Most cases of hemolytic uremic syndrome (HUS)1 occur as a complication of intestinal infections with Shiga toxin-producing Escherichia coli (STEC), a group of pathogenic E. coli that represents a major public health concern worldwide.2–5 The pathogenesis of STEC-associated HUS includes the colonization of the intestinal mucosa, the release of Shiga toxins (Stx) in the intestinal lumen, their passage into the blood, and their delivery to the renal and cerebral endothelia endowed with toxin receptors.6–9 The Stx-induced renal endothelial injury is the primary pathogenetic event in HUS.7 Although the macropinocytotic mechanism of toxin uptake by polarized intestinal epithelial cells and the transcytotic process that allows Stx to reach the circulation are well-known,10–12 the following 2 different hypotheses have been proposed to explain the mode of toxin delivery to the kidney: (i) soluble Stx circulate free in the blood stream; and (ii) Stx bind to some blood components (cells, macromolecules), transporting them to the target cells. Because free Stx have never been found in the serum of patients with HUS,3,13 some authors have proposed that Stx may have a very rapid blood clearance.14 Conversely, the 2 main Stx variants (Stx1 and Stx2) bind to human polymorphonuclear leukocytes (PMN) to the same extent, through a Gb3-independent mechanism as demonstrated in different studies in vitro.15–18 Moreover, Stx have been detected on the surface of circulating PMN in the blood of patients with HUS by cytofluorimetric assays with monoclonal antibodies to Stx.17,19,20
Independent from the toxin mode of delivery, the endothelial cells intoxicated by Stx show a wide spectrum of responses such as apoptotic cell death,21,22 and production of proinflammatory cytokines23,24 that might contribute to the pathogenesis of HUS by inducing inflammation in the kidney.7,8 Although STEC may produce 2 main types of AB5 exotoxins, the cytokine upregulation events induced by Stx2 are much more efficient than those triggered by Stx1.24,25 This finding has been proposed24 as the molecular explanation of the strong epidemiologic association between Stx2-producing E. coli strains and HUS.26,27
Our group has previously demonstrated that either free Stx25 or PMN-bound Stx28 are able to impair protein synthesis and, at the same time, trigger the production of proinflammatory cytokines in human endothelial cells. In this study, we investigate the relationship between the amounts of Stx detectable on the PMN or in the sera of HUS patients and their clinical features.
In Italy, cases of HUS are notified on a National Registry by centers for pediatric nephrology, and clinical specimens are submitted to a reference laboratory (Istituto Superiore di Sanità) for diagnosis of STEC infection.4,29 A total of 60 consecutive HUS cases (<15 year) admitted between 2003 and 2007 at 3 hospitals participating in the Registry were enrolled in the study after obtaining an informed consent from their parents, and were examined for laboratory evidence of STEC infection. A case of HUS was defined as a patient with acute microangiopathic hemolytic anemia, thrombocytopenia (platelet count <150,000/mm3), and acute renal injury.4,29 The definition includes both complete (acute renal failure) and incomplete (acute renal injury with preserved or only slightly impaired renal function) HUS.30 Feces and blood specimens were collected as soon as possible after admission to the hospital. Feces and sera were stored at −20°C. Ethylenediaminetetraacetic acid-treated blood samples for detection of Stx on PMN were stored at room temperature and were processed within 24 hours of sampling. Information on hemolytic anemia, platelet and leukocyte counts, serum creatinine concentrations, dialysis and/or transfusions, and neurologic complications were collected for all patients. Serum creatinine concentrations were defined pathologic whether the values were above the upper limit for age (<2 years: pathologic values >0.66 mg/dL; >2 years: pathologic values >0.77 mg/dL).
