Tracer Studies in the Rat Demonstrate Misdirected Filtration and Peritubular Filtrate Spreading in Nephrons with Segmental Glomerulosclerosis

Experiments with Lissamine Green

Lissamine green (Lissamine Green SF, Gurr Ldt., London, UK) is a water-soluble dye that is handled by the kidney like inulin (7). Under pentobarbital anesthesia, a 3% lissamine green (prepared according to reference 8), 0.9% NaCl solution was slowly administered into the tail vein in doses between 0.25 and 0.75 ml/100 g body wt. To localize lissamine green in tissue sections, we modified a method used previously to measure electrolytes in tissue sections (9). Immediately after the dye was applied, the left kidney was removed and shock-frozen in liquid N₂ cooled melting isopentane. Blocks of cortical tissue were freeze-dried at -50°C, embedded in paraffin, sectioned into 12- μm-thick sections, fixed to glass slides without water contact, deparaffinized, and observed by light microscopy (LM).

Experiments with Ferritin

Ferritin is an iron-containing protein of high molecular weight (up to 750,000 Da) (10). In healthy kidneys, ferritin is not filtered but, when exogenously administered, passes in small amounts through the glomerular basement membrane (GBM) and eventually is picked up through endocytosis by podocytes (11) and has access to the mesangium (12,13). In injured glomeruli with damage to the filtration barrier, ferritin passes in bulk through a leaky GBM (14,15,16) and subsequently is taken up through endocytosis by podocytes or travels with the tubule fluid. From there, most of the ferritin is reabsorbed within the proximal tubule and the rest is excreted with the urine (17).

In the present experiments, we used a commercial preparation of native, i.e., anionic, horse spleen ferritin (F-4503; Sigma, Heidelberg, Germany) containing 100 mg protein/ml. Because it is known that such solutions may contain cadmium, the preparation was dialyzed for 24 h against three changes of a 0.1 M ethylenediaminetetraacetate in 0.07 M phosphate buffer, (pH 7.2). Animals that were treated with these preparations showed no ill effects. The solution was administered slowly into the tail vein in anesthetized rats in a dose of 100 mg ferritin/100 g body wt. Three protocols were tested: removal of the kidneys (1) immediately up to 2 min, (2) 20 to 90 min, and (3) 24 h after the injection; each group included a control rat. The clearest labeling in the tissues was seen with the second protocol.

With respect to the handling of the renal tissue, previously applied protocols were used (15,18,19). The kidneys were removed, and slices were fixed by immersion in 4% formaldehyde and embedded in paraffin by standard procedures. Several series of 8- μm sections were cut. Ferritin was visualized in the sections by the potassium ferrocyanide method for iron (Prussian blue reaction) (20). The sections were counterstained with nuclear-fast red and observed by LM.

In some cases, the serial sections were stained alternatively with Prussian blue followed by Goldner trichrome (21). In other cases, after the documentation of the ferritin labeling, the sections were stained a second time by Goldner's trichrome technique.

Electron Microscopy

Tissues that were fixed by immersion with formaldehyde were also used for transmission electron microscopy (TEM). Small pieces (approximately 2 × 2 mm) of renal cortex were postfixed in 2% glutaraldehyde, processed, and embedded into Epon by standard procedures. In addition, perfusion-fixed material was used. For this purpose, in some of the animals only one kidney was fixed by immersion and the second kidney was fixed by vascular perfusion with a 2% formaldehyde solution as described previously (5). Pieces of cortical tissue were postfixed in 2% glutaraldehyde and processed as described above. From blocks of both immersion and perfusion-fixed material, 1-μm section series were cut, stained with methylene blue and Azur II (22), and observed by LM. From sites of interest, ultrathin sections were cut with a diamond knife, stained with lead citrate and uranyl acetate, and observed by TEM.

