anti-claudin-6 antibody was obtained from Immuno-Biological Laboratories (Takasaki, Gunma, Japan). The antibody was produced against an oligopeptide corresponding to the cytoplasmic domains of mouse claudin-6 (amino acids 207–219), and showed no cross-reactivity with claudin-1, -2, -3, -4, -5, -7, -8, -12, or -15 (21). Murine monoclonal antibodies against zonula occludens-1 (ZO-1) and paxillin were purchased from Zymed Laboratories (South San Francisco, CA).
The procedures for the present study were approved by the Animal Committee at Niigata University School of Medicine, and all animals were treated according to the guidelines for animal experimentation of Niigata University.
Adult and neonatal whole kidneys and isolated glomeruli were homogenized in TRIzol (GIBCO BRL, Grand Island, NY), and total cellular RNA was extracted from these samples essentially according to manufacturer's instructions. Glomeruli were isolated from kidneys pooled from the same group by the conventional sieving method described previously (27). Five micrograms of RNA was reverse transcribed into cDNA with random hexamers as primers and Superscript II reverse transcriptase (Invitrogen, Carlsbad, CA). Amplification was carried out by using Thermal Cycler Dice (Takara Bio, Shiga, Japan) through 20 cycles (GAPDH) or 28 cycles (claudin-6) of denaturation at 94°C for 30 s, annealing at 62°C for 30 s, and extension at 72°C for 1 min. The optimum number of amplification cycles used for semiquantitative PCR was chosen on the basis of the preliminary trial in the linear phase of amplification. Primers were as follows: rat claudin-6, forward 5′-cagagccctctgtgtcatca-3′ corresponding to 277 to 296 of rat claudin-6 (XM_220202), reverse 5′-ccctccacctatcagcaaaa-3′ corresponding to 555 to 574 of rat claudin-6; GAPDH, forward 5′-taaagggcatcctgggctacact-3′ corresponding to 807 to 829 of rat GAPDH (XM_008463), reverse 5′-ttactccttggaggccatgtagg-3′ corre sponding to 984 to 1006 of rat GAPDH.
Real-time RT-PCR was performed with a LightCycler 1.5 PCR thermal cycler (Roche Applied Science, Penzberg, Germany) to determine relative mRNA abundance. Amplification was performed with the primers for claudin-6 and GAPDH by using FastStart DNA MasterPLUS SYBER Green I according to the manufacturer's instructions. Recombinant plasmids containing PCR products of claudin-6 or GAPDH were used as standards. The results of real-time RT-PCR were represented as the ratio of claudin-6 to GAPDH.
Whole kidneys, isolated glomeruli, cortices, and medullae were homogenized in lysis buffer (8 M urea, 1 mM dithiothreitol, 1 mM EDTA, 50 mM Tris·HCl, pH 8.0) on ice. The protein in samples was quantified by Lowry's method after precipitation by trichloroacetate with sodium deoxycholate. Samples of 20 μg of protein each were loaded on 15% sodium dodecyl sulfate-polyacrylamide gel, and the bands were transferred to a polyvinylidine difluoride membrane. The membranes were preincubated 1 h with 5% nonfat milk in PBS containing 0.05% Tween 20, incubated with 0.25 μg/ml of anti-claudin-6 antibody overnight, and washed in PBS containing 0.05% Tween 20. Then they were incubated with a 1:500-diluted goat anti-rabbit immunoglobulin-conjugated peroxidase-labeled dextrane polymer (rabbit EnVision; DAKO, Carpinteria, CA), and the immunoreactivity was visualized by an ECL Plus Western blotting detection system (GE Healthcare UK Ltd, Little Chalfont, UK).
Kidneys were removed from rats and were processed according to a trichloroacetic acid (TCA) fixation protocol (6). The reasons why we chose the TCA fixation were that staining intensity was more intense in TCA-fixed tissues than in paraformaldehyde-fixed ones and that there was no difference of staining pattern between both the fixatives. Small pieces of renal cortices were soaked in ice-cold 10% TCA for 1 h, washed three times with PBS, and then frozen with liquid nitrogen. The rat kidneys were sectioned at a thickness of 3 μm in a cryostat, treated in 0.2% Triton X-100 in PBS for 15 min, and washed with PBS three times. These sections were soaked 1 h in blocking solution (5% normal goat serum in PBS) and were incubated with primary antibodies. For double-labeled immunofluorescence microscopy, rabbit anti-claudin-6 antibody (5 μg/ml) was mixed with murine monoclonal antibody against ZO-1 or paxillin and was applied as simultaneous primary antibodies. After being washed with PBS, the sections were stained with FITC-conjugated anti-rabbit IgG, rewashed with PBS, and subsequently reacted with TRITC-conjugated anti-mouse IgG. PBS or normal rabbit serum was used as a negative control for the primary antibodies. Immunofluorescence of the sections was observed with a laser-scanning confocal microscope (MRC-1024; Bio-Rad Laboratories, Hercules, CA).
