Acute inflammation in the pathogenesis of hemolytic-uremic syndrome

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

NEPHROLOGY FORUM

Acute inflammation in the pathogenesis ofhemolytic-uremic syndrome

Principal discussant: Andrew J. King

Scripps Clinic, La Jolla, California, USA

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 exami-nation were significant only for mildly icteric sclera. The lungswere clear to auscultation and percussion. The jugular veins werenot distended, and no murmurs, rubs, or gallops were heard.Initial abdominal examination was notable for hypoactive bowelsounds, although the abdomen was soft and non-tender topalpation. Rectal examination revealed brown stool that was4� positive for occult blood and later found to be positive forfecal leukocytes. The neurologic examination was limited dueto the confused state of the patient, but it did not reveal anyfocal defects.

Laboratory testing revealed: serum sodium, 135 mEq/L; po-tassium, 3.4 mEq/L; chloride, 102 mEq/L; and bicarbonate,19 mmol/L. The serum creatinine was 3.7 mg/dL; blood urea

CASE PRESENTATION nitrogen, 110 mg/dL; and total bilirubin, 4.2 mg/dL with normalThe patient, a 68-year-old white woman, presented to her hepatic transaminases. Hematocrit was 25% with a hemoglobin

primary physician with two to three days of bloody diarrhea; of 8.1 g/dL. The leukocyte count was 29,700/mm3 with 86%bilateral, crampy, lower abdominal pain, and fever to 101�F. polymorphonuclear cells, 3% bands, 8% lymphocytes, and 3%Initially, she was treated with clear liquids as well as ciproflox- mononuclear cells. The platelet count was 49,000/mm3. Urinaly-acin and loperamide. Over the next 24 hours, diarrhea, nausea, sis revealed a specific gravity of 1.011; pH 5; 2� protein, 2�and vomiting increased, causing her to go to the emergency blood, and 5 to 7 red blood cells/high-power field in the sedi-ward. Her family had noted progressive confusion over the ment. No cellular casts were present. A chest radiograph wasprevious 24 hours. There was no history of recent travel, nor unremarkable, and a KUB radiograph revealed an ileus with-was there any known contact with infected individuals. Her

out evidence of free air.dietary history was significant for the ingestion of hamburgerOver the following 12 hours, the patient’s abdominal exami-meat three days prior to the onset of symptoms. She had had

nation changed, with the development of overt peritoneal signstype II diabetes mellitus for eight years, hypertension for 25leading to an exploratory laparotomy. Extensive edema of theyears, and hypothyroidism. Her medications on admission weredistal ileum and colon with evidence of infarction was found.hydrochlorothiazide, 25 mg/day; metoprolol, 25 mg twice daily;A partial colectomy, partial ileal resection, and ileostomy wereand thyroxine in addition to the ciprofloxacin and loperamide.performed. Pathology of the colon revealed extensive bowelThere was no family history of renal or hematologic disease.wall edema with regions of infarction, sloughing of mucosa,A retired secretary who lived alone, she was too confused toand local microangiopathy. The bowel had marked infiltration ofgive a review of systems, although her family stated that prior

to this acute illness she had been doing well. polymorphonuclear cells with crypt abscesses. Her postoperativePhysical examination in the emergency ward revealed a mark- course was complicated by oliguria and rapidly worsening renal

edly confused woman unable to state her name or the date. Her function. Hemodialysis was initiated. She was treated withbroad-spectrum antibiotics and underwent a total of six plasmaexchange treatments using fresh frozen plasma as the replace-

The Nephrology Forum is funded in part by grants from Amgen, ment fluid. A stool culture from her initial visit with her primaryIncorporated; Merck & Co., Incorporated; Dialysis Clinic, Incorpo- physician grew Escherichia coli O157:H7. Her hospital courserated; and Bristol-Myers Squibb Company.

was prolonged (4 months total) and complicated by adult respi-ratory distress syndrome (ARDS), pancreatitis, seizures, Clos-Key words: Shiga toxin, enterohemorrhagic E. coli, von Willebrand

factor-cleaving protease, hemorrhagic enterocolitis, thrombotic mi- tridium difficile colitis, and Staphylococcus aureus bacteremia.croangiopathy. She never regained renal function, but her mental status re-

turned to normal. 2002 by the International Society of Nephrology

1553

Nephrology Forum: Inflammation and HUS1554

DISCUSSION Shiga toxin-producing Escherichia coli

Shiga toxin (Stx)-producing Escherichia coli has beenDr. Andrew J. King (Division Head, Division of Ne-phrology, Scripps Clinic and Green Hospital, La Jolla, designated by the U.S. Centers for Disease Control and

Prevention as an “emerging pathogen” [6]. As manyCalifornia, USA): This unfortunate woman provides adramatic illustration of the devastating nature of adult as 90% of children with HUS have antecedent STEC,

typically of the serotype O157:H7 [7]. The association ofhemolytic-uremic syndrome (HUS) induced by entero-hemorrhagic Escherichia coli (EHEC). She had multi- Shiga toxin-producing Shigella dysenteriae with HUS was

first made by Koster et al in 1978 [8], followed by obser-organ involvement, including the gut, central nervoussystem, and kidney, a presentation not too divergent vations indicating an association of E. coli cytotoxins

with the syndrome by Karmali et al in 1985 [9]. Thefrom the patient with thrombotic microangiopathy firstdescribed by Eli Moschowitz in 1924 [1]. Thrombotic implicated E. coli cytotoxins, Stx1 and Stx2, belong to a