Laboratory Diagnosis of STEC Infection
For STEC isolation, feces were streaked onto MacConkey agar plates and colony sweeps were tested for Stx production by the Vero cells assay and for the presence of stx genes by polymerase chain reaction amplification.4 Stool specimens were also examined for the presence of free Stx by the Vero cell cytotoxicity assay, as described previously.13,31 Serum samples were tested for antibodies to the lipooligosaccharide (LPS) of 5 major STEC serogroups (O157, O26, O103, O111, and O145) by ELISA (enzyme-linked immunosorbent assay) and immunoblotting, as previously described.13,32 Sera from patients were also tested for the presence of free Stx by a protein synthesis inhibition assay33 conducted on the culture of human umbilical vein endothelial cells (HUVEC) as previously described.21,34 HUVECs were challenged for 16 hours with free Stx35,36 in the presence of 10% human sera, with 10% sera from Stx-treated blood from healthy donors, and with 10% sera from patients with HUS. The inhibition value of each single serum sample was obtained from 4 independent experiments.
Detection of Stx Bound to PMN
Stx bound to PMN were detected by flow cytometry, as reported in a specific methodological article in which all technical details were carefully described.17 The assay was validated by comparing control subjects and HUS patients in a blinded manner17 and by challenging Stx-positive PMN with a negative control antibody.19 Briefly, leukocytes isolated after erythrocyte lysis from 100 μL of blood were incubated with equal amounts of mouse monoclonal antibodies (IgG) against Stx1 and Stx2, in the presence of human serum to saturate Fc receptors on PMN. After incubation with FITC-goat antimouse IgG, flow cytometric analysis was used to reveal the PMN-bound fluorescence, allowing a highly sensitive single-step detection of both Stx. The mean channel value of fluorescence (MCV) was chosen as an objective parameter to measure the extent of binding of Stx to PMN.17 The single values were calculated by subtracting the control MCV, that is, the MCV of white cells from the same patient incubated with the secondary antibody alone. An MCV equal to 0.3 was considered the detection limit and MCV between 0.3 and 0.6 were considered borderline.
Personal, clinical, and laboratory data were stored in a Microsoft Excel file. Data analysis was performed using SPSS version 14.0. Differences in continuous variables were tested with Student t test after controlling for homoscedasticity. Differences in proportions were assessed through the χ2 of Fisher exact test when appropriate.
Study Population and Evidence of STEC Infection
Of the 60 HUS cases consecutively observed, 46 had evidence of STEC infection and were further investigated. Their median age was 23 months (range: 4–180) and the gender ratio (number of males per 100 females) was 156. All cases but 3 underwent the complete diagnostic panel for STEC infection, although fecal samples were not available for 3 patients.
The results of the different assays by presence of prodromal diarrhea are reported in Table 1. Free Stx were detected in the feces of 15 patients: 12 had Stx2, 1 had Stx1, and 2 had both toxins. STEC strains possessing stx2 genes and belonging to serogroups O157, O26, O114, and O121 were isolated from 7 of the patients with fecal Stx2. Of the 46 patients examined, LPS antibodies were detected in the sera of 30 patients. Antibodies to O26 were detected in 12 cases, to O157 in 7, to O103 in 6, to O111 in 3, and to O145 in 2 cases. Flow cytometric analysis showed the presence of Stx on the circulating PMN from 27 of the 46 HUS cases (58.7%). In 9 of these patients, the detection of Stx on PMN was the sole evidence of STEC infection. In contrast, PMN were negative in 19 patients with other evidence of STEC infection. Those included 3 positive for STEC, fecal Stx, and LPS antibodies; 4 positive for fecal Stx and LPS antibodies; and 12 positive for LPS antibodies only.