Experiments with Lissamine Green

In nephrons of control animals and in healthy (nonsclerotic) nephrons of FHH and MN rats, lissamine green, in addition to a very faint staining of extracellular matrices, was seen as a diffuse material in the lumen of proximal tubules and, in more prominent deposits, in distal tubules and collecting ducts (Figure 1a). In damaged nephrons, the distribution of the dye depended on the nature and the severity of the lesion. In the case of the presence of a tuft adhesion to Bowman's capsule, there was dense accumulation of lissamine green in the tuft adhesion, often together with some staining of the mesangial matrix in “intact” tuft portions (Figure 1, b through e). The staining of the adhesion was not homogeneous but varied from area to area, suggesting that the dye binds to some extracellular material. At the periphery of an adhesion, the staining extended into expanded spaces associated with the basement membrane of the parietal epithelium, frequently showing up as a complete circumferential staining of the glomerular profile sharply demarcated from the interstitium. Moreover, in many of the injured glomeruli, the dye—via the urinary pole—encroached onto the outer aspect of the corresponding tubule and encircled the tubule with an intensively green-stained collar (Figure 1, b through e). Again, these green envelopes were sharply delimited from the interstitium, and as seen in Figure 1, b and d, there was also a clear separation from the tubule lumen.

Intrarenal Distribution of Ferritin as Observed by LM after the Prussian Blue Reaction

We used three protocols that differed in the time span between the injection of ferritin and removal of the kidneys (up to 2 min, 20 to 90 min, and 24 h). In both rat strains, all three time courses showed fundamentally the same results with respect to the distribution of ferritin; the lag times between 30 and 90 min yielded the most intensive staining.

In kidneys of control rats and in healthy nephrons of FHH and MN rats, ferritin was not encountered in tubules (no traces in the lumen, no endocytotic uptake) and did not escape in visible amounts from peritubular capillaries (Figure 2). However, in a small proportion of glomeruli, some capillaries exhibited a faint outline of staining, indicating penetration of ferritin into the GBM as well as more intensive accumulations in axial regions suggesting uptake by mesangial cells or macrophages or both (Figure 2).

As known from previous studies (14,15,18,17) and as observed in this study, in injured glomeruli with damage to the filtration barrier, ferritin passed through the barrier, was reabsorbed by proximal tubules, and may appear in the urine. In partially injured nephrons with an “intact” glomerular tuft (or at least “intact” tuft portions) showing a strong Prussian blue staining of capillary walls (suggesting that ferritin has penetrated into a leaky GBM), the corresponding proximal tubules exhibited a granular accumulation of ferritin, indicating endocytotic uptake (Figure 3).

In nephrons with advanced injury, two patterns must be distinguished: (1) nephrons with accumulation of ferritin in the tuft adhesion to Bowman's capsule and (2) nephrons without accumulation of ferritin.

Tuft Adhesions with Ferritin Accumulation. Ferritin accumulation in adhesions was identified by an irregular blue staining of the extracellular matrix (Figures 4 and 5); in areas of hyalinosis, the staining was more intense, suggesting enhanced accumulation of ferritin in the hyaline material. Generally, in such glomeruli, ferritin accumulation was also dense in the expanded basement membrane of the parietal epithelium (PBM) in a pattern that was either restricted to a certain segment associated with the adhesion or spread over the entire circumference (Figures 4 and 5). On the interstitial surface, the blue-stained PBM was sharply delimited. In the histologic controls, i.e., in the same sections that, subsequent to the documentation of the ferritin distribution, were stained by Goldner's technique (Figure 4), it could be seen clearly that the sites of ferritin accumulation corresponded to the “sclerotic” portions of the tuft enclosed in paraglomerular spaces on the outside of the parietal epithelium.

Alongside the glomerulotubular junction (urinary pole), the glomerular ferritin staining extended to the outer aspect of the proximal tubule. This extension of the staining was restricted to one side of the glomerulotubular junction (not shown) or involved the entire circumference (Figures 4 and 5); the staining either ceased immediately at or after the junction (not shown) or extended further along the proximal tubule (Figure 4), frequently involving a group of tubular profiles in the vicinity of the affected glomerulus (Figure 5). As documented in serial sections, these profiles belonged to the same injured nephron; eventually, the entire proximal convolution of an injured nephron was enclosed by a blue staining cylinder. As seen subsequently in trichrome-stained sections (Figures 4 and 5) as well as in corresponding transmission electron micrographs (see below), these ferritin-stained regions represented subepithelial peritubular spaces.

