Main

HUS, a clinical syndrome consisting of hemolytic anemia, acute renal failure, and thrombocytopenia, is a major cause of pediatric acute renal failure(1). The pathogenetic status of HUS is preceded by a prodrome of bloody diarrhea in infants and young children(2). Since the beginning of the 1980s, it has been reported that STEC infections are the main cause of HUS(2,3).

Stx, also called verotoxin, was first discovered as a novel cytotoxin in vero cells, which were derived from the kidney of the African green monkey(4). Since the prototype toxin Stx1 was purified in 1983(5), two types of Stx, Stx1 and Stx2, and two Stx2 variants (subtypes), Stx2vh and Stx2vp, have been identified(6). The Stx are a complex of proteins consisting of A and B subunits(6). The A subunit is a 30-kD cytotoxic chain, which has ribonucleic acid N-glycohydrolase activity and cleaves a specific adenine residue on the 28S ribosomal ribonucleic acid, resulting in inhibition of protein synthesis(7–10). The 7-kD B subunit forms a pentamer and binds to the terminal digalactose of Gb3/CD77, a functional receptor for Stx(11–13), with high affinity(7,14). It is thought that the holotoxin binds via B subunits to the cells expressing its receptor Gb3/CD77, is internalized, and then damages the cells via inhibition of protein synthesis by the A subunit(7).

Based on experimental(15–19) and histopathologic observations(20), it is widely believed that endothelial cell damage mediated by Stx plays a central role in HUS(15,21). However, recent studies have shown that a specific portion of tubules(22,23) and glomeruler mesangial cells(24) express receptor and bind to Stx as well as endothelium. In addition, we have recently found that distal tubular epithelial cells in the kidney express Gb3/CD77 and are susceptible to Stx-mediated cytotoxicity via apoptosis(25,26). These findings suggest that endothelial cells are not the sole target for Stx-mediated cell injury.

Although experimental data indicate Stx-mediated direct injury of a variety of cell species in the kidney(5,15–19,21,24–27), details of the pathogenesis of HUS remain unclear largely because the kinetics of Stx in humans is not known. Although Stx, secreted by STEC in intestine, is thought to migrate to the kidney via blood flow, direct evidence has not been demonstrated that shows the presence of Stx either in the blood flow or in the damaged organ(3,27).

In this study, we used immunohistochemistry to detect the presence of residual Stx1 and Stx2 in the renal sections of an autopsied fatal case of HUS. Stx was found to bind to the specific portion of distal tubules. Our results suggest that during STEC infection, Stx migrates to the kidney from the intestine and binds to renal tubular epithelium, a target of Stx.

METHODS

Case presentation. A 1-y 9-mo-old girl primarily showed enteric symptoms of diarrhea and fever (38 °C) on July 16, 1996. On the 2nd d of the disease, bloody diarrhea occurred. On d 6 of the disease, enterocolitis progressed to HUS consisting of hemolytic anemia, thrombocytopenia, and symptoms of renal failure (severe edema, anuria, and hypertension), and the patient was therefore admitted to the hospital for intensive care. STEC O157:H7 (Stx1 +, Stx2+) was isolated in stool culture on d 3 and 7 of the disease. Peritoneal dialysis was performed during her entire admission period. Her general condition remained stable, although anuria continued until she died. Peritonitis due to perforation of the gall bladder occurred and progressed to respiratory distress. The patient died in d 25. Renal function was not improved during her clinical course. Autopsy was performed, and renal tissue was frozen. Her laboratory data during admission are summarized in Table 1.

Table 1 Laboratory findings of the case associated with Escherichia coli O157 infection
Full size table

Immunohistochemical analysis. Frozen sections were prepared and fixed in 100% acetone at 4°C for 15 min. After washing with PBS and blocking with 5% normal rabbit serum for 20 min at room temperature, samples were overlaid with monoclonal anti-Stx1 (13C4, ECACC No. 95032701)(28) or anti-Stx2 antibody (VTm1.1)(29) for 30 min at room temperature. These anti-Stx antibodies were used at a 300-fold dilution of ascites forms. Isotype-matched mouse monoclonal immunoglobulin was used as a control. After washing with PBS, samples were treated with horseradish peroxidase-conjugated rabbit anti-mouse antibody (DAKO). A final wash with PBS was followed by color development using 3,3′ -diaminobenzidine, tetrahydrochloride as described previously(30). Expression of CD molecules on the renal tissue was detected by using anti-CD10 (IF-6)(31), anti-CD24 (L30)(32), anti-CD34 (QBEnd10, Coulter/Immunotec, Inc., Westbrook, MA), and anti-Gb3/CD77 (38.13, Coulter)(33) antibodies. The method of staining for CD molecules was identical to that described above. As negative control, six renal tissues from patients without HUS, including IgA nephropathy, nephrotic syndrome, and Wilms' tumor (nontumor part), were also examined. All materials were used after obtaining informed consent and have been approved by our Institutional Review Board.

