Main

Formation of the permanent mammalian kidney involves a complex interaction between the cells of the metanephric mesenchyme and the epithelial cells of the branching ureteric buds. This bidirectional inductive interaction proceeding centrifugally from deeper tissue compartments toward the surface leads to the formation of a closely fixed amount of functioning glomeruli and secretory nephrons(1–4). After induction, further development involves cell growth and spatial arrangements guided in part by finely tuned removal of excess cells by apoptosis and a scheduled appearance of morphologic and functional features ultimately yielding the distinct nephron segments. Changes in the organization of the metanephric mesenchyme and the subsequent epithelial cells as well as local extracellular matrices including the appearance pattern of the individual matrix molecules during nephrogenesis have been extensively characterized(1, 5, 6). Recent identification of several new genes acting coordinately during induction(7–12), together with deregulation experiments of these genes in mice are revealing the complex interplay of the early acting genes in the morphogenetic process. Interestingly, new understanding of the developmental processes has simultaneously revealed new mechanisms of tissue injury and pathogenetic mechanisms of diseases(2). Thus, a regenerative pathway of, e.g. liver cells after tissue damage has been documented in detail(13, 14), and similar mechanisms seem to be involved in various kidney diseases. Wizgall et al.(15) showed a typical expression pattern of distinct genes in the postischemic kidney along the nephron epithelium, evidencing of dedifferentiation pathway, and similar mechanisms have been proposed also for other renal diseases(16–18), although not yet for those limited to the glomerulus.

CNF is a rare autosomal recessive kidney disease of the newborn characterized by massive, treatment-resistant proteinuria(19, 20). Previously, CNF led to death within early months, but the current treatment with early nephrectomy followed by renal transplantation cures all symptoms(21). As no symptoms are found in other organ systems of the treated CNF patients, even after several years of follow-up, in spite of its rarity, CNF is a unique model disease of pure proteinuria and can be used to test various hypotheses of the still poorly known glomerular filtration mechanisms.

CNF kidneys may show areas of irregular tissue organization(19, 20). We have also detected cortical kidney areas with abnormal clustering and remarkably enlarged glomeruli in CNF kidneys as well as areas of atypically organized peritubular compartments (H. Holthöfer and A. Haltia, unpublished observations). Here we used newly identified markers of various differentiation stages of the nephron to characterize the structural abnormalities found especially in the CNF glomeruli in detail. Together with our earlier results evidencing of aberrant differentiation of CNF kidneys(22), the findings suggest abnormal early mesenchymal-epithelial interaction, which may be associated with a generalized functional immaturity, reflected also by proteinuria.

METHODS

CNF patients and tissue samples. Diagnosis of CNF (n= 6) was based on the typical clinical picture (large placenta, massive proteinuria at birth), and exclusion of other types of congenital nephroses and later by the typical pathology at nephrectomy(20, 21). Before nephrectomy, treatment of the patients included daily albumin infusions, aggressive nutrition, and as the weight of 8-9 kg was achieved (at an age of 9-18 mo), nephrectomy of both kidneys was performed(21). The kidneys at nephrectomy were immediately perfused with Ringer's buffer solution (Orion Pharmaceuticals, Espoo, Finland) via the renal arteries and processed for samples for immunohistochemistry (see below).

Cadaver kidneys (ages 3, 12, and 22 y) unsuitable for transplantation due to vascular anatomic reasons (Department of Surgery, University of Helsinki) were used as controls, and processed as the CNF samples.

After removal, the kidneys were cut into two halves along the longitudinal axis across the renal pelvis to get the gross anatomic appearance of the corticomedullary relationship. For immunohistochemistry, samples of cortical tissue as well as samples from the corticomedullary areas were cut with a scalpel into small cubes in a drop of 3.5% paraformaldehyde fixative, and frozen in isopentane cooled with liquid nitrogen until used. Slices were cut routinely at 3-4 μm with a cryostat.

