Acute inflammation in the pathogenesis of hemolytic-uremic syndrome

Kidney International, Vol. 61 (2002), pp. 1553–1564

NEPHROLOGY FORUM

Acute inflammation in the pathogenesis of hemolytic-uremic syndrome Principal discussant: Andrew J. King Scripps Clinic, La Jolla, California, USA

CASE PRESENTATION The patient, a 68-year-old white woman, presented to her primary physician with two to three days of bloody diarrhea; bilateral, crampy, lower abdominal pain, and fever to 101⬚F. Initially, she was treated with clear liquids as well as ciprofloxacin and loperamide. Over the next 24 hours, diarrhea, nausea, and vomiting increased, causing her to go to the emergency ward. Her family had noted progressive confusion over the previous 24 hours. There was no history of recent travel, nor was there any known contact with infected individuals. Her dietary history was significant for the ingestion of hamburger meat three days prior to the onset of symptoms. She had had type II diabetes mellitus for eight years, hypertension for 25 years, and hypothyroidism. Her medications on admission were hydrochlorothiazide, 25 mg/day; metoprolol, 25 mg twice daily; and thyroxine in addition to the ciprofloxacin and loperamide. There was no family history of renal or hematologic disease. A retired secretary who lived alone, she was too confused to give a review of systems, although her family stated that prior to this acute illness she had been doing well. Physical examination in the emergency ward revealed a markedly confused woman unable to state her name or the date. Her

The Nephrology Forum is funded in part by grants from Amgen, Incorporated; Merck & Co., Incorporated; Dialysis Clinic, Incorporated; and Bristol-Myers Squibb Company. Key words: Shiga toxin, enterohemorrhagic E. coli, von Willebrand factor-cleaving protease, hemorrhagic enterocolitis, thrombotic microangiopathy.

 2002 by the International Society of Nephrology

temperature was 38.6⬚C, her blood pressure was 140/80 mm Hg, and her heart rate was 86 beats/min. The head and neck examination were significant only for mildly icteric sclera. The lungs were clear to auscultation and percussion. The jugular veins were not distended, and no murmurs, rubs, or gallops were heard. Initial abdominal examination was notable for hypoactive bowel sounds, although the abdomen was soft and non-tender to palpation. Rectal examination revealed brown stool that was 4⫹ positive for occult blood and later found to be positive for fecal leukocytes. The neurologic examination was limited due to the confused state of the patient, but it did not reveal any focal defects. Laboratory testing revealed: serum sodium, 135 mEq/L; potassium, 3.4 mEq/L; chloride, 102 mEq/L; and bicarbonate, 19 mmol/L. The serum creatinine was 3.7 mg/dL; blood urea nitrogen, 110 mg/dL; and total bilirubin, 4.2 mg/dL with normal hepatic transaminases. Hematocrit was 25% with a hemoglobin of 8.1 g/dL. The leukocyte count was 29,700/mm3 with 86% polymorphonuclear cells, 3% bands, 8% lymphocytes, and 3% mononuclear cells. The platelet count was 49,000/mm3. Urinalysis revealed a specific gravity of 1.011; pH 5; 2⫹ protein, 2⫹ blood, and 5 to 7 red blood cells/high-power field in the sediment. No cellular casts were present. A chest radiograph was unremarkable, and a KUB radiograph revealed an ileus without evidence of free air. Over the following 12 hours, the patient’s abdominal examination changed, with the development of overt peritoneal signs leading to an exploratory laparotomy. Extensive edema of the distal ileum and colon with evidence of infarction was found. A partial colectomy, partial ileal resection, and ileostomy were performed. Pathology of the colon revealed extensive bowel wall edema with regions of infarction, sloughing of mucosa, and local microangiopathy. The bowel had marked infiltration of polymorphonuclear cells with crypt abscesses. Her postoperative course was complicated by oliguria and rapidly worsening renal function. Hemodialysis was initiated. She was treated with broad-spectrum antibiotics and underwent a total of six plasma exchange treatments using fresh frozen plasma as the replacement fluid. A stool culture from her initial visit with her primary physician grew Escherichia coli O157:H7. Her hospital course was prolonged (4 months total) and complicated by adult respiratory distress syndrome (ARDS), pancreatitis, seizures, Clostridium difficile colitis, and Staphylococcus aureus bacteremia. She never regained renal function, but her mental status returned to normal.

1553

1554

Nephrology Forum: Inflammation and HUS

DISCUSSION

Shiga toxin-producing Escherichia coli

Dr. Andrew J. King (Division Head, Division of Nephrology, Scripps Clinic and Green Hospital, La Jolla, California, USA): This unfortunate woman provides a dramatic illustration of the devastating nature of adult hemolytic-uremic syndrome (HUS) induced by enterohemorrhagic Escherichia coli (EHEC). She had multiorgan involvement, including the gut, central nervous system, and kidney, a presentation not too divergent from the patient with thrombotic microangiopathy first described by Eli Moschowitz in 1924 [1]. Thrombotic microangiopathy (TMA) is a term applied to a group of disorders that present with thrombocytopenia, evidence of red blood cell destruction, and a characteristic pathologic appearance in the microvasculature, including endothelial cell injury and platelet deposition. Thrombotic microangiopathy is associated with a wide range of infections, drugs, tumors, and other disorders. The focus of this report, however, is HUS triggered by EHEC. First described by Gasser et al in 1955 [2], HUS is a term applied to a TMA with predominant renal involvement. Although there had been speculation that HUS and thrombotic thrombocytopenic purpura (TTP) were differing manifestations of the same disease, it recently has become clear that for the majority of adults with TTP, the key pathologic feature is an immunoglobulin that inactivates circulating von Willebrand factor (vWF)cleaving protease [3, 4]. Failure to properly metabolize naturally occurring ultralarge multimers of vWF leads to increased binding to platelet glycoproteins under conditions of high shear stress, that is, that which exists in the microvasculature. These features of TTP explain the remarkable efficacy of plasma exchange in this disorder [5]. By contrast, vWF-cleaving protease activity is normal in patients with HUS, and the efficacy of plasma exchange is much less clear. Unlike TTP, patients with diarrhea-induced HUS have abundant evidence of an acute inflammatory response, the magnitude of which predicts clinical outcome. In this Forum, I will review evidence suggesting that inflammatory cells and their byproducts play a major role in (1) loss of intestinal barrier function, thereby promoting movement of endotoxin and Shiga toxins into the circulation; (2) delivery of Shiga toxins to target organs; (3) sensitization of target organs to Shiga toxins by increasing glycolipid receptor expression; and (4) direct injury of target organ endothelium. Although definitive evidence for a role of inflammatory cells and their mediators is lacking in humans, on circumstantial evidence, the case is compelling. Patients with HUS due to Shiga toxinproducing Escherichia coli (STEC) rarely present in the first 24 to 48 hours and thus, as with other forms of acute renal failure (ARF), we are left looking for the “smoking gun,” the mediators that cause the ARF.

