Detection of Verocytotoxin Bound to Circulating Polymorphonuclear Leukocytes of Patients with Hemolytic Uremic Syndrome

The epidemic form of hemolytic uremic syndrome (HUS) is the most common cause of acute renal failure in children (1,2). Endothelial damage of predominantly glomeruli and to a lesser extent arterioles of the kidney appears to play an important role in the pathogenesis of HUS. In severe cases, however, endothelial cell damage is not limited to the kidney but extends to other organs, such as the brain and pancreas (3). In 90% of HUS cases, infection with Escherichia coli producing a verocytotoxin (VT) (also termed “Shiga” or “Shiga-like toxin”), especially serotype O157:H7, has been strongly implicated. E. coli can produce VT-1, VT-2, or both. Most common among patients with HUS are infections caused by VT-2-producing E. coli (4,5). VT can cause inhibition of overall protein synthesis after binding to the globotriaosylceramide (Gb3) receptor, which is found on endothelial cells after stimulation with inflammatory mediators (6,7). Serologic testing of patients with HUS for antibodies to the lipopolysaccharide of E. coli O157:H7 is often used to detect infection by VT-producing E. coli O157 (8,9,10). However, VT itself has never been detected in the blood of patients with HUS.

Recently, we investigated the binding of VT-1 and VT-2 in whole blood (11). In that study, we demonstrated that VT binds rapidly and exclusively to polymorphonuclear leukocytes (PMN) when incubated in whole blood in vitro. This binding occurs via selective binding to a specific receptor on PMN. Thin-layer chromatography demonstrated that the receptor exhibited an R_f value between those of Gb4 and Gb5. In addition, the receptor on PMN exhibited a 100-fold lower affinity, compared with that of the Gb3 receptor, which is found on endothelial cells. PMN previously loaded with VT were able to pass the ligand VT to tumor necrosis factor-α-stimulated glomerular microvascular endothelial cells in vitro, which then caused inhibition of protein synthesis (11). In this study, we demonstrate for the first time, to our knowledge, the presence of VT-2 in the systemic circulation of patients during the acute phase of HUS. Here we confirm that VT binds to PMN, as suggested by our in vitro studies, supporting our theory that PMN are responsible for transferring the ligand VT from the intestine to target organs.

Materials and Methods

Results

On the basis of the results of our in vitro work (11), we focused in this study on the binding of VT-2 to PMN. Indirect immunofluorescence assays demonstrated positive staining for VT-2 on PMN from nine of 10 patients in the acute phase of HUS (Table 1). The results of an indirect immunofluorescence assay for one representative patient are presented in Figure 1. VT-2 binding was observed only to PMN. To confirm that PMN were the only cells binding VT in blood, VT binding to purified blood cells from patients with HUS was studied. As suggested by our previous findings for blood cells from healthy donors (11), VT-2 bound only to PMN and not to lymphocytes, monocytes, or erythrocytes from patients with diarrhea-associated (D+) HUS in the acute phase of the disease.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Indirect immunofluorescence assay of verocytotoxin-2 (VT-2) bound to polymorphonuclear leukocytes (PMN) for one representative patient (patient 9) in the acute phase of hemolytic uremic syndrome (HUS). Magnification, ×1000.

FACS analysis was used to quantify the indirect immunofluorescence results. No VT-2 binding was detected in control blood samples from any of the healthy volunteers (Figure 2, A, C, and E). In addition, no binding of VT-2 was observed for six patients with infectious diseases (see above).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Flow cytometric analysis for detection of VT-2 bound to PMN. (A and C) PMN isolated from control blood (from two different healthy volunteers) and incubated with a monoclonal antibody against VT-2, followed by incubation with FITC-conjugated goat anti-mouse IgG. No positive binding was observed. (E) Interphase sample from the same control subject, containing lymphocytes (L), monocytes (M), and PMN (P), incubated with a monoclonal antibody against VT-2. No binding of VT-2 was observed. (B) Results for the first sample of blood from a patient with HUS (patient 8), using the same incubation steps as used for the control sample. Of all PMN, 90% were positive for VT-2 binding. (D) Results showing that, 5 d later, VT-2 binding to PMN was reduced to 15%. (F) Results showing that, interestingly, monocytes were positive for VT-2 binding on day 5, whereas no binding was observed on day 1.