An attempt to detect free Stx in the blood was carried out for 34 of the 46 patients by using HUVEC cultivated 16 hours in the presence of 10% patients' sera. The very low inhibition of protein synthesis observed (3.9% ± 6.4%) was close to the detection limit. It should be noted that in the presence of 10% human serum, HUVEC are very sensitive to added Stx1 (IC50 = 0.9 pM; ∼20 pg/well) and Stx2 (IC50 = 2.5 pM; ∼50 pg/well). Moreover, the IC50 of both toxins changed neither after 2 days incubation with human sera at room temperature, nor after 3 cycles of freezing and thawing, excluding loss of activity due to the storage of patients' sera. Finally, the cytotoxic activity of Stx added to blood from healthy donors was completely recovered after the preparation of sera. We conclude that patients' sera contained only negligible amounts of free Stx, not sufficient to induce injury to endothelial cells and the associated responses.25
Amounts of Stx Bound to PMN
The MCV represents an objective parameter to quantify the amount of Stx bound to PMN by indirect flow cytometry.17,19 Therefore, the 27 patients with Stx-positive PMN were divided into 2 groups according to the MCV of their PMN. Patients in Group 1 (high Stx amount, 11 cases) had MCV values ≥1.5; patients in Group 2 (low Stx amount, 16 cases) had MCVs of <1.5 (P < 0.001) (Table 2), whereas the remaining patients (Group 3) had MCVs lower than the detection limit 0.3. The 1.5 cut-off point chosen to discriminate between Group 1 and Group 2 is the MCV corresponding to 50% mean saturation of PMN receptors,19 and in HUVEC experimentally challenged with PMN loaded with different amounts of Stx,28 it represents the breaking point between the endothelial production of high level of proinflammatory cytokines and the nearly complete block of protein synthesis and cytokine production.
The median time interval between the sampling of PMN and the onset of prodromal diarrhea or the diagnosis of HUS was similar for the 2 groups with positive PMN (Group 1: 8 days, range: 1–22; Group 2: 8.5 days, range: 1–20), whereas it was slightly lower for the group with negative PMN (Group 3: 6 days, range: 1–21). Nearly all the patients were infected by Stx2-producing E. coli and the minority of patients (3 cases) infected by strains producing Stx1 alone or in combination with Stx2 was equally distributed in the groups.
Clinical Features of Patients According to Stx Amounts on PMN
The clinical presentation of STEC-associated HUS may vary from mild cases without renal failure (incomplete form) to full-blown HUS needing dialysis and with severe clinical manifestations and complications.37 Therefore, the 3 groups of patients, defined according to the amounts of Stx on PMN, were compared for platelet and neutrophil counts, hemoglobin and serum creatinine concentrations, need for dialysis and/or transfusions, and neurologic complications. All patients showed thrombocytopenia, acute hemolytic anemia, and elevated neutrophil counts in the days preceding the PMN assay (Fig. 1, day 0). In the following days, platelet increased (Fig. 1A) and neutrophil decreased (Fig. 1B) without significant differences among the groups. Conversely, mean serum creatinine concentrations in Group 1 was lower than those measured in Groups 2 and 3 (Table 2), as confirmed by the trends of the relative curves until day 4 (Fig. 1C). Notably, serum creatinine concentrations were abnormal in only 5 patients in Group 1 (45%), whereas nearly all of the children in Group 2 (88%, P < 0.05 vs. Group 1) and all those in Group 3 (100%, P < 0.001 vs. Group 1) had values above the upper limit for age during the entire period of observation (days: 0–7). Accordingly, the proportion of patients needing peritoneal dialysis or hemodialysis was slightly higher in Groups 2 and 3, and the mean duration of the dialysis was longer (Table 2).
Neurologic complications including seizures and coma were more frequent in Group 1 and in Group 3 compared with Group 2 (Table 2).
Stratification of clinical data according to gender within each Group or in all of the patients did not reveal significant correlations between the gender of patients and the severity of HUS, even though females showed a more powerful trend in having neurologic complications.
The present investigation extends our previous studies on the presence of Stx on circulating PMN of children with HUS17,19 and confirms that its detection can be a useful integration into the panel of microbiologic and serologic assays currently used for the diagnosis of STEC infections. Conversely, only negligible amounts of Stx were observed in patients' sera in agreement with previously reported findings.3,13 This observation supports the view that PMN plays a prominent role in the delivery of Stx.