Immunoelectron microscopic observations of rat kidneys were carried out as reported previously (15, 28). In brief, kidneys were perfused and fixed with paraformaldehyde-lysine-periodate (PLP), because the ultrastructure of the tissue was not well preserved in TCA fixation. Tissue blocks (1 mm3) from PLP-perfused kidneys were placed in the PLP fixative for 4 h at 4°C, hydrated, and then embedded in hydrophilic methacrylate resin. The ultrathin sections collected on nickel grids were stained with an immunogold technique. The sections were incubated 1 h with 5% normal goat serum and then incubated 1 h with rabbit anti-claudin-6 antibody (5 μg/ml) at room temperature, followed by incubation with goat anti-rabbit IgG coupled to 15 nm colloidal gold (GE Healthcare, Little Chalfont, UK; 1:100 dilution) for 2 h. After being washed in PBS and fixed with 2.5% glutaraldehyde, the sections were contrasted with uranyl acetate and lead citrate and were then viewed with a Hitachi H-600A electron microscope. PBS, normal rabbit serum, or purified rabbit immunoglobulins (5 μg/ml) were used as a negative control for the primary antibodies.
Perfusion with PS.
Rats were anesthetized with diethylether, the abdominal aorta was cannulated with 23-gauge polyethylene tubing (cat. no. SV-23DLK; Terumo, Tokyo, Japan), and the left kidneys were perfused in situ as described by Seiler et al. (24) with some modifications. In brief, the perfusion schedule consisted of 1) 2 min of perfusion with HBSS at 37°C followed by 2) 10 min of perfusion with HBSS alone or HBSS containing 500 mg/ml of PS (Nakalai Tesque, Kyoto, Japan) at 37°C, and 3) 2 min of perfusion of HBSS at 37°C. For immunofluorescence microscopy, renal cortices were cut into small pieces and were processed according to a TCA fixation protocol. For immunoelectron microscopy, kidneys were perfused additionally with PLP fixative.
Induction of experimental nephrosis.
Puromycin aminonucleoside (PAN) nephrosis was induced in rats by intravenous injection of PAN (Sigma, Saint Louis, MO) at a dose of 5 mg/100 g body wt in PBS. Nephrotic rats were killed at day 4 after the PAN injection. The glomeruli isolated from four or six kidneys were pooled and used as one sample of glomerular protein or RNA. For immunofluorescence microscopy, renal cortices were cut into small pieces and were processed according to a TCA fixation protocol. For immunoelectron microscopy, kidneys were perfused with PLP fixative.
Localization of claudin-6 in neonatal and adult kidneys.
RT-PCR and Western blot analysis showed that claudin-6 expression in adult kidneys was far less than in kidneys of 2-day-old rats at both the transcriptional and protein levels (Fig. 1). Real-time RT-PCR showed that expression of caludin-6 was more than sixfold higher in 2-day-old neonates than in adults. In addition, Western blot analysis confirmed the specificity of anti-claudin-6 antibody reacting with a protein with an apparent molecular mass of 23 kDa.
Claudin-6 distribution in the kidney was examined by double-labeled immunofluorescence by using rabbit anti-claudin-6 antibody in combination with murine monoclonal anti-ZO-1 antibody. Because ZO-1 is concentrated at cell-cell contact sites of podocytes under both physiological and pathological conditions and during development (11, 23), ZO-1 staining was used to locate the glomerular capillary wall. Most of the tubular and glomerular epithelial cells exhibited significant staining for claudin-6 in neonatal kidneys (Fig. 2A). Consistent with the results of Western blot analysis, the claudin-6 staining in tubules declined distinctly in adult kidneys, whereas the staining in glomeruli was well preserved even after the maturation of the glomerulus (Fig. 2B). In all developmental stages of the glomerulus, from the S-shaped body stage to the maturing stage, claudin-6 of podocytes always colocalized with ZO-1. The claudin-6 staining was detected as distinct dots in presumptive podocytes at the S-shaped body stage (Fig. 2, C–C″) and diffuse along the glomerular capillary wall at the capillary loop stage (Fig. 2, D–D″) and in the mature glomerulus in the adult (Fig. 2, E–E″). Paxillin has been demonstrated to be restricted to the cell-matrix contacts at the base of the foot processes of podocytes (12). Double-labeled immunofluorescence microscopy using anti-paxillin antibody showed that claudin-6 colocalized with paxillin in the glomerular capillary wall (Fig. 2, F–F″).