Shiga toxin (Stx)-producing Escherichia coli has been designated by the U.S. Centers for Disease Control and Prevention as an “emerging pathogen” [6]. As many as 90% of children with HUS have antecedent STEC, typically of the serotype O157:H7 [7]. The association of Shiga toxin-producing Shigella dysenteriae with HUS was first made by Koster et al in 1978 [8], followed by observations indicating an association of E. coli cytotoxins with the syndrome by Karmali et al in 1985 [9]. The implicated E. coli cytotoxins, Stx1 and Stx2, belong to a family of bacterial protein toxins related to the parent Shiga toxin, which share the same enzymatic action and recognize related host-neutral glycolipid receptors. Interest in this disease has been fostered by considerable media attention stemming from several notable outbreaks, including one in 1993 related to undercooked hamburgers at “Jack in the Box” restaurants in the northwest United States and one more recently in Japan. Depending on the outbreak, patients with documented STEC have a risk of developing HUS of approximately 5% to 10% [7]. Details regarding the genetics, regulation, structure, enzymatic specificity, and physiology of the Stx family are found in several reviews [10–12]. Briefly, Stx is a heterodimer consisting of an enzymatically active A subunit surrounded by five identical B subunits. The receptor for the toxin is a blood group active glycolipid, globotriaosylceramide (Gb3). The distribution and fatty acid chain composition of Gb3 are believed to play a role in susceptibility of various cells to the toxin [13]. Upon binding, internalization and activation of the A subunit lead to depurination of a specific adenosine in 28S ribosomal RNA; the end result is irreversible inhibition of protein elongation [14]. Glomerular endothelial cells and, to a greater extent, renal tubular epithelial cells, express Gb3, thereby contributing to the propensity of patients infected with STEC to develop acute renal failure. Clinical markers of acute inflammation The woman presented today had high fever and marked leukocytosis, common findings in patients presenting with STEC. Leukocytosis can be extreme, as in leukemoid reactions, and it is a positive predictor of acute mortality and residual nephropathy [15–18]. Children with “diarrhea-positive” HUS (D⫹HUS) have significantly higher leukocyte counts on presentation than do children with atypical “diarrhea-negative” HUS (D-HUS); this finding suggests that the intestinal disease is an important factor in generating the leukocytosis [15]. In the Osaka outbreak of 1996, both leukocyte count and C-reactive protein were higher in the group of children with STEC who developed HUS compared to infected children who did not [19]. Severe gastrointestinal disease also portends a poor prognosis in children with STEC.

Nephrology Forum: Inflammation and HUS

1555

Fig. 1. Potential role of resident and circulating inflammatory cells in Shiga toxin-producing Escherichia coli (STEC)-induced enterocolitis. Abbreviations are: Stx, Shiga toxin; LPS, lipopolysaccharide; M⌽, macrophages; ROI, reactive oxygen intermediates.

Role of inflammatory cells in hemorrhagic enterocolitis The gastrointestinal disease in patients with STEC varies from watery diarrhea to the most severe form, hemorrhagic colitis. Colonic inflammation might play a role in local intestinal microangiopathy, the transfer of Stx from the bowel lumen through the lamina propria and into the circulation, and the generation of a systemic inflammatory response (Fig. 1). Although STEC was not previously thought to provoke marked intestinal inflammation, Slutsker et al noted that patients with STEC frequently have fecal leukocytes [20]. Indeed, colonic biopsy specimens from patients with STEC often have crypt abscesses and infiltrates of monocytes and polymorphonuclear cells (PMNs) [21–23]. In rabbits, infection with E. coli O157:H7 leads to severe inflammatory colitis with PMN infiltrates [24]. In this model, infusion of monoclonal antibodies directed toward the PMN ad-

hesion molecule CD-18 prevented intestinal infiltration of PMN and rendered histologic and functional protection [25]. Shiga toxin traverses polarized intestinal human epithelial cells grown in tissue culture and retains biologic activity [26]. Recent evidence suggests that Stx induces a chemokine response from human intestinal epithelial cells [27]. Using a human colonic epithelial cell line, Stx treatment led to “superinduction” of interleukin-8 (IL-8) with increase in both IL-8 mRNA and protein, the latter occurring despite the known inhibitory effects of Stx on protein elongation [27]. Secretion of IL-8 into the lamina propria might create a chemokine gradient sufficient to recruit circulating PMNs. Both IL-8 and tumor necrosis factor-␣ (TNF-␣) are more elevated in stools of patients with Stx-producing S. dysenteriae compared to patients infected with non-Stx-producing S. flexneri [28]. Several investigators have found that patients with STEC-induced

1556

Nephrology Forum: Inflammation and HUS Table 1. Blood and urinary cytokines in HUS

Cytokine IL-8

Levels

TNF-␣

Blood Urine Blood

IL-1␤

Urine Blood

IL-6

Urine Blood

IL-10

Urine Blood Urine

Increased Increased Increased Unchanged or infrequent No or scant data Increased Unchanged or infrequent No or scant data Increased Unchanged or infrequent Increased Increased Decreased No or scant data

References [18, 29-33, 35] [33] [36] [18, 34, 35, 37] [35] [37] [30-32, 34, 37, 38] [35] [34] [30-32] [29]

HUS have elevated circulating levels and urinary excretion of IL-8 (Table 1). The magnitude of IL-8 elevation correlates with markers of PMN degranulation and with patient outcome [18]. Other potential intestinal sources of chemokine production are the resident and influxing leukocytes in the lamina propria. Murine macrophages exposed to Stx increase production of both TNF-␣ and IL-6 [39]. Van Setten et al have shown in human monocytes that Stx increases production of IL-1␤, TNF-␣, IL6, and IL-8 and does not inhibit protein synthesis [40]. Binding sites for Stx on the monocytes were increased with lipopolysaccharide exposure. By contrast, Ramegowda and Tesh found that human monocytes and monocytic cell lines were quite resistant to Stx unless the cells were pretreated with phorbol esters [41]. Several potential consequences of a local inflammatory response in the lamina propria exist. Transepithelial migration of PMNs can lead to transient loss of colonic barrier function and cause the passage of luminal contents into the circulation [42, 43]. In one recent study, movement of PMNs from the basolateral to the apical side of a human intestinal epithelial cell line allows increasing translocation of both Stx1 and Stx2 in the opposite direction [44]. Loss of barrier function could contribute to direct entry of Stx into the circulation, as well as endotoxemia that would promote a systemic cytokine response. To date, no one has been able to detect circulating Stx in patients with STEC. Recent studies, however, have revealed Stx bound to circulating PMNs in patients with D⫹HUS, thus confirming entry into the circulation [45]. Most cellular responses in vitro, however, are elicited at Stx doses well below the lower limits of detection in plasma. Sepsis syndrome is very uncommon in children with STEC, except in the setting of bowel infarction. In this regard, several investigators have found specific antibodies to the lipopolysaccharide of Stx-producing E. coli [46–48]. Endotoxemia is much more prominent in patients with S. dysenteriae, an infection that is more likely to cause HUS and which is associ-

Fig. 2. Indirect immunofluorescence assay of Stx-2 bound to polymorphonuclear cells (PMNs) in a patient presenting with hemolytic-uremic syndrome (HUS). Reprinted with permission from the American Society of Nephrology [45].