Some authors have claimed that the cytofluorimetric assay developed for Stx detection is nonspecific because of false-positive results after hemodialysis, possibly caused by activation of PMN.38 We did not observe any correlation between positive cytofluorimetric assays and concomitant hemodialysis, as only 27% of the patients in Group 1 and 19% of those in Group 2 underwent hemodialysis, which, conversely, was administered to 42% of patients with negative PMN. This excludes the presence of artifacts in Stx detection on PMN related to dialysis. The cytofluorimetric assay also provides a quantitative evaluation of the toxin amounts on PMN.17,19 Therefore, we investigated the clinical relevance of the concentrations of Stx on PMN and clinical features in HUS cases. When patients were divided into 3 groups according to the amounts of Stx detected on PMN, no significant differences were observed in platelet and neutrophil counts. However, most cases in Group 1 (high Stx amount) presented incomplete HUS, with preserved or slightly impaired renal function (normal serum creatinine), whereas most children in Group 2 (low Stx amount) and all patients in Group 3 (negative PMN) had acute renal failure (creatinine values above the upper limit for age). The trends of mean serum creatinine concentrations over time were consistent with these findings.
This apparent paradox could be explained by the following hypotheses: (i) circulating PMN are not involved in toxin delivery and might capture free toxins preventing their binding to renal endothelium (sponge effect); (ii) PMN are involved in toxin transport, and the amounts of Stx on their membrane could reflect the differential avidity for toxins of renal endothelia of different patients or, conversely; (iii) PMN shuttle the gut-associated toxins to the kidney and the endothelial responses to intoxication, such as proinflammatory cytokines release, might be inversely related to the amounts of toxins delivered to endothelia. The last explanation is validated by our previous experimental observations, which showed that PMN loaded with different amounts of Stx, comparable with those observed in our groups of HUS patients, induced strikingly different responses in endothelial cells in terms of inflammatory cytokines released.28 We found that the transmigration of PMN carrying low amounts of Stx, comparable with those found in patients in Group 2, caused an inhibition of protein synthesis lower than 60% in endothelial cells, accompanied by a strong upregulating effect on IL-8 and MCP-1 production. However, the transmigration of PMN with high MCV values, similar to those of patients in Group 1, caused a block of translation greater than 60%, with the concomitant impairment of IL-8 and MCP-1 induction. The production of proinflammatory cytokines is indeed almost completely blocked by the strong inhibition of translation caused by high amounts of Stx. Therefore, HUS patients with high Stx amounts may have developed sudden endothelial injuries, which triggered apoptosis in intoxicated cells, without producing the complete set of cytokines necessary to amplify the process and, finally, to induce the inflammatory damage of the kidney. We cannot exclude that the differences observed in the amount of Stx could have also been influenced by the phase of the infection in which PMN were collected. However, the median intervals between the onset of disease and PMN sampling were similar in the 3 groups of patients. Moreover, our findings are in keeping with those obtained by Taylor et al39 in baboons experimentally treated with intravenous infusion of high and low doses of Stx1. In this model, the histopathologic changes observed in the kidney were less pronounced in the animals challenged with high concentrations of toxin, and the authors concluded that the low-dose group accurately reflects what is usually observed in childhood postdiarrheal HUS.39 In the same study, the necropsy of animals treated with the high-dose toxin also revealed edema of the brains. It is noteworthy that, in our study, neurologic complications were observed more frequently in patients with high amounts of Stx in PMN (Table 2). We speculate that the neurologic impairment observed in the HUS patients belonging to Group 1 could have been derived by Stx-underlined pathogenetic mechanisms. Rather than a slowly self-amplifying proinflammatory circle as in the fenestrated renal endothelia, the sudden brain endothelial injuries imposed by large amounts of toxins might alter the blood-brain barrier, eventually inducing neurologic complications.
The lack of detection of Stx in the blood of patients in Group 3 is surprising, as the interval between the onset of disease and blood sampling was not different from those of the other 2 groups, as well as the presence of STEC or free Stx in the stool samples. We can hypothesize that these patients might have had a rapid blood clearance of Stx, preventing their detection.