The precise localization of claudin-6 in the glomerular capillary wall of the adult kidney was examined by immunoelectron microscopy (Fig. 3). Immunogold particles for claudin-6 were so rare that only a few particles could be found per cross-section of glomerulus. When ultrathin cryosections were used, the immunogold particles were also extremely sparse in normal glomeruli (data not shown). They were predominantly located on the basal membrane of foot processes (Fig. 3, A and B) and occasionally at the base of slit diaphragms and in focal regions of close cell-cell contacts of podocytes (Fig. 4, C and D). We could count 61 immunogold particles for claudin-6 in glomeruli. The relative distribution of the immunogold particles along the glomerular capillary wall was summarized in Table 1. The immunogold particle distribution was not observed in the negative controls by using normal rabbit serum or PBS instead of rabbit anti-claudin-6 antibody (data not shown).
Localization of claudin-6 in PS-perfused glomeruli and in PAN nephrosis.
In the control HBSS-perfused kidneys, the glomerular capillary wall showed the same immunofluorescence staining for claudin-6 as the normal kidneys without HBSS perfusion (Fig. 2, E–E″). When kidneys were perfused with PS for 10 min, the claudin-6 staining was apparently more conspicuous than that in the control kidneys (Fig. 4, A–A″). Western blot analysis detected no increase in signals for claudin-6 in PS-perfused isolated glomeruli (data not shown).
In the PS-perfused kidney, a substantial number of immunogold particles were found along every capillary loop (Fig. 4, B–D), an observation in sharp contrast to that in the control kidneys with or without HBSS perfusion, in which particles were very rare (Fig. 3). Most of the particles were obviously localized at close cell-cell contact sites of podocytes. We could count 149 particles in the PS-perfused kidneys. The relative distribution of the immunogold particles revealed that the particles were predominantly localized at close cell-cell contact sites in the PS-perfused glomeruli (Table 1).
A single intravenous injection of PAN induced massive proteinuria at day 6 (116 ± 53 mg/day) and at day 10 (201 ± 39 mg/day). No increase or only a slight increase in urinary protein was detected at day 2 or 4 when compared with control rats (2.0 ± 1.0 mg/day). RT-PCR, Western blotting, and immunolocalization studies were performed 4 days after PAN injection when junctions with close contact are most frequently observed in podocytes (15, 28). The levels of the claudin-6 transcripts in the PAN nephrotic glomeruli did not show significant changes: the ratios (×10−4) of claudin-6 to GAPDH in real-time RT-PCR were 0.766 ± 0.174 in the controls (n = 5) vs. 0.894 ± 0.310 in PAN-treated groups (n = 5; P = 0.38; Fig. 5A). Western blotting showed anti-claudin-6 antibody reacting with a protein with an apparent molecular mass of 23 kDa in glomeruli (Fig. 5B). Immunofluorescence for claudin-6 was observed distinctly along the glomerular capillary wall (Fig. 5C). Immunoelectron microscopy revealed that cytoplasmic processes of podocytes were broadened and that junctions with close contact were newly formed between the processes. The junctions were extensively decorated with the immunogold particles for claudin-6 (Fig. 5, D and E).
With the anti-claudin-6 antibody reactive with a protein with an apparent molecular mass of 23 kDa in Western blot analysis (Figs. 1 and 5), which is in good agreement with that previously reported (21), double-labeled immunofluorescence microscopy and, more precisely, immunoelectron microscopy clearly indicated the expression of claudin-6 in podocytes under both physiological and pathological conditions. The present study is the first to report claudin expression in podocytes, because expression of neither claudin nor occludin, proteins known to be components of tight junctions, has been demonstrated in the podocyte (1, 10, 13, 20). In mouse kidneys, we observed the same claudin-6 staining in glomeruli, although one previous study using a different antibody did not (1). In human glomeruli, we could not detect any signal by immunohistochemistry or Western blotting. There are species differences of claudin expression between mammalian glomeruli (Yaoita E, unpublished data).
Immunoelectron microscopy demonstrated that claudin-6 was concentrated at intercellular junctions of podocytes where adjacent cell membranes were in close contact. The junctions with close cell-cell contact were occasionally encountered in normal podocytes and were newly emerged and increased in PS-perfused and PAN-treated kidneys. They have been shown morphologically to contain tight junctions (2, 8, 16, 19, 24). Based on the findings of immunoelectron microscopy coupled with the idea that claudins are major structural components of tight junctions (5), it is reasonable to conclude that claudin-6 is a component of tight junctions in podocytes.