In addition to the (more or less) homogeneous ferritin staining of extracellular compartments (paraglomerular and subepithelial peritubular spaces), ferritin was frequently encountered in intracellular granules in proximal tubule cells of injured nephrons. As noted above, intracellular deposits in proximal tubule cells may derive from luminal uptake (Figure 3). Moreover, in nephrons with peritubular ferritin staining, those intracellular deposits were also found in cases of obstruction of the tubular lumen, suggesting uptake from the peritubular side (Figures 4 and 5).

Tuft Adhesions without Ferritin Accumulation. In the case of tuft adhesions without ferritin staining (Figures 3 and 6), the expanded parietal basement membrane and the subepithelial peritubular spaces all were devoid of ferritin, even in cases with injury to the filtration barrier of the “intact” tuft portion and leakage of ferritin into the tubule fluid (Figure 6). In those cases, ferritin was taken up into proximal tubule cells by endocytosis (leading to its degradation (17)), but there was no evidence of paracellular transport to the subepithelial side (Figure 6). The intercellular junctions between proximal tubule cells, even in more advanced stages of tubule damage (see Figure 7), remained intact. These findings suggest that the major route of access of ferritin to those peritubular spaces (Figures 4 and 5) is through passage along the glomerulotubular junction. Nephron remnants that consisted of globally collapsed glomeruli and tubular matrix remnants did not show any accumulation of ferritin (not shown). Nephrons with ferritin accumulation in the adhesion and in peritubular spaces and those without ferritin were found side by side in the same section.

Spatial Relationships between Degenerating and Intact Nephrons. The ferritin staining represented a conspicuous marker of injured nephrons. This was helpful when tracing an injured nephron in serial sections; its tubular profiles could be distinguished easily from those that belonged to an adjacent healthy nephron As shown in Figures 4 and 5, severely injured and healthy tubules may be located in close contact to each other; structurally intact profiles of proximal tubule may be fully surrounded by heavily damaged tubules and vice versa.

Distribution of Ferritin Depicted by TEM

Ferritin was seen directly by TEM as black dots smaller than 10 nm in diameter representing the iron core of the molecule. Immersion and perfusion-fixed tissue samples were studied. Compared with immersion fixation, perfusion fixation led to better tissue preservation, including the prevention of tubular collapse. However, accumulations of ferritin in extracellular locations tended to be washed out by the perfusion procedure.

At all sites where ferritin accumulation was observed using the Prussian blue reaction by LM, ferritin was also found by TEM. Ferritin leakage through an injured glomerular filter into an intact tubule was followed by endocytosis by proximal tubule cells and accumulation within lysosomes (Figures 3a and 7a). The LM data (Figures 4 and 5) strongly suggested that in advanced stages of nephron degeneration there was tubular uptake of ferritin from the subepithelial side. This was supported by the TEM findings. Figure 7, a through c, depict three stages of tubular degeneration with established subepithelial peritubular spaces: in Figure 7a, the tubular lumen was still patent, in Figure 7b, it was obstructed, and in Figure 7c, a solid assembly of just a few cells remained. The subepithelial spaces consisted of a cylinder of an electron-dense matrix that appeared as a tremendously expanded basement membrane and, in addition (Figure 7c), of electron lucent spaces that seemed to develop by the separation of the matrix cylinder from the epithelial remnant. Toward the interstitium, the subepithelial peritubular spaces were delimited by a layer of fibroblast processes that tightly adhered to the matrix cylinder. By TEM, ferritin was encountered densely accumulated in cytoplasmic granules (Figure 7, a through c) and homogeneously distributed in the subepithelial matrix cylinders (Figure 7, d through f). In conjunction with the LM findings, these TEM observations suggest that the subepithelial peritubular spaces are the distribution routes of a misdirected filtrate. In totally obstructed tubules (Figure 7b) and even in epithelial residues (Figure 7c), there were ferritin-filled granules, suggesting the possibility that endocytotic uptake in these structures was probably via uptake from the subepithelial side.