ated with high circulating levels of IL-1 and TNF-␣. Increases in circulating inflammatory cytokines likely are triggered by the colonic infection and possibly by the low-grade endotoxemia (Table 1). When assessing the degree of cytokine elevation, note that renal failure independently contributes to high levels [49]. Futhermore, patients with uncomplicated STEC also have increased levels. Despite these caveats, it appears that patients with HUS have elevated circulating inflammatory cytokines. I will discuss the role of these cytokines as possible “coconspirators” in vascular injury in a few moments. However, local production of cytokines in the gut wall might be an important second step in the generation of intestinal inflammation because of their ability to activate intestinal endothelial cells, thereby increasing adhesion molecule expression and PMN diapedesis. Direct effects of Shiga toxins on inflammatory cells Increasing evidence, both in vitro and in vivo, suggests that Stx binds to PMN, triggers PMN activation, and promotes adherence of PMNs to target endothelium. Te Loo et al recently found that Stx binds to PMNs (not via Gb3) and raised the possibility that these cells are direct carriers of the toxin from the gut to target organs [50]. Further, these investigators report that circulating PMNs and, to a lesser extent monocytes from patients with D⫹HUS, have Stx bound to the surface (Fig. 2) [45]. As I said before, these data demonstrated unequivocally that Stx penetrates the gut in STEC-induced HUS and suggest that PMNs transport Stx to target cells. Activated PMNs contribute to tissue injury by releasing proteolytic enzymes, reactive oxygen intermediates, and phospholipid derivatives. Fitzpatrick et al detected in children with D⫹HUS sustained increases in ␣1-antitrypsin-complexed elastase, a PMN degranulation product that peaked approximately 3 to 4 days into admission [18]. Interestingly, this rise was preceded by an increase in

Nephrology Forum: Inflammation and HUS

1557

Fig. 3. Renal injury in patients with STEC proceeding from normal (top) to overt HUS (bottom). RTE, renal tubular epithelium; RBC, red blood cell; TNF, tumor necrosis factor; IL-1, interleukin-1; Gb3, globotriaosylceramide; GEC, glomerular endothelial cell; GepC, glomerular epithelial cell; PMN, polymorphonuclear cell; mes cell, mesangial cell.

IL-8, a chemokine that also activates PMNs. Others have confirmed the increase in elastase but were unable to detect activation of peripheral blood PMNs using chemiluminescence to assess for oxidative burst [51]. Evidence exists that children with HUS are under oxidative stress. Malonyldialdehyde blood levels are increased and there is oxidative damage to erythrocytes [52–55]. Patients with acute renal failure from other causes, however, also are under oxidative stress, and the clinical literature on HUS in general rarely includes appropriate controls. We have shown that Stx induces release of superoxide from PMNs in vitro in a dosedependent fashion and suggested that the toxin directly triggers oxidative burst in PMNs (Fig. 3) [56]. These cells were subsequently less able to respond to phorbol esters

or to undergo phagocytosis. Noris et al noted an approximate twofold increase in superoxide release of PMNs isolated from adults with recurrent TMA compared to controls [54]. These investigators also noted higher plasma nitrite/nitrate levels, a rise indicative of increased production of another reactive oxygen species, nitric oxide (NO). The source of NO remained speculative, although it was observed that the serum of patients with TMA induced higher levels of NO production from reporter human umbilical vein endothelial cells (HUVEC) compared to control serum [54]. The question remains open as to whether toxin-stimulated PMNs actually contribute to tissue injury, either in the gut or in remote target organs. The first evidence supporting this possibility came from Forsyth et al, who

1558

Nephrology Forum: Inflammation and HUS

noted that PMNs isolated from children with diarrheainduced HUS are more adherent to reporter HUVEC than are control PMNs [57]. Further, these PMNs induced morphologic evidence of endothelial cell injury in the form of fibronectin breakdown, an effect that was blocked in a subset of the patients by a CD-18 antibody. Our laboratory has examined this issue in vitro using a co-culture model, whereby normal PMNs are stimulated by Stx and placed in co-culture with reporter human saphenous vein endothelial cells (SVEC) (abstract, Murray SL et al, J Am Soc Nephrol 10:533A, 1999). We demonstrated that Stx-stimulated PMNs induced apoptosis in the co-cultured SVEC, whereas untreated PMNs or Stx alone did not. Other investigators have confirmed human microvascular endothelial cell cytotoxicity of PMNs from patients with TMA, noting significant protection by a metalloprotease inhibitor (BB3103), and several reactive oxygen species inhibitors (abstract, Ruiz-Torres et al, J Am Soc Nephrol 11:506A, 2000). Taken together, these results suggest that Stx-stimulated PMNs in the lamina propria (resident and recruited) contribute to endothelial cell damage in the intestine and lead to hemorrhage, loss of barrier function, and local microangiopathy. Thorpe et al also have shown that human intestinal microvascular endothelial cells express high levels of Gb3 and are exquisitely sensitive to the ribotoxic effects of Stx1 and 2 [27]. Whether Stx-activated PMNs act on target organs is an intriguing question that I will discuss later. The findings of Forsyth et al that PMNs from patients with TMA are more adherent to endothelial cells raises the possibility that Stx might alter adhesion molecular expression [57]. Indeed, Morigi and colleagues found that Stx1 pretreatment of HUVEC increased the binding of leukocytes flowing across the endothelial monolayer in a dose-dependent fashion [58]. The magnitude of effect was comparable to IL-1␤ and was inhibited by monoclonal antibodies to various adhesion molecules (E-selectin, ICAM-1, and VCAM). Further, TNF-␣ together with Stx led to greater binding than with either TNF-␣ or Stx alone. In preliminary studies, we have observed that flowing PMNs treated with Stx bind better to untreated HUVEC than do vehicle-treated PMNs, an effect that is magnified by pretreatment of HUVEC with IL-1 (personal observations). In the acute phase of D⫹HUS, soluble VCAM levels are increased [59, 60]. Once again, the effects of acute renal failure per se might play a role, as VCAM levels in the HUS patients were lower than in dialyzed controls without HUS [59]. Our ability to detect endothelial cell activation in vivo is limited. Nevertheless, these in vitro studies do support the possibility that increased levels of pro-inflammatory cytokines have an accelerant effect on the binding of toxin-laden PMNs to target endothelium.