Studies of the PS-perfused kidney have suggested that components of tight junctions are present in proximity to the slit diaphragms before the formation of tight junctions (8, 11). The immunoelectron microscopy in this study showed that immunogold particles were mainly located along the basal cell membrane of foot processes in addition to being located at slit diaphragms in control rats, although the density of the immunogold particles was low. In contrast, perfusion of kidneys with PS brought about a substantial increase of immunogold particles and their predominant localization to the close cell-cell contact sites of podocytes. The difference in distribution and the increase in immunogold particles on perfusion of PS may imply that neutralization of negatively charged groups on the surface of podocytes by polycations causes redistribution of claudin-6 from basal membrane to the sites below the slit diaphragms to form tight junctions. Because a decrease in the negatively charged polyanions on the surface of podocytes has been reported in PAN nephrosis (9, 14), similar redistribution of claudin-6 may contribute to the tight junction formation in the nephrotic podocytes.
Ten-minute perfusion with PS prominently enhanced glomerular staining for claudin-6 in both immunofluorescence and immunoelectron microscopy in spite of no difference in claudin-6 immunoreactivity between HBSS- and PS-perfused kidneys in Western blot analysis. The most plausible explanation of the enhanced immunostaining is the change of antigen distribution. When antigen density is low, antigens distributed diffusely throughout a certain area would cause a weak antibody binding to the area of the tissue, resulting in a small number of antibodies remaining in the tissue. When, in contrast, antigens are concentrated into a restricted area to increase their density, the antibody binding would be strengthened, resulting in a number of remaining immunogold particles. Diffuse distribution of claudin-6 proteins along the basolateral membrane of podocytes and their accumulation into newly formed tight junctions of podocytes may produce the difference in immunostaining between HBSS- and PS-perfused kidneys.
Previous immunoelectron microscopic observations have revealed that tight junctions of podocytes contain coxsackievirus and adenovirus receptor (CAR), Fat1, and podocin as integral membrane proteins and ZO-1 as a cytoplasmic protein (7. 11. 15. 25, 29). Fat 1, podocin, and ZO-1 have been shown to be shared by slit diaphragms. In this study, a significant number of immunogold particles for claudin-6 were detected at the base of slit diaphragms, suggesting that parts of claudin-6 protein are also associated with slit diaphragms. In addition, CAR has been suggested to be distributed along the basolateral membrane similarly to claudin-6 (15). It is likely that all or most of components of tight junctions in podocytes are shared by slit diaphragm. Sharing the same components may facilitate the dynamic changes of intercellular junctions in podocytes.
The kidneys of newborn rats are not fully developed but display a developmental gradient, with immature glomeruli located toward the renal capsule and more mature glomeruli toward the corticomedullary junction. The differentiation of podocytes parallels the development of glomerulus, which is divided into four stages: vesicle, S-shaped body, capillary loop, and maturing stage (17). It has been reported that transcriptional levels of claudin-6 in the mouse kidney decrease remarkably during development (1). We also found the same maturational decrease in claudin-6 in the rat kidney: staining intensity reduced dramatically in tubular epithelial cells in the adult kidney. In contrast, significant claudin-6 staining of podocytes persisted throughout all developmental stages of the glomerulus. The S-shaped body stage is the earliest stage in which the presumptive podocyte can be identified. In this stage, podocytes are connected by the junctional complexes containing tight junctions, which were stained with anti-claudin-6 antibody as distinct dots colocalizing with ZO-1. In the capillary-loop stage, disappearance of tight junctions and appearance of slit diaphragms occur simultaneously. Persistent expression of claudin-6 throughout the capillary-loop stage indicates that redistribution of claudin-6 also participates in the transition from tight junctions to slit diaphragm in the development.
Perspectives and Significance
We demonstrated for the first time that claudin-6 is a component of tight junctions of podocytes of the rat kidney under physiological and pathological conditions. Claudin-6 was also identified as a component of tight junctions in immature podocytes in the neonatal kidney. Claudin-6 exists in proximity to the slit diaphragm, and its redistribution may contribute to formation of tight junctions that emerged in association with nephrotic conditions. Together, these findings strongly suggest that distribution of claudin-6 corroborates the intimate relationship between tight junctions and slit diaphragms of podocytes. The paracellular space of podocytes is sealed by slit diaphragms and tight junctions. These intercellular junctions are the terminal elements of the glomerular filtration barrier. Elucidation of the mechanism underlying tight junction formation in podocytes will provide a new insight into regulation of glomerular filtration.
This work was supported by a Grant-in-Aid for Scientific Research (C) from the Japanese Ministry for Education, Science, and Culture (19590940) to E. Yaoita.
We thank Dr. H. Kurihara (Juntendo University, Tokyo, Japan) for critical discussion and technical help for immunogold labeling using ultrathin cryosections and to Dr. T. Morioka (Niigata University, Niigata, Japan) for technical assistance for real-time RT-PCR.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 the American Physiological Society