Figure 2B presents the results of VT binding to PMN from one patient (patient 8) on the first day after admission to the hospital; on the first day, 90% of the PMN were positive for VT-2 staining. Five days later, only 15% of PMN were still positive (Figure 2D), suggesting that VT was transferred to target cells (data not shown). Interestingly, at that time not only PMN but also monocytes (Figure 2E) were positive for VT binding, which is highly indicative of activated monocytes, as described by van Setten et al. (12). For patient 9, VT-2 binding to monocytes was also observed 5 d after collection of the first sample. No VT-2 binding to monocytes was observed for the other patients, and patient 10 exhibited no binding to PMN 5 d after collection of the first sample.

The time course of VT-2 binding to PMN was studied only for the five patients with the greatest numbers of VT-2-positive cells (patients 5, 8, 9, 10, and 11) (Figure 3A,3B). For patients 5, 8, 10, and 11, large decreases in the numbers of PMN positive for VT-2 were observed in the 5 d after collection of the first sample. The last sample for patient 9 was obtained 5 d after the first sample, and no significant decrease in VT-2 binding was observed. A possible explanation for this may be that additional toxin was absorbed from the circulation and transferred to the white blood cells as fast as the cells could transfer the toxin to Gb3-containing glomerular cells. Patient 9 had a very severe form of HUS, with neurologic involvement, and died on the day the last sample was obtained.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

(A) Flow cytometric analysis for detection of VT-2 bound to PMN from patients 9, 10, and 11 on days 1 and 5 after admission. In all determinations performed, blood from a healthy volunteer was used as a control sample (left panels), to exclude nonspecific binding of the monoclonal antibody against VT-2. For patient 9, 96% of PMN were positive on the first day and 92% of the cells were still positive after 5 d. The fluorescence intensity decreased, indicating that less VT-2 was present on PMN. However, VT-2 binding to monocytes was simultaneously observed for this patient.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

(B) Flow cytometric analysis of interphase samples from patients 9, 10, and 11 on days 1 (left) and 5 (right). For patient 10, 14% of the PMN (P) were positive for VT-2 binding on day 1. On day 5, no positive binding was observed (2% of the cells were nonspecifically labeled with FITC-conjugated goat anti-mouse Ig; this value was subtracted from the total numbers of positive cells, leading to the conclusion that no cells were positive for this patient). For patient 11, VT-2 binding to PMN was observed for 80% of the cells on day 1. On day 5, 20% of the cells were still positive. For patients 10 and 11, no binding to monocytes (M) or lymphocytes (L) was observed.

No positive staining for VT-2 was observed for patients 1 and 4. All patients who exhibited positive results for VT-2 binding to PMN were still experiencing sometimes bloody diarrhea at the time the blood samples were obtained. Patient 1 exhibited no prodromal phase of bloody diarrhea and experienced a HUS relapse just a few weeks later. He responded to treatment with plasmapheresis. Therefore, he was considered to have an atypical form of HUS.

Discussion

In this study, we demonstrated the presence of VT-2 in the systemic circulation of nine of 10 patients during the acute phase of the epidemic form of HUS. In 85% of all cases of D+ HUS in Western Europe, infections with VT-2-producing E. coli are found (13). Therefore, we investigated the presence of VT-2 in the systemic circulation of patients with D+ HUS, in the acute phase of the disease.

VT-2 bound to PMN and not to erythrocytes, monocytes, or lymphocytes during the period of diarrhea. There was a strong association between the detection of VT-2 bound to PMN and the simultaneous presence of bloody diarrhea. One patient with atypical HUS repeatedly exhibited no positive staining during the acute phase of the disease. In the other case in which no VT-2 binding was detected (patient 4), the result was probably negative because of late admission to the hospital. That patient had been without diarrhea for 7 d at the time the blood sample was obtained. An explanation for this observation may be that the toxin had already been transferred from the intestine to the kidney. Alternatively, this patient may have had an infection with VT-1-producing E. coli, which we could not detect with the VT-2-specific monoclonal antibody used in this study.