Potential role of inflammation on target organ injury Whereas inflammatory cells are abundant in the gut of patients with STEC-induced HUS, their presence is far more subtle in the primary target organs, the central nervous system, and the kidney. In a case control study of 15 autopsy specimens from children who died during the acute phase of D⫹HUS, however, the median number of PMNs per 100 glomeruli was 929 compared to 156 for controls (children who died with normal renal function) [61]. The median peripheral neutrophil count in this autopsy series was 21 ⫻ 109/L. Although animal models have been problematic in the study of this disease, a naturally occurring disease in the greyhound dog, “Greentrack disease” (a.k.a. “Alabama rot”), bears a striking resemblance to the human disease; in this model, PMNs are present in the glomerulus [62]. In a recently described rat model of HUS induced by ricin, a toxin with similar ribotoxic effects to Stx, all the features seen in human HUS exist, including glomerular thrombosis and injury [63]. The glomeruli of these rats had a substantial increase in glomerular macrophages within eight hours of ricin infusion and increased circulating monocyte chemotactic protein-1, TNF-␣, IL-1, and IL-6. The recent finding by te Loo et al increases the likelihood that PMNs are an important carrier of Stx to the target endothelium [45]. The number of Stx-positive cells dropped off dramatically after 3 to 5 days in the patients with sequential measurements. Although the fate of those Stx-laden PMNs is unclear, it was shown convincingly that transfer of toxin occurs from the as-yet-undefined PMN binding site to Gb3 on glomerular endothelial cells in vitro [50]. By definition, the kidney is the primary target organ in STEC-induced HUS. This vulnerability is in large part due to the high density of Gb3 receptors in the renal microvasculature and renal tubular epithelial cells [64]. Indeed, renal tubular epithelial cells are perhaps the richest cellular source of Gb3 yet identified, equaling or perhaps even exceeding that of Vero cells. The open question is how much this vulnerability is due to ambient cytokine effects (Table 2). The laboratory of Hughes et al [67] and our laboratory have shown that these cells are exquisitely sensitive to the ribotoxic effects of Stx; this work raises the possibility that acute tubular necrosis plays a role in the renal failure (abstract, King et al, J Am Soc Nephrol 9:580A, 1998). Human glomerular microvascular endothelial cells (GMEC) express lesser amounts of Gb3, and in our studies and those of van Setten et al [72], GMEC were less sensitive to the ribotoxic effects of Stx than were other microvascular endothelial cells, such as human intestinal or foreskin endothelial cells. The variation in the literature with respect to human GMEC sensitivity is likely due to differences in source of cells, method of isolation, culture conditions,

1559

Nephrology Forum: Inflammation and HUS Table 2. Effects of cytokines on human kidney cell sensitivity to Stx Cell

Cytokine

Gb3 density/binding

Ribotoxicity

Apoptosis

Reference

TNF-␣ TNF-␣ TNF-␣ LPS

Increased density ND No change No change

Increased ND No change No change

ND Induceda ND ND

[40] [65] [64]

IL-1 TNF-␣ IL-6 LPS

Increase Increase No change Increase

Increase Increase No change Increase

ND ND ND ND

[66]

IL-1 TNF-␣

ND ND

Increase Increase

ND ND

[68]

IL-1 TNF-␣ IL-6 LPS TNF-␣ TNF-␣ TNF-␣

Increased binding No change No change No change ND ND ND

Increase No change No change Increase ND ND ND

ND ND ND ND Increased Increased Increased

[67]

Glomerular endothelium

Glomerular epithelium

Mesangium Renal tubular epithelium Proximal tubules

Proximal tubules Cortical cells ACHN cell line

[69] [70] [71]

ND, not done. a Apoptosis required pretreatment with TNF-␣

and inter-human variability (perhaps related to Gb3 fatty acid chain variance). Pretreatment of GMEC with TNF-␣ substantially increased both receptor density and sensitivity to Stx [72]. In keeping with these findings, recent data suggest that Stx induces apoptosis of GMEC only if cells are pretreated with TNF-␣ [65]. Taken together, these data suggest that the magnitude of cytokine response determines the expression of the glomerular disease. However, investigators who have carefully measured levels of inflammatory cytokines have shown an inconsistent relationship with disease severity, particularly when appropriate controls are included, such as patients with uncomplicated STEC or other causes of renal failure (Table 1). Other possible sources of inflammatory cytokines include activated monocytes adherent to glomerular cells or kidney cells. In support of the latter possibility, Stx infusion into transgenic mice bearing a chloramphenacol acetyl transferase (CAT) gene indicative of TNF-␣ synthesis revealed increased CAT activity exclusively in the kidney [73]. Furthermore, renal levels of TNF-␣, IL-1, and IL-6 are increased in gnotobiotic mice inoculated with E. coli O157:H7 [74]. Hughes and colleagues make a strong case for renal tubular epithelial cells as a possible source of local cytokine production [67]. They demonstrated potent induction and release of TNF-␣ and IL-1 from non-transformed human proximal tubule cells exposed to Stx, effects that were magnified by lipopolysaccharide. Taken together, these data suggest that inflammatory cytokines contribute to glomerular injury by increasing Gb3 expression on GMEC, thereby heightening the sensitivity of these cells to the ribotoxic effects of Stx and their

vulnerability to undergo apoptosis. If the finding that Stx increases adhesion molecule expression holds true for human GMEC, this mechanism could serve as an accelerant to glomerular injury by increasing Stx delivery. Conclusion Shiga toxin-producing Escherichia coli is an “emerging pathogen” that causes a devastating systemic disease, hemolytic-uremic syndrome. The gastrointestinal disease is associated with fecal leukocytes and a robust inflammatory response in the intestinal lamina propria. Shiga toxin itself likely triggers chemokine production by intestinal epithelial cells and/or resident macrophages, which leads to the influx of circulating inflammatory cells into the lamina propria. Once established, the local inflammation could promote injury both to the intestinal epithelium and endothelium and lead to intestinal hemorrhage, loss of barrier function, and generation of inflammatory cytokines. Loss of barrier function allows Stx to penetrate into the circulation, where it circulates freely or bound to PMN and is delivered to target organs. Injury of target cells is promoted by increased ambient inflammatory cytokines, possibly as a result of increased receptor density, and it culminates in organ dysfunction. The excessive vulnerability of the glomerulus likely relates to direct contact with toxin-loaded PMNs and is potentially increased by high local concentrations of proinflammatory cytokines generated by renal tubular epithelial cells. The fact that toxin-loaded PMNs are present when the patient is admitted to the hospital raises the possibility that leukopheresis might be a therapeutic option that will reduce target organ injury.