Staining procedure. For immunofluorescence microscopy(23) the tissue sections were washed with PBS (pH 7.2) and incubated with the appropriate dilutions of the primary antibodies overnight at 4 °C. Then the sections were washed thoroughly in PBS and incubated for 1 h with the respective FITC-coupled anti-rabbit (R) or anti-mouse (M) IgG second antibodies (Jackson Laboratories, Westgrowe, Pa), washed, and mounted in a nonfading mounting medium (Mowiol; Calbiochem Corp., La Jolla, Ca).

Antibodies and lectins used. The following antibodies were used to characterize the tissue samples: rabbit (R) antifibronectin (diluted at 1:100 in PBS; Serotec, Oxford, UK), R anti-laminin (1:50; Bethesda Research Laboratories, Gaithersburg, MD), mouse (M) anti-cytokeratin (Pkk1, 1:50; Labsystems, Helsinki, Finland), M anti-human monocytes/macrophages (M 718, 1:10; Dakopatts, Glostrup, Denmark), M anti-human leukocyte common antigen(M701, 1:20; Dakopatts) M anti-stem cell factor (SCF, used at 1:50; R&D Systems, Minneapolis, Mn), M anti-Bcl-2 (1:20; Upstate Biologicals), R anti-Wilms' tumor gene protein product [clone HC17 raised against the amino-terminal 173 amino acids of WT-1, used at 1:250; kindly provided by Dr. Frank Rauscher, The Wistar Institute, Philadelphia; production of the antibody has been described by Amin et al.(24)], M anti-PCNA (Boehringer Mannheim, Mannheim, Germany), R anti-HB-GAM, used at 1:20; kindly provided by Dr. Heikki Rauvala, Institute of Biotechnology, University of Helsinki), R anti-MK [kindly provided by Dr. Takashi Muramatsu, University of Nagoya(25, 26)], and FITC-Triticum vulgaris-lectin (WGA; Vector Laboratories, Burlingame, Ca). To detect cells with compacted nuclei typical of apoptosis, the staining method described by Abrams et al.(27) was used, employing 5 μg/mL acridine orange (Sigma Chemical Co., St. Louis, MO).

An Olympus Ox50 microscope equipped with an appropriate filter system for FITC fluorescence was used for microscopy.

RESULTS

At macroscopic examination, the longitudinal cutting surface of CNF kidneys regularly revealed a cortical thickness of 3-6 mm and also areas of poor demarcation between the cortical and medullary compartments with cortical tissue extending to the medulla (Fig. 1).

Figure 1
figure 1

Longitudinal cutting surface of a kidney from a patient with congenital nephrotic syndrome of the Finnish type (CNF) showing the poor demarcation of cortical area (arrows).

Full size image

Immuno- and lectin histochemical stainings of CNF kidneys revealed normal appearing tubular profiles with morphologically distinguishable epithelial cells of the proximal tubules (cells with apical brush border), loops of Henle, distal tubules (number of tubular profiles within kidney cortex, no brush border), and collecting ducts (variability in cell size and location of nucleus, bifurcation of ducts) as in the controls. When compared with normal kidney cortex (Fig. 2a) characteristic areas of tubular dilatations in CNF as reported earlier(19) were seen (Fig. 2b). In all CNF kidneys studied, tubules were focally surrounded by areas of poorly organized cells (Figs. 2b and 3b). Staining for human lymphocytes and monocyte/macrophages only partially colocalized with these areas (data not shown), which were faintly positive for anti-fibronectin (Fig. 2b), anti-stem cell factor (Fig. 3b; Fig. 3a shows reactivity of normal kidney cortex for stem cell factor) and also for anti-cytokeratin antibodies, whereas no laminin-specific staining was found in these areas (Fig. 4). These areas contained only occasional cells staining with the anti-Bcl-2 (Fig. 5) and anti-PCNA antibodies (Fig. 6) or cells with acridine orange reactive compacted nuclei (Fig. 7). Together these results indicate of no aberrations in the apoptotic or cell proliferation pathways in this compartment. No reactivity with antibodies specific for WT-1, MK or HB-GAM (Fig. 8) could be seen.