RESULTS

Hematoxylin and eosin staining of renal sections showed focal necrosis (Fig. 1A) and tissue regeneration (Fig. 1, B and C). The regenerating tissue mostly consisted of small tubules (Fig. 1B), but dilated tubules could also be identified (Fig. 1C).

Figure 1
figure 1

Histologic findings of the kidney of the patient. With hematoxylin and eosin staining, focal necrosis including glomeruli was observed (A). In addition, regenerating tissue consisting small tubules (B) as well as dilated tubules (C) was also found. Magnification, × ∼100 (A and B) and × ∼400 (C).

Full size image

In frozen sections, an Stx1 MAb clearly stained epithelial cells of dilated tubules (Fig. 2A). With precise observation, it was evident that staining was mainly located on the apical portion of epithelia. Monoclonal anti-Stx2 antibody also stained the epithelia of dilated tubules with an identical staining pattern (Fig. 2C). No staining was observed by using an isotype-matched control antibody (Fig. 2D). Six kidney tissues obtained from patients without HUS were not stained by anti-Stx1 nor Stx2 (an example is presented in Fig. 2, E and F). Because the specificity of these antibodies has been confirmed previously(26,28,29, Uchida H et al., unpublished observations.), these results show the specific detection of Stx1 and Stx2 on the renal tubules. In addition to distal tubules, a few damaged glomeruli (Fig. 3A) and endothelia (Fig. 3, A and B) were also stained with anti-Stx antibodies.

Figure 2
figure 2

Identification of Stx in the renal tubules of the patient. MAb against Stx1 stained epithelia of dilated tubules (A) but not glomeruli (B) in the renal tissue of the patient. Anti-Stx2 also stained dilated tubules (C) as anti-Stx1. No staining was observed when control antibody was used on the same tissue preparation (D). Normal renal tissue, obtained from a surgical resection of Wilms' tumor of a 5-y-old patient, was not stained by either anti-Stx1 (E) or anti-Stx2 (F) antibody. Magnification, ×∼400.

Full size image
Figure 3
figure 3

Identification of Stx in damaged glomerulus and endothelia of the patient. MAb against Stx1 also stained a few of damaged glomerulus (A) and endothelia (A and B). By using anti-Stx2 antibody, identical result was observed (data not shown). Magnification, ×400.

Full size image

It has been reported that the cellular components of the kidney can be classified according to the pattern of expression for CD leukocyte antigens(30,34–36). The renal distribution of CD molecules was tested to localize the sites of Stx binding. CD10 was reported to be expressed on glomeruler and proximal tubular epithelia(30,34). In the patient's tissue, CD10 was expressed on glomeruli and on proximal tubules that were twisted in shape (Fig. 4A). These CD10-positive tubules were mostly distributed around glomeruli in this specimen. In addition, no staining of CD10 was observed in dilated tubules (Fig. 4B), which were positive for Stx staining. In normal tissue, CD24 is expressed on a part of Henle's loop, distal tubules, and Bowman's capsule(30,35). In the patient's sample, CD24 was expressed on a considerable number of tubules (Fig. 4C) as well as on Bowman's capsules (data not shown). These CD24-positive tubules were mostly small, dilated, and round (Fig. 4D). Collecting tubules, which assembled to form a nest, were negative for CD24 (Fig. 4C). Gb3/CD77, a functional receptor for Stx(11–13), was expressed on dilated tubules and on some of the tubules that were small in diameter (Fig. 4F). It was thus likely that some of the CD24-positive tubules expressed Gb3/CD77. Gb3/CD77 was also positive on endothelial cells of small vessels. CD34 is a hematopoietic progenitor marker but is also known to be expressed on vascular endothelial cells(36). Consistent with this, small vessels were visualized by CD34 in the patient's kidney (Fig. 4E).

Figure 4
figure 4

Distribution of CD molecules in the renal tissue of the patient. Distribution of CD molecules in the renal tissue of the patient was examined by immunohistochemistry. CD10 was expressed on glomeruli and on proximal tubules (A) but not on dilated tubules (B). CD24 was expressed on tubules (C), including dilated ones (D). CD34 was positive on endothelial cells surrounding tubules (E). CD77 was present on some tubules with large and small diameters (F). Magnification, ×∼100 (A and C) and ×∼400 (B, D, E, and F).