1560

Nephrology Forum: Inflammation and HUS

QUESTIONS AND ANSWERS Dr. Nicolaos E. Madias (Executive Academic Dean, Tufts University School of Medicine, Boston, Massachusetts): You mentioned that approximately 5% to 10% of children with documented Shiga toxin producing E. coli develop HUS. In view of the importance of pro-inflammatory cytokines in the pathogenesis of the syndrome, have polymorphisms in cytokine genes been examined regarding susceptibility to HUS? Dr. King: That is an interesting question, given the potential impact of these polymorphisms on the acute inflammatory response. However, there are no data in patients with HUS. This raises the larger question as to why only a minority of patients infected with STEC get HUS. Lingwood et al have raised the possibility that differences among individuals in the fatty acid composition of the glycolipid receptor alter binding characteristics of Stx and thus susceptibility to its toxic effects [75]. Support for this hypothesis remains limited. I think it is more compelling that the children with the most robust evidence of systemic inflammation appear to be those at greater risk. I suspect those are the children who have greater movement of toxin into the circulation (a higher dosage of Stx) and higher concentrations of pro-inflammatory cytokines making target organs more vulnerable. Dr. Madias: What do we know about the mechanism of binding of Shiga toxin to polymorphonuclear cells? Dr. King: These recent data are based on two reports [45, 50]. The nature of the receptor is unclear; however, it is not Gb3. Dr. John T. Harrington (Dean, Tufts University School of Medicine, Boston): You told us that 70% of the cases of HUS are related to O157:H7. Do we know whether the residual 30% of E. coli share the same characteristics? Dr. King: A large number of serotypes of E. coli are capable of producing Stx1 and Stx2. To my knowledge, the serotype is not a major factor in determining the relative risk of acquiring HUS. However, it appears that the majority of patients with HUS have STEC that produce Stx2. Thus, this subtype probably is more likely to induce systemic disease. Dr. Harrington: Andy, you said that “cleaving protease activity” was normal in HUS but absent in TTP, yet we see the same pathologic findings in both diseases. How do you account for that discrepancy? Dr. King: I think that thrombotic microangiopathies are diseases associated with endothelial cell injury. Persons with rare forms of recurrent TTP are not in a state of unrelenting thrombotic microangiography. Thus there must be an additional provocateur triggering the response. The deficiency in cleaving protease activity in these patients makes them susceptible to the full expression of microvascular disease. I managed a woman with

recurrent HUS who would become anuric whenever she acquired an infection with a fever greater than 101⬚F. Whatever the insult is to the endothelium, three manifestations of this injury are the loss of the non-thrombogenic properties of the monolayer, exposure of the basement membrane, and platelet aggregation, all of which lead to a characteristic pathologic finding. Dr. Andrew S. Levey (Chief, Division of Nephrology, New England Medical Center, Boston): You alluded to a therapy that could “clean up” the toxins from the white cells. Do you have any speculations on ways that that might be done? Dr. King: Leukopheresis would directly remove the toxin-coated white cells. Potentially, you could create an affinity column using Gb3 or another molecule with a Gal-Gal disaccharide. In most forms of acute renal failure in outpatients, we first see the patient two or three days into the disease, when many cellular responses might have occurred. Perhaps in the first 24 hours of STEC infection there is a robust influx of polymorphonuclear cells in the glomerulus. What I find intriguing about the report of te Loo et al was that even days after presentation there was evidence of toxin-bearing polymorphonuclear cells in the blood [45]. This finding would imply ongoing endothelial cell injury early in the hospitalization and a potential therapeutic window of opportunity for us to lessen the toxin exposure. Dr. Madias: Could you please address other therapeutic approaches, such as Shiga toxin binding residue to prevent absorption of toxin, hyperimmune antitoxin antiserum, and Shiga toxin receptor antagonists? Also, would you comment on the utility of antibiotic treatment of enterohemorrhagic E. coli? Dr. King: A large clinical trial using a product named “Synsorb” was recently halted. It is a form of dirt that bears the receptor for toxin that is ingested, binds the toxin, and lowers the concentration of toxin within the gut. The jury is out on the specifics of the trial. Also, newer carbohydrate ligands have been reported that effectively bind the toxin and could potentially be used therapeutically in the future. Antibiotic therapy remains somewhat controversial, yet some very important data suggest that ciprofloxacin actually increases production of the Shiga toxin. I think at this time the use of antibiotics is best avoided. Dr. Lawrence Milner (Chief, Division of Pediatric Nephrology, Floating Hospital for Infants and Children, New England Medical Center): You mentioned that patients with E. coli O15:H7 with a higher white cell count have a worse prognosis. You also demonstrated a mechanism of transportation of toxin in the body by white cells. Are children with the higher white cell count at greater risk because they have more toxin being delivered to the kidneys, or is it the high white cell numbers arriving at the kidney that are causing the damage?

Nephrology Forum: Inflammation and HUS

Dr. King: My answer has to be purely speculative; however, the association of leukocytosis with a poor outcome is well documented. I would guess that the degree of leukocytes is a reflection of the magnitude of bowel disease, and that the greater the gut inflammation, the greater the influx of Shiga toxin. The parent compound of this family of toxins comes from Shigella dysenteriae. Although we in the United States don’t see Shigella dysenteriae very often, it is a violent disease that has much more hemorrhagic colitis and bowel wall inflammation. Dr. Dana Miskulin (Nephrology Fellow, New England Medical Center): Have any in-vitro studies shown that the Shiga toxin itself causes lymphocyte degranulation? Dr. King: Little is known of the effects of Stx on lymphocytes. It is noteworthy that the B subunit can induce apoptosis in B lymphocytes. Our finding that Stx induces oxidative burst in PMNs was the first in-vitro evidence of PMN oxidative burst triggered by Stx. As I alluded to earlier, data are conflicting regarding the direct effects of toxin on monocytes/macrophages. Tesh et al clearly found that there was an increase in IL-1 production [39]. Dr. Bertrand L. Jaber (Division of Nephrology, New England Medical Center): Since Shiga toxin binds to circulating leukocytes, which can be detected by immunofluorescence, do you envision the commercialization of this assay for early detection of the hemolytic-uremic syndrome? Dr. King: I think detection of Stx bound to PMNs should be relatively straightforward. This could be achieved using standard FACS analysis with appropriate immunofluorescent probes. Dr. Gopesh K. Modi (Nephrology Fellow, New England Medical Center): What is the normal function of Gb3 receptors in the body? The density of these receptors is high in the kidneys. What is the density of these receptors in the central nervous system, and does it correlate with the relative absence of neurologic involvement in diarrhea-associated HUS? Dr. King: The natural role of Gb3 in cell function is unknown. The richest source of Gb3 is the cortical tubular cell. However, the renal vasculature and mesangial cells also express the glycolipid. The brain is another abundant source of Gb3, but I suspect that the kidney is more vulnerable to injury for several reasons. The glomerular microvasculature is distinctive because of its high flow rates, hydraulic permeability, and pressure. High blood flow rates directly increase local shear stress, and ultrafiltration increases shear stress by increasing viscosity within the glomerular capillary. Higher shear stress tends to increase platelet binding and activation. It’s a lot harder to quantitate neurologic injury than renal injury. There’s no obvious quantifiable marker, such as