Figure 2
figure 2

Frozen section of normal (a) and CNF kidney(b) stained with anti-fibronectin antibodies. Note the dilated tubular profiles and peritubular cellularity positive for anti-fibronectin in b (×120).

Full size image
Figure 3
figure 3

Section of normal (a) and CNF kidney(b), staining for anti-stem cell factor. Note the lack of peritubular staining in a, whereas in b the peritubular cells appear positive (a, ×120; b, ×360).

Full size image
Figure 4
figure 4

Anti-laminin antibodies decorate only the basement membranes in the CNF kidney (cd, collecting duct) (×360).

Full size image
Figure 5
figure 5

Antibodies to Bcl-2 fail to stain cells within the peritubular areas (indicated by arrows) of CNF kidneys(×260).

Full size image
Figure 6
figure 6

Antibodies to proliferating cell nuclear antigen (PCNA) decorate only occasional cells (indicated by arrows) within the CNF kidney (×260).

Full size image
Figure 7
figure 7

Acridine orange staining of CNF kidney detects only occasional cells with compacted, acridine orange-positive nucleus(arrow) (×480).

Full size image
Figure 8
figure 8

Frozen section of CNF kidney stained for HB-GAM. No reactivity with the peritubular areas is seen (delineated by arrows)(×180).

Full size image

Several glomeruli were found in each cortical section in all kidney samples, but their arrangement differed characteristically from that of the control kidneys: densely clustered glomeruli separated by narrow Bowman's capsules were seen (Fig. 9) in all the CNF samples especially at the corticomedullary zone. In addition, large glomeruli suggesting fusion of two to three individual glomeruli were found (Fig. 10). These “fused” glomeruli were typically surrounded by a continuous laminin-containing Bowman's capsule. The cellular composition within these fused glomeruli appeared balanced, with no prominence of podocyte, mesangial or endothelial elements as decided by morphology and reactivity for WT-1 antibodies and WGA lectin (Fig. 11). In these CNF glomeruli, immunoreactivity for WT-1 protein (Fig. 12b; Fig. 12a shows reactivity of normal glomeruli for WT-1) appeared exclusively in the visceral epithelial cells in a pattern as in normal kidneys; the WT-1 positive cells with nuclear but also cytoplasmic reactivity were seen facing the urinary space. No signs of cell proliferation as demonstrated by PCNA positive cells were seen in the visceral epithelial areas of the CNF glomeruli although in a few glomeruli PCNA positive cells were detected in the more central mesangial zone. No cells with aberrations in the apoptotic pathway (positivity for Bcl-2, compacted nuclei positive for acridine orange) could be seen within CNF or control kidney glomeruli. In addition, the functionally important anionic charge of CNF glomeruli appeared normal as decided by the comparable intensity and distribution of WGA lectin binding sites within the epithelial cells and glomerular basement membranes of the “fused” glomeruli (Fig. 11), compared with those of the controls.

Figure 9
figure 9

Area of a CNF kidney showing compacted glomeruli separated by Bowman's capsules, staining for anti-laminin antibodies. G, glomerulus (×260).

Full size image
Figure 10
figure 10

FITC-WGA staining of CNF kidney cortex reveals large glomeruli (G) appearing as fused (×120).

Full size image
Figure 11
figure 11

FITC-WGA staining of CNF kidney shows a remarkably enlarged, WGA-positive glomerulus. Compare also the size of the tubular profiles (×260).

Full size image
Figure 12
figure 12

Staining for Wilms' tumor antigen WT-1 in control kidney (a) and CNF kidney cortex (b) shows similar reactivity in the glomerulus (G) (×260).