It has often been suggested that PMN may play a critical role in the pathogeneses of D+ HUS. PMN levels are elevated in HUS, and it has been suggested that the number of PMN is a predictive factor for the outcome of the disease (14,15). In addition, increased numbers of PMN were found in the glomeruli of kidney autopsy samples from patients with D+ HUS (16,17). Furthermore, PMN in patients with HUS are activated, and elevated levels of interleukin-8 and elastase are found (18). It has been suggested that PMN of patients with HUS may damage the endothelium through release of intracellular components, such as elastase, or formation of superoxide (19,20). In our in vitro experiments, we indeed demonstrated that PMN loaded with VT-1 were able to induce endothelial cell death, whereas VT-1 or PMN alone had no effect (11). In addition, a recently published study showed that apoptosis of PMN was inhibited by VT; those authors concluded that longer survival of neutrophils might aggravate neutrophil-mediated tissue damage (21). Finally, Zoja and colleagues (22) reported that VT-1 can cause increased adhesion of PMN to the endothelium under flow conditions, by up-regulating adhesive proteins. Administration of antibodies against E-selectin, intercellular adhesion molecule-1, or vascular cell adhesion molecule-1 reduced the adhesion of PMN to the endothelium.

Karmali and colleagues (23) investigated the presence of VT-1 in the serum of rabbits and demonstrated a short serum half-life (2 min) for VT-1. From those results, the authors concluded that VT is probably rapidly cleared from the systemic circulation and thus is not detectable in the circulation of patients with HUS. They did not consider a role for PMN in transporting VT, however. In our in vitro experiments, we demonstrated that VT binds rapidly and exclusively to PMN (11). No binding to other components of blood was observed. Interestingly, it has been suggested that PMN also play a very important role in the pathogenesis of Kawasaki disease, and it has been reported that lipopolysaccharide is bound and transported by PMN (24).

In line with our in vitro data are the findings that VT-2 bound to PMN in vivo in nine of 10 patients with the epidemic form of HUS. We think that VT has never been detected in the serum of patients with HUS because VT binds rapidly to PMN after entering the systemic circulation and thus is absent from the plasma or serum of patients with HUS.

The finding of VT-2 bound to PMN from patients with D+ HUS not only represents the missing link between intestinal infection and damage to target organs but also provides new approaches for therapy. Two patients with severe HUS whom we studied exhibited high percentages of PMN positive for VT-2 binding; in a later phase, binding to monocytes was also observed. We think that, during the phase of bloody diarrhea, VT traverses the intestine-blood barrier, binds to PMN, and thus is transferred to target organs.

Synsorb Pk, a synthetic analog of the Gb3 receptor, can bind VT in vitro and can neutralize VT when mixed in vitro with VT-positive stools from children with HUS (25,26). When Synsorb Pk is administered orally to patients with the epidemic form of HUS while they still have diarrhea, it reduces the amount of VT available; this reduction might prevent and/or decrease the binding of VT to PMN and the subsequent transfer of VT from the intestine to target organs. However, additional therapy is required in severe cases. We think that, on the basis of the finding that VT is bound to PMN in patients with D+ HUS, leukopheresis would be a possible solution in such cases. The presence of VT bound to PMN can be assessed rapidly, using a simple, reproducible, quick method, as we have described in this report. PMN loaded with VT can be removed using leukopheresis, thus preventing the transfer of VT to target organs or protecting target organs from severe damage.

In conclusion, our data demonstrate a new and crucial aspect in the pathogenesis of HUS, namely the specific binding and transfer of VT by PMN in the systemic circulation. We think that this observation is important for clarification of the pathogenesis of HUS and for treatment of patients with HUS. We suggest that in severe cases of D+ HUS, effective leukopheresis therapy should be considered.