1561

serum creatinine level, proteinuria, or hematuria. I suspect that the majority of these children have subclinical neurologic involvement. The small number of children who die are frequently those with severe neurologic involvement. Dr. Modi: Besides E. coli, what are the other infectious causes of HUS? Does the pathogenesis in these cases involve Shiga-toxin-like substances? Dr. King: A variety of infections are associated with thrombotic microangiopathies other than E. coli and Shigella, most notably HIV. Furthermore, rarely other enteric gram-negative rods can express Shiga toxins. Dr. Ronald D. Perrone (Division of Nephrology, New England Medical Center): I found the hypothesis that polymorphonuclear cells carry the Shiga toxin to the glomerulus very interesting. Are there experimental models with cytokines in animals in situ? Dr. King: Shiga toxin has been infused into a variety of species with differing responses. Perhaps the baboon is one of the most relevant models, and on a recent visit to Utah I suggested that Dr. Seigler admix blood and Stx prior to infusion. Potentially, the carrier cell might be a critical determinant of target organ injury. Dr. V. Balakrishnan (Division of Nephrology, New England Medical Center, Boston): I still have some conceptual problems with the neutrophils. These cells seem to be the least suited to be transporters of toxin. Your lab has shown that Shiga toxin is a potent activator of neutrophils. I find it difficult to accept that these activated neutrophils can migrate beyond the microvasculature, as they are likely to be adherent to the endothelium. You showed some autopsy data that reported approximately 900 neutrophils per 100 glomeruli [63]. Given this low ratio of neutrophils to glomeruli, it is difficult to postulate a role for significant neutrophil-induced endothelial injury in this disease. Dr. King: I suspect that the movement of PMNs into the lamina propria generates a local inflammatory response disrupting barrier function of the intestinal wall and leading to movement of toxin into the circulation and thus, exposure of the circulating cells. I think it is very unlikely that the cells move in and out of the lamina propria. I think the cells are firmly attached to various adhesion molecules. Regarding the small number of PMN cells in the glomeruli found in autopsy specimens, there are several potential explanations. First, it is possible that unbound or “free” toxin in the plasma causes direct injury to the glomerular cell and the tubular cell, that is, the PMN plays no role in injury. Second, the PMN might function as a carrier for toxin with transfer from the lower affinity binding site on the PMN to the higher affinity Gb3 on the target cell. This possibility does not implicate the PMN in the tissue injury. The third possibility is that a binding event occurs, toxin is delivered to the target,

1562

Nephrology Forum: Inflammation and HUS

and the PMN contributes to cell injury either by release of reactive oxygen species, proteases, or phospholipids. When you analyze autopsy specimens, by definition the children have died of the disease. The timing of death is usually about one week into the hospitalization and thus the glomerular disease might have passed through several phases. Dr. Balakrishnan: Most of the pathologic data indicate endothelial swelling and separation from the basement membrane. Most of the data that we have seen describe apoptosis, a very controlled form of cell death. Dr. King: There is only one report that Shiga toxins induce apoptosis in glomerular endothelial cells [76]. Whether apoptosis plays a dominant role in the glomerular injury is speculative. As you alluded to, the most common histologic manifestation is endothelial swelling and uplifting from the basement membrane, not the typical findings of apoptosis. I suspect that this is a manifestation of direct toxicity of Stx on the glomerular endothelial cell. It’s interesting to note in the studies by Forsyth et al that PMNs obtained from children with HUS caused disruption of the fibrillar architecture of reporter endothelial cells [57]. I think these cells are injured in ways other than just apoptosis. Although apoptosis might not be a dominant feature in glomerular injury, more evidence indicates that apoptosis is important in tubular injury. These cells are exquisitely sensitive to toxin, at least in part because of the high concentration of Gb3. Dr. Madias: How important is the attachment of E. coli to intestinal epithelial cells for development of pathology, and what do we know about the biology of this process? Dr. King: There are very specific proteins involved in intestinal attachment, which is essential for the disease to occur. I cannot quantify how important that is for the disease. Dr. Klemens B. Meyer (Division of Nephrology, New England Medical Center): The Gb3 receptor is found on the endothelial cells but also on mesangial and tubular cells. Do we know what its normal function is? Is there any relationship to the receptor density and does that have anything to do with susceptibility? Dr. King: I am unaware of the role that Gb3 plays in the normal function of the cell. Susceptibility does appear to differ among individuals as only 5% of infected individuals acquire HUS. I am unaware of data relating receptor density to disease risk. However, the fatty acid chain length of Gb3 does vary among individuals. Thus, this variable might alter susceptibility to the disease. Dr. Jeanine Carlson (Department of Medicine, New England Medical Center): Why don’t people get HUS twice? The Gb3 and Shiga toxin are there; why doesn’t this happen again? I’m sure that all these people don’t become vegetarians. Dr. King: First, this type of infection fortunately re-

mains relatively uncommon, and thus a child might never be exposed again. Children with E. coli also generate antibodies to specific components of the bacterial lipopolysaccharide, which might afford protection from future infection. It is clear that there are patients with recurrent TTP or HUS, the pathogenesis of which has been established for some of these patients, for example, those with defective vWF cleaving protease activity or Factor H deficiency. However, I am unaware of data confirming whether children with E. coli-induced HUS are protected from subsequent exposures. Dr. Harrington: Has the disease decreased in incidence since 1993? Did the new public health regulations make any difference? Dr. King: The disease is increasing in frequency and has been labeled as an emerging pathogen by the CDC. Dr. Harrington: Why don’t we just cook the hamburgers? Dr. King: There are no rare hamburgers in our house. Nothing pink! Reprint requests to Dr. A. King, Division Head, Scripps Clinic, Division of Nephrology, 10666 North Torrey Pines, N 239, La Jolla, California, 92037, USA. E-mail: [email protected]

REFERENCES 1. Moschowitz E: Hyaline thrombosis of the terminal arterioles and capillaries: A hitherto undescribed disease. Proc NY Pathol Soc 29:21–24, 1924 2. Gasser C, Gautier E, Steck A, et al: Hamolytisch-uramische Syndrome: bilaterale Nierenrindennekrosen bei akuten erwobenen hamolytischen anamien. Schweiz Med Wochenshr 85:905–909, 1955 3. Furlan M, Robles R, Galbusera M, et al: von Willebrand factorcleaving protease in thrombotic thrombocytopenic purpura and the hemolytic uremic syndrome. N Engl J Med 339:1578–1584, 1998 4. Tsai HM, Lian EC: Antibodies to von Willebrand factor-cleaving protease in acute thrombotic thrombocytopenic purpura. N Engl J Med 339:1585–1594, 1998 5. Rock GA, Shumak KH, Buskard NA, et al: Comparison of plasma exchange with plasma infusion in the treatment of thrombotic thrombocytopenic purpura. N Engl J Med 325:393–397, 1991 6. Feng P: Escherichia coli serotype O157:H7 infection in humans. Emerg Infect Dis 1:47–52, 1995 7. Siegler RL, Pavia AT, Christofferson RD, Milligan MK: A 20-year population-based study of postdiarrheal hemolytic uremic syndrome in Utah. Pediatrics 94:35–40, 1994 8. Koster F, Levin J, Walker L, et al: Hemolytic-uremic syndrome after shigellosis. Relation to endotoxemia and circulating immune complexes. N Engl J Med 298(17):927–933, 1978 9. Karmali MA, Petric M, Lim C, et al: The association between idiopathic hemolytic uremic syndrome and infection by verotoxinproducing Escherichia coli. J Infect Dis 151:775–782, 1985 10. Acheson DWK, Donohue-Rolfe A, Keusch GT: The family of Shiga and Shiga-like toxins, in Sourcebook of Bacterial Protein Toxins, edited by Alouf JE, Freer JH, London, Academic Press, 1991, pp 415–433 11. Takeda T, Dohi S, Igarahsi T, et al: Impairment by verotoxin of tubular function contributes to the renal damage seen in haemolytic uraemic syndrome. J Infect 27:339–341, 1993 12. Paton AW, Paton JC: Detection and characterization of Shiga toxigenic Escherichia coli by using multiplex PCR assays for stx1, stx2, eaeA, enterohemorrhagic E. coli hlyA, rfb0111 and rfb0157. J Clin Microbiol 36:598–602, 1998 13. Newburg DS, Chaturvedi P, Lopez EL, et al: Susceptibility of

Nephrology Forum: Inflammation and HUS

14. 15. 16. 17.