In conclusion, our data suggest that the extent of renal damage in children with HUS could depend on the amount of Stx on their PMN and presumably delivered by them to the renal endothelium. Large amounts of Stx could induce a reduced release of cytokines by endothelia, with consequent lower degree of inflammation. However, low amounts of toxin can trigger the cytokine cascade, provoking inflammation that ultimately contributes to tissue damage. This hypothesis could be directly assessed in clinical studies aimed at investigating serum and urine cytokine levels in patients with different amounts of Stx on their PMN.
1. Trompeter RS, Schwartz R, Chantler C, et al. Haemolytic-uraemic syndrome: an analysis of prognostic features. Arch Dis Child
2. Griffin PM, Tauxe RV. The epidemiology of infections caused by Escherichia coli
O157:H7, other enterohemorrhagic Ecoli
, and the associated hemolytic uremic syndrome. Epidemiol Rev
3. Karmali MA, Petric M, Lim C, et al. The association between idiopathic hemolytic uremic syndrome and infection by verotoxin-producing Escherichia coli
. J Infect Dis
4. Tozzi AE, Caprioli A, Minelli F, et al. Shiga toxin-producing Escherichia coli
infections associated with hemolytic uremic syndrome, Italy, 1988–2000. Emerg Infect Dis
5. Caprioli A, Morabito S, Brugere H, et al. Enterohaemorrhagic Escherichia coli
: emerging issues on virulence and modes of transmission. Vet Res
6. Paton JC, Paton AW. Pathogenesis and diagnosis of Shiga toxin-producing Escherichia coli
infections. Clin Microbiol Rev
7. Ray PE, Liu XH. Pathogenesis of Shiga toxin-induced hemolytic uremic syndrome. Pediatr Nephrol
8. Noris M, Remuzzi G. Hemolytic uremic syndrome. J Am Soc Nephrol
9. Nataro JP, Kaper JB. Diarrheagenic Escherichia coli
. Clin Microbiol Rev
10. Acheson DW, Moore R, De Breucker S, et al. Translocation of Shiga toxin across polarized intestinal cells in tissue culture. Infect Immun
11. Hurley BP, Thorpe CM, Acheson DW. Shiga toxin translocation across intestinal epithelial cells is enhanced by neutrophil transmigration. Infect Immun
12. Malyukova I, Murray KF, Zhu C, et al. Macropinocytosis in Shiga toxin 1 uptake by human intestinal epithelial cells and transcellular transcytosis. Am J Physiol Gastrointest Liver Physiol
13. Caprioli A, Luzzi I, Rosmini F, et al. The HUS Italian Study Group. Hemolytic-uremic syndrome and Vero cytotoxin-producing Escherichia coli
infection in Italy. J Infect Dis
14. Flagler MJ, Strasser JE, Chalk CL, et al. Comparative analysis of the abilities of Shiga toxins 1 and 2 to bind to and influence neutrophil apoptosis. Infect Immun
15. Griener TP, Mulvey GL, Marcato P, et al. Differential binding of Shiga toxin 2 to human and murine neutrophils. J Med Microbiol
16. te Loo DM, Monnens LA, van Der Velden TJ, et al. Binding and transfer of verocytotoxin by polymorphonuclear leukocytes in hemolytic uremic syndrome. Blood
17. Tazzari PL, Ricci F, Carnicelli D, et al. Flow cytometry detection of Shiga toxins in the blood from children with hemolytic uremic syndrome. Cytometry B Clin Cytom
18. Brigotti M, Carnicelli D, Ravanelli E, et al. Interactions between Shiga toxins and human polymorphonuclear leukocytes. J Leukoc Biol
19. Brigotti M, Caprioli A, Tozzi AE, et al. Shiga toxins present in the gut and in the polymorphonuclear leukocytes circulating in the blood of children with hemolytic-uremic syndrome. J Clin Microbiol