Full size image

The sites of Stx binding could be specified in relation to the distributions of these CD molecules. It was thus apparent that there was no binding of either Stx1 or Stx2 on proximal tubules in the specimen because the patterns of distribution of Stx and CD10 were completely different. Rather, Stx was mostly localized to dilated distal tubules, which were positive for CD24 and Gb3/CD77.

DISCUSSION

We have shown the specific binding of anti-Stx1 and anti-Stx2 MAb to epithelia of a portion of tubules, which corresponded to distal tubules in the renal section of a patient with HUS with STEC infection. The reagents used in this investigation are highly specific, as confirmed previously (26,28,29, Uchida H et al., unpublished observations) and by the fact that control kidneys were not stained. Although it has been postulated that Stx produced by STEC in the intestine migrate into kidney via the blood stream(3,27) and then injure a variety of renal cell species(5,15–19,21,24–27,37), this has not been proven. To our knowledge, this is the first report describing the in situ immunostaining of Stx1 and Stx2 in the kidney of a patient with STEC infection, indicating that Stx migrates into the kidney during STEC infection.

The tubules on which Stx were deposited were characterized by comparing the distribution of various CD molecules. It was apparent that Stx bound to some of the cells that were positive for CD24, a marker for the ascending part of Henle's loop, and distal tubules(30,35). In addition, Stx staining was clearly positive on some of the tubular epithelial cells that expressed Gb3/CD77, a functional receptor for Stx(11–13). It has been reported that tubular epithelia expressed Gb3/CD77 and bound Stx(22,23), although this relationship was not described precisely. The Gb3/CD77-positive tubules were similar to the CD24-positive ones in this case. Thus, it is likely that a portion of distal tubules expressed Gb3/CD77. Our preliminary study on normal renal tissue obtained from children with Wilms' tumor also indicated that a portion of CD24-positive epithelia expresses Gb3/CD77 and binds Stx (Mori T et al., unpublished observations). These data correlate with our previous report that Stx can bind to cells derived from distal tubular epithelial cells in the kidney and directly injure them via induction of apoptosis(25,26), suggesting that the distal tubule is another primary target for Stx.

Certain important questions arise from our observations. Were Stx-bound tubules in this sample undergoing cellular damage? Is it really possible for Stx to be detected so late? We cannot currently address these questions. Indeed, it is curious that Stx was detected on tissue obtained as late as d 25 of the patient's illness because it is unlikely that STEC was still present. In addition, most of the cells should be destroyed if they were attached to the toxin. Although we did not have a chance to examine the patient's renal tissue at the early stage of her disease, the laboratory data suggest that the kidney was severely damaged and the histologic examination showed considerable regeneration at autopsy.

One possible interpretation concerning the unresponsiveness of the cells to toxin may be that Stx can bind to Gb3/CD77 on cells but is not internalized. Details of this mechanism are still unknown, but the following reports may support this idea. First, Jacewicz et al.(38) reported that Gb3/CD77 is required for Stx binding but not sufficient for cytotoxicity and suggested the existence of a toxin translocation mechanism linked to surface glycolipids. Thus, Stx binding without internalization could be possible where the toxin translocation mechanism is absent or dysfunctional while Stx receptor is present. Second, it was reported that Gb3/CD77 receptors have fatty acid heterogeneity, which modulates receptor function(39–41). Thus, the diversity of Gb3/CD77 receptors may account for the different behavior of Stx after binding. It was reported that Stx1 binds to human monocytes or mouse peritoneal macrophages via a Gb3/CD77 species that is different from that found on endothelial cells and induces cytokine synthesis but not cell death(42,43). In addition, the cytotoxic effect of Stx does not always parallel the expression level of Gb3/CD77 receptors and Stx binding. For example, it is well known that HUS due to STEC infection occurs more frequently in infants and young children than in adults among the patients, even though the total renal contents of Gb3/CD77 is lower in children and the binding capacity of Stx is proportional to the level of Gb3/CD77 expression(44). Moreover, we also reported a significant difference between Stx1 and Stx2 in terms of efficiency for cytotoxicity in ACHN cells even though their binding affinity to ACHN cells is comparable(25).

Taken together, these results suggest that a small level of Gb3/CD77 receptor heterogeneity may affect the internalization efficiency of Stx. If this is the case, Stx may bind to renal tubular cells that express Gb3/CD77 but fail to be internalized in some cells. Although further study must be conducted, the determination of the mechanism of internalization of Stx after binding to the cell surface, especially as this relates to Gb3/CD77 receptor diversity, should provide new direction of research as well as of clinical management to overcome this disorder, HUS.