18. 19.

20. 21. 22. 23. 24. 25. 26. 27. 28. 29.

30. 31. 32. 33.

34. 35. 36.

hemolytic-uremic syndrome relates to erythrocyte glycosphingolipid patterns. J Infect Dis 168:476–479, 1993 Endo Y, Tsurugi K, Yutsudo T: Site of action of a Verotoxin (VT2) from Escherichia coli O157:H7 and of Shiga toxin on eukaryotic ribosomes. Eur J Biochem 171:45–50, 1988 Walters MD, Matthel U, Jay R, et al: The polymorphonuclear count in childhood hemolytic uremic syndrome. Pediatr Nephrol 3:130–134, 1989 Coad NAG, Marshall T, Rowe B, Taylor CM: Changes in the postenteropathic form of the hemolytic uremic syndrome in children. Clin Nephrol 35:10–16, 1991 Milford DV, Taylor CM, Guttridge B, et al: Haemolytic uremic syndromes in the British Isles, 1985–8: Association with Verocytotoxin-producing Escherichia coli. Part 1: clinical and epidemiological aspects. Arch Dis Child 65:716–721, 1990 Fitzpatrick MM, Shah V, Trompeter RS, et al: Interleukin-8 and polymorphonuclear leukocyte activation in hemolytic uremic syndrome. Kidney Int 42:951–956, 1992 Ikeda K, Ida O, Kimoto K, et al: Predictors for the development of haemolytic uraemic syndrome with Escherichia coli O157:H7 infections: With focus on the day of illness. Epidemiol Infect 124: 343–349, 2000 Slutsker L, Ries AA, Greene KD, et al: Escherichia coli O157:H7 diarrhea in the United States: clinical and epidemiologic factors. Ann Intern Med 126:505–513, 1997 Remis R, MacDonald KL, Riley LW, et al: Sporadic cases of hemorrhagic colitis associated with Escherichia coli O157:H7. Ann Intern Med 101:624–626, 1984 Kelly J, Oryshak A, Wenetsek M: The colonic pathological variability of infection by O157:H7 infection. Am J Surg Pathol 14:87– 92, 1990 Hunt CM, Harvey JA, Youngs ER, Reid TM: Clinical and pathological variability of infection by enterohaemorrhagic (Verocytotoxin producing) Escherichia coli. J Clin Pathol 42:847–852, 1989 Zhe L, Bell C, Buret A, et al: The effect of enterohemorrhagic Escherichia coli O157:H7 on intestinal structure and solute transport in rabbits. Gastroenterology 104:467–474, 1993 Elliott E, Li Z, Bell C, et al: Modulation of host response to Escherichia coli O157:H7 infection by anti-CD18 antibody in rabbits. Gastroenterology 106:1554–1561, 1994 Hurley BP, Jacewicz M, Thorpe CM, et al: Shiga toxins 1 and 2 translocate differently across polarized intestinal epithelial cells. Infect Immun 67:6670–6677, 1999 Thorpe CM, Hurley BP, Lincicome LL, et al: Shiga toxins stimulate secretion of interleukin-8 from intestinal epithelial cells. Infect Immun 67:5985–5993, 1999 Raqib R, Wretlind B, Anderson J, Lindberg AA: Cytokine secretion in acute shigellosis is correlated to disease activity and directed more to stool than to plasma. J Infect Dis 171:376–384, 1995 Westerholt S, Hartung T, Tollens M, et al: Inflammatory and immunological parameters in children with haemolytic uremic syndrome (HUS) and gastroenteritis–pathophysiological and diagnostic clues. Cytokine 12:822–827, 2000 Yamamoto T, Nagayama K, Satomura K, et al: Increased serum IL-10 and endothelin levels in hemolytic uremic syndrome caused by Escherichia coli O157. Nephron 84:326–332, 2000 Litalien C, Proulx F, Mariscalco MM, et al: Circulating inflammatory cytokine levels in hemolytic uremic syndrome. Pediatr Nephrol 13:840–845, 1999 Proulx F, Turgeon JP, Litalien C, et al: Inflammatory mediators in Escherichia coli O157:H7 hemorrhagic colitis and hemolyticuremic syndrome. Pediatr Infect Dis J 17:899–904, 1998 Van Setten PA, van Hinsbergh VW, van den Heuvel LP, et al: Monocyte chemoattractant protein-1 and interleukin-8 levels in urine and serum of patients with hemolytic uremic syndrome. Pediatr Res 43:759–767, 1998 Karpman D, Andreasson A, Thysell H, et al: Cytokines in childhood hemolytic uremic syndrome and thrombotic thrombocytopenic purpura. Pediatr Nephrol 9:694–699, 1995 Inward CD, Varagunam M, Adu D, et al: Cytokines in haemolytic uraemic syndrome associated with verocytotoxin-producing Escherichia coli infection. Arch Dis Child 77:145–147, 1997 Lopez EL, Contrini MM, Devotos S, et al: Tumor necrosis factor

37. 38. 39. 40.

41.

42. 43. 44. 45.

46. 47. 48. 49.

50. 51. 52. 53. 54. 55.

56.

57. 58.