20. Te Loo DM, van Hinsbergh VW, van den Heuvel LP, et al. Detection of verocytotoxin bound to circulating polymorphonuclear leukocytes of patients with hemolytic uremic syndrome. J Am Soc Nephrol
21. Brigotti M, Alfieri R, Sestili P, et al. Damage to nuclear DNA induced by Shiga toxin 1 and ricin in human endothelial cells. FASEB J
22. Nakao H, Takeda T. Escherichia coli
Shiga toxin. J Nat Toxins
23. Zoja C, Angioletti S, Donadelli R, et al. Shiga toxin-2 triggers endothelial leukocyte adhesion and transmigration via NF-kappaB dependent up-regulation of IL-8 and MCP-1. Kidney Int
24. Matussek A, Lauber J, Bergau A, et al. Molecular and functional analysis of Shiga toxin-induced response patterns in human vascular endothelial cells. Blood
25. Brigotti M, Carnicelli D, Ravanelli E, et al. Molecular damage and induction of proinflammatory cytokines in human endothelial cells exposed to Shiga toxin 1, Shiga toxin 2, and alpha-sarcin. Infect Immun
26. Persson S, Olsen KE, Ethelberg S, et al. Subtyping method for Escherichia coli
Shiga toxin (verocytotoxin) 2 variants and correlations to clinical manifestations. J Clin Microbiol
27. Friedrich AW, Bielaszewska M, Zhang WL, et al. Escherichia coli
harboring Shiga toxin 2 gene variants: frequency and association with clinical symptoms. J Infect Dis
28. Brigotti M, Tazzari PL, Ravanelli E, et al. Endothelial damage induced by Shiga toxins delivered by neutrophils during transmigration. J Leukoc Biol
29. Gianviti A, Tozzi AE, De Petris L, et al. Risk factors for poor renal prognosis in children with hemolytic uremic syndrome. Pediatr Nephrol
30. Tarr PI, Gordon CA, Chandler WL. Shiga-toxin-producing Escherichia coli
and haemolytic uraemic syndrome. Lancet
31. Caprioli A, Luzzi I, Gianviti A, et al. Pheno-genotyping of verotoxin 2 (VT2)-producing Escherichia coli
causing haemorrhagic colitis and haemolytic uraemic syndrome by direct analysis of patients' stools. J Med Microbiol
32. Caprioli A, Luzzi I, Rosmini F, et al. Community-wide outbreak of hemolytic-uremic syndrome associated with non-O157 verocytotoxin-producing Escherichia coli
. J Infect Dis
33. Petronini PG, Tramacere M, Mazzini A, et al. Hyperosmolarity-induced stress proteins in chick embryo fibroblasts. Exp Cell Res
34. Maier JA, Voulalas P, Roeder D, et al. Extension of the life-span of human endothelial cells by an interleukin-1 alpha antisense oligomer. Science
35. Ryd M, Alfredsson H, Blomberg L, et al. Purification of Shiga toxin by alpha-D-galactose-(1–4)-beta-D-galactose-(1–4)-beta-D-glucose-(1-) receptor ligand-based chromatography. FEBS Lett
36. Downes FP, Barrett TJ, Green JH, et al. Affinity purification and characterization of Shiga-like toxin II and production of toxin-specific monoclonal antibodies. Infect Immun
37. Gerber A, Karch H, Allerberger F, et al. Clinical course and the role of Shiga toxin-producing Escherichia coli
infection in the hemolytic-uremic syndrome in pediatric patients, 1997–2000, in Germany and Austria: a prospective study. J Infect Dis
38. Geelen JM, van der Velden TJ, Te Loo DM, et al. Lack of specific binding of Shiga-like toxin (verocytotoxin) and non-specific interaction of Shiga-like toxin 2 antibody with human polymorphonuclear leucocytes. Nephrol Dial Transplant
39. Taylor FB Jr, Tesh VL, DeBault L, et al. Characterization of the baboon responses to Shiga-like toxin: descriptive study of a new primate model of toxic responses to Stx-1. Am J Pathol