Full size image

DISCUSSION

Here we wanted to apply newly established markers for early and later stages of nephron differentiation to analyze whether evidence of either arrest of differentiation or dedifferentiation could be found in the massively proteinuric kidneys of congenital nephrotic syndrome patients. The results revealed enlarged single glomeruli but also areas of tightly compacted clusters of glomeruli of normal size. These results are in agreement with the results of Tryggvason and Kouvalainen(28), who could show an increase in the total number of glomeruli in CNF kidneys. The number of functional nephrons and glomeruli is determined by the number of branches of the ingrowing ureteric buds. These then interact with the metanephric mesenchyme to reach the final number of glomeruli(1, 4, 29). Thus, the finding of clustered glomeruli may suggest a failure in the coordinated mesenchymoepithelial interaction during nephrogenesis in CNF. The terminally differentiated podocytes(30) or other intraglomerular cells did not, however, show signs of continued cell proliferation or excess cell death by apoptosis in these glomeruli. Furthermore, the peritubular cellular areas expressing markers of early differentiation (fibronectin, stem cell factor, cytokeratin) further support the conclusions of arrested differentiation in this disease. The results with Bcl-2 and PCNA antibodies as well as acridine orange staining suggest that excess rescue from apoptosis or increased cellular proliferation, mechanisms normally controlling the ultimate cell number of the mature kidney(31) may not be perturbed in CNF kidneys. Thus, the proteinuria of CNF could be due to a developmental arrest at a stage characterized by a nonfunctional glomerular filtration barrier.

The frequency of CNF in Finland has been estimated to be 8 per 100 000 newborn with only sporadic cases outside Finland or without Finnish descendence. However, it is evident that the recently recognized gene locus of CNF on the long arm of chromosome 19(32) is shared with cases outside Finland(33, 34) but the CNF gene has not been identified as yet. In support to our interpretation of a developmental arrest in CNF the treatment-resistant proteinuria in this disease starts already in utero, as evidenced by the exceedingly high content of α-fetoprotein in amniotic fluid and in maternal circulation(35). The typical morphologic finding in CNF glomeruli is the flattening of podocytes with retraction of the foot processes(20); whether this is secondary to the heavy proteinuria or reflects immaturity of podocytes is not known.

The MAb for WT-1 (clone HC17, against the aminoterminal 173 amino acids) showed remarkable and unexpected cytoplasmic reactivity in podocytes, in addition to the nuclear staining, although WT-1 has been considered a strictly nuclear protein(11). One possibility for this might be that a specific splice variant of WT-1 is detected. However, cytoplasmic reactivity of WT-1 has been repeatedly observed though not systematically reported [see e.g. Amin et al.(24) and Menssen et al.(36)]. Furthermore, the protein itself, before transport to the nucleus is produced in the cytoplasm and could thus explain its cytoplasmic detection in podocytes. A disease-specific feature of WT-1 expression does not appear probable, because the control kidneys used also showed variable cytoplasmic reactivity in addition to the nuclear one. A phenomenon due to a fixation artifact does not appear true, either, as additional fixation experiments showed similar reactivity.

The possibility that abnormal early inductive events may lead to a distinct functional abnormality of glomeruli is intriguing but well supported by the findings of gene deregulation studies in mice. The early developmental gene Pax-2 overexpression in mice results in proteinuria and morphologic changes typical for congenital nephrosis(8). Our results have shown that Pax-2-specific mRNA as well as the corresponding protein product is intact in CNF kidneys(37). Furthermore, Kestiläet al.(38) have excluded any gene defects in the human homologue of Pax-2 in CNF. Mice lacking the functional S-laminin gene similarly produce a phenotype closely resembling that of human congenital nephroses(39). However, the recently recognized disease locus of CNF excludes also S-laminin to be defective in this disease(32). We have earlier shown that the cell type-specific glomerular gangliosides found transiently during early stages of glomerulogenesis(23) remain atypically expressed in the CNF kidney glomeruli(22), supporting the possibility of developmental arrest in CNF glomeruli.

In conclusion, morphologic changes suggesting abnormal mesenchymal-epithelial tissue interaction was found in the CNF cortical kidney and this was also supported by the markers of nephron differentiation stages used. However, the markers used did not appear optimal to characterize the glomerular functional disturbance in CNF kidneys, proteinuria, which could be the consequence of such a proposed developmental arrest.