1563

concentrations in hemolytic uremic syndrome patients and children with bloody diarrhea in Argentina. Pediatr Infect Dis J 14:594– 598, 1995 Van de Kar NC, Sauerwein RW, Demacker PN, et al: Plasma cytokine levels in hemolytic uremic syndrome. Nephron 71:309– 313, 1995 Martin AA, Woolven BL, Harris SJ, et al: Plasminogen activator inhibitor type-1 and interleukin-6 in haemolytic uraemic syndrome. J Paediatr Child Health 36:327–331, 2000 Tesh VL, Ramegowda B, Samuel JE: Purified Shiga-toxins induce expression of proinflammatory cytokines from murine peritoneal macrophages. Infect Immun 62:5085–5094, 1994 Van Setten PA, Monnens LA, Verstraten RG, et al: Effects of verocytotoxin-1 on nonadherent human monocytes: Binding characteristics, protein synthesis, and induction of cytokine release. Blood 88:174–183, 1997 Ramegowda B, Tesh V: Differentiation-associated toxin receptor modulation, cytokine production and sensitivity to Shiga-like toxins in human monocytes and monocytic cell lines. Infect Immun 64:1173–1180, 1996 Nash S, Stafford J, Madara JL: Effects of polymorphonuclear leukocyte transmigration on the barrier function of the cultured intestinal epithelial monolayers. J Clin Invest 80:1104–1113, 1987 Parsons PE, Sugahara K, Cott GR, et al: The effect of neutrophil migration and prolonged neutrophil contact on epithelial permeability. Am J Pathol 129:302–312, 1987 Hurley BP, Thorpe C, Acheson DWK: Neutrophil translocation across intestinal epithelial cells is enhanced by neutrophil transmigration. Infect Immun 69:6148–6155, 2001 Te Loo DM, Hinsbergh VW, Heuvel LP, Monnens LA: Detection of verocytotoxin bound to circulating polymorphonuclear leukocytes of patients with hemolytic uremic syndrome. J Am Soc Nephrol 12:800–806, 2001 Chart H, Scotland SM, Smith HR, Rowe B: Antibodies to Escherichia coli O157 in patients with haemorrhagic colitis and haemolytic uraemic syndrome. J Clin Pathol 42:973–976, 1989 Bitzan M, Moebius E, Ludwig K, et al: High incidence of serum antibodies to Escherichia coli O157 lipopolysaccharide in children with hemolytic-uremic syndrome. J Pediatr 119:380–385, 1991 Caprioli A, Luzzi I, Rosmini F, et al: Hemolytic-uremic syndrome and Vero cytotoxin-producing Escherichia coli infection in Italy. The HUS Italian Study Group. J Infect Dis 166:154–158, 1992 Pereira BJG, Shapiro L, King AJ, et al: Plasma levels of IL-1␤ and TNF␣ and their specific inhibitors in undialyzed chronic renal failure, CAPD and hemodialysis patients. Kidney Int 45:8490– 8896, 1994 Te Loo DM, Monnens LA, van Der Velden TJ, et al: Binding and transfer of verocytotoxin by polymorphonuclear leukocytes in hemolytic uremic syndrome. Blood 95:3396–3402, 2000 Hughes DA, Smith GC, Davidson JE, et al: The neutrophil oxidative burst in diarrhoea-associated haemolytic uraemic syndrome. Pediatr Nephrol 10:445–447, 1996 Brown RE, Alade SL, Knight JA, Evans BJ: Serum lipoperoxidation products in an infant with hemolytic uremic syndrome. Clin Chem 34:2382–2384, 1988 Turi S, Nemeth I, Vargha I, Matkovics B: Oxidative damage of red blood cells in haemolytic-uraemic syndrome. Pediatr Nephrol 8:26–29, 1994 Noris M, Ruggenenti P, Todeschini M, et al: Increased nitric oxide formation in recurrent thrombotic microangiopathies: A possible mediator of microvascular injury. Am J Kidney Dis 27:790–796, 1996 Facorro G, Aguirre F, Florentin L, et al: Oxidative stress and membrane fluidity in erythrocytes from patients with hemolytic uremic syndrome. Acta Physiol Pharmacol Ther Latinoam 47:137– 146, 1997 King AJ, Sundaram S, Cendoroglo M, et al: Shiga toxin induces superoxide production in polymorphonuclear cells with subsequent impairment of phagocytosis and responsiveness to phorbol esters. J Infect Dis 179:503–507, 1999 Forsyth SD, Simpleson AC, Fitzpatrick MM, et al: Neutrophilmediated endothelial injury in haemolytic uraemic syndrome. Lancet 2:411–414, 1989 Morigi M, Micheletti G, Figliuzzi M: Verotoxin-1 promotes

1564

59.

60.

61.

62.

63.

64.

65.

66.

67.

Nephrology Forum: Inflammation and HUS

leukocyte adhesion to cultured endothelial cells under physiologic flow conditions. Blood 86:4553–4558, 1995 Nevard CH, Blann AD, Jurd KM, et al: Markers of endothelial cell activation and injury in childhood haemolytic uraemic syndrome. Pediatr Nephrol 13:487–492, 1999 Inward CD, Pall AA, Adu D, et al: Soluble circulating cell adhesion molecules in haemolytic uraemic syndrome. Pediatr Nephrol 9:574–578, 1995 Inward CD, Howie AJ, Fitzpatrick MM, et al: Renal histopathology in fatal cases of diarrhoea-associated haemolytic uraemic syndrome. Pediatr Nephrol 11:556–559, 1997 Hertzke DM, Cowan LA, Schoning P, Fenwick BW: Glomerular ultrastructural lesions of idiopathic cutaneous and renal glomerular vasculopathy of greyhounds. Vet Pathol 32:451–459, 1995 Taylor CM, Williams JM, Lote CJ, et al: A laboratory model of toxin-induced hemolytic uremic syndrome. Kidney Int 55:1367– 1374, 1999 Obrig TG, Louise CB, Lingwood CA, et al: Endothelial heterogeneity in Shiga toxin receptors and responses. J Biol Chem 268: 15484–15488, 1993 Pijpers AH, van Setten PA, van den Heuvel LP, et al: Verocytotoxin-induced apoptosis of human microvascular endothelial cells. J Am Soc Nephrol 12:767–778, 2001 Hughes AK, Stricklett PK, Schmid D, Kohan DE: Cytotoxic effect of Shiga toxin-1 on human glomerular epithelial cells. Kidney Int 57:2350–2359, 2000 Hughes AK, Stricklett PK, Kohan DE: Shiga toxin-1 regulation

68. 69. 70. 71. 72. 73.

74. 75. 76.

of cytokine production by human proximal tubule cells. Kidney Int 54:1093–1106, 1998 Simon M, Cleary TG, Hernandez JD, Abboud HE: Shiga toxin 1 elicits diverse biologic responses in mesangial cells. Kidney Int 54:1117–1127, 1998 Kodama T, Nagayama K, Yamada K, et al: Induction of apoptosis in human renal proximal epithelial cells by Escherichia coli verocytotoxin 1 in vitro. Med Microbiol Immunol (Berl) 188:73–78, 1999 Karpman D, Hakansson A, Perez MT, et al: Apoptosis of renal cortical cells in the hemolytic-uremic syndrome: In vivo and in vitro studies. Infect Immun 66:636–644, 1998 Taguchi T, Uchida H, Kiyokawa N, et al: Verotoxins induce apoptosis in human renal tubular epithelium derived cells. Kidney Int 53:1681–1688, 1998 Van Setten PA, van Hinsbergh VWM, van der Velden TJAN, et al: Effects of TNF␣ on verocytotoxin in purified human glomerular microvascular endothelial cells. Kidney Int 51:1245–1256, 1997 Harel Y, Silva M, Giroir B, et al: A reporter transgene indicates renal-specific induction of tumor necrosis factor (TNF) by Shigalike toxin. Possible involvement of TNF in hemolytic uremic syndrome. J Clin Invest 92:2110–2116, 1993 Isogai E, Isogai H, Kimura K, et al: Role of tumor necrosis factor alpha in gnotobiotic mice infested with an Escherichia coli O157:H7 strain. Infect Immun 66:197–202, 1998 Lingwood CA: Role of verotoxin receptors in pathogenesis. Trends Microbiol 4:147–153, 1996 Pijpers AH, van Setten PA, van den Heuvel LP, et al: Verocytotoxin-induced apoptosis of human microvascular endothelial cells. J Am Soc Nephrol 12:767–778, 2001