family of bacterial protein toxins related to the parentmicroangiopathy (TMA) is a term applied to a group ofdisorders that present with thrombocytopenia, evidence Shiga toxin, which share the same enzymatic action and

recognize related host-neutral glycolipid receptors. In-of red blood cell destruction, and a characteristic patho-logic appearance in the microvasculature, including en- terest in this disease has been fostered by considerable

media attention stemming from several notable out-dothelial cell injury and platelet deposition. Thromboticmicroangiopathy is associated with a wide range of infec- breaks, including one in 1993 related to undercooked

hamburgers at “Jack in the Box” restaurants in the north-tions, drugs, tumors, and other disorders. The focus ofwest United States and one more recently in Japan. De-this report, however, is HUS triggered by EHEC. Firstpending on the outbreak, patients with documenteddescribed by Gasser et al in 1955 [2], HUS is a termSTEC have a risk of developing HUS of approximatelyapplied to a TMA with predominant renal involvement.5% to 10% [7]. Details regarding the genetics, regulation,Although there had been speculation that HUS andstructure, enzymatic specificity, and physiology of thethrombotic thrombocytopenic purpura (TTP) were dif-Stx family are found in several reviews [10–12]. Briefly,fering manifestations of the same disease, it recently hasStx is a heterodimer consisting of an enzymatically activebecome clear that for the majority of adults with TTP,A subunit surrounded by five identical B subunits. Thethe key pathologic feature is an immunoglobulin thatreceptor for the toxin is a blood group active glycolipid,inactivates circulating von Willebrand factor (vWF)-globotriaosylceramide (Gb3). The distribution and fattycleaving protease [3, 4]. Failure to properly metabolizeacid chain composition of Gb3 are believed to play anaturally occurring ultralarge multimers of vWF leadsrole in susceptibility of various cells to the toxin [13].to increased binding to platelet glycoproteins under con-Upon binding, internalization and activation of the Aditions of high shear stress, that is, that which exists insubunit lead to depurination of a specific adenosine inthe microvasculature. These features of TTP explain the28S ribosomal RNA; the end result is irreversible inhibi-remarkable efficacy of plasma exchange in this disordertion of protein elongation [14]. Glomerular endothelial[5]. By contrast, vWF-cleaving protease activity is nor-cells and, to a greater extent, renal tubular epithelial cells,mal in patients with HUS, and the efficacy of plasmaexpress Gb3, thereby contributing to the propensity of pa-exchange is much less clear.tients infected with STEC to develop acute renal failure.Unlike TTP, patients with diarrhea-induced HUS have

abundant evidence of an acute inflammatory response,Clinical markers of acute inflammationthe magnitude of which predicts clinical outcome. In this

Forum, I will review evidence suggesting that inflamma- The woman presented today had high fever and markedleukocytosis, common findings in patients presenting withtory cells and their byproducts play a major role in (1)

loss of intestinal barrier function, thereby promoting STEC. Leukocytosis can be extreme, as in leukemoidreactions, and it is a positive predictor of acute mortalitymovement of endotoxin and Shiga toxins into the circula-

tion; (2) delivery of Shiga toxins to target organs; (3) and residual nephropathy [15–18]. Children with “diar-rhea-positive” HUS (D�HUS) have significantly highersensitization of target organs to Shiga toxins by increas-

ing glycolipid receptor expression; and (4) direct injury leukocyte counts on presentation than do children withatypical “diarrhea-negative” HUS (D-HUS); this findingof target organ endothelium. Although definitive evi-

dence for a role of inflammatory cells and their mediators suggests that the intestinal disease is an important factorin generating the leukocytosis [15]. In the Osaka out-is lacking in humans, on circumstantial evidence, the

case is compelling. Patients with HUS due to Shiga toxin- break of 1996, both leukocyte count and C-reactive pro-tein were higher in the group of children with STECproducing Escherichia coli (STEC) rarely present in the

first 24 to 48 hours and thus, as with other forms of acute who developed HUS compared to infected children whodid not [19]. Severe gastrointestinal disease also portendsrenal failure (ARF), we are left looking for the “smoking

gun,” the mediators that cause the ARF. 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 hesion molecule CD-18 prevented intestinal infiltrationhemorrhagic enterocolitis of PMN and rendered histologic and functional protec-

tion [25].The gastrointestinal disease in patients with STECShiga toxin traverses polarized intestinal human epi-varies from watery diarrhea to the most severe form,

thelial cells grown in tissue culture and retains biologichemorrhagic colitis. Colonic inflammation might play aactivity [26]. Recent evidence suggests that Stx inducesrole in local intestinal microangiopathy, the transfer ofa chemokine response from human intestinal epithelialStx from the bowel lumen through the lamina propriacells [27]. Using a human colonic epithelial cell line, Stxand into the circulation, and the generation of a systemictreatment led to “superinduction” of interleukin-8 (IL-8)inflammatory response (Fig. 1). Although STEC was notwith increase in both IL-8 mRNA and protein, the latterpreviously thought to provoke marked intestinal in-occurring despite the known inhibitory effects of Stx onflammation, Slutsker et al noted that patients with STECprotein elongation [27]. Secretion of IL-8 into the laminafrequently have fecal leukocytes [20]. Indeed, colonicpropria might create a chemokine gradient sufficient tobiopsy specimens from patients with STEC often haverecruit circulating PMNs. Both IL-8 and tumor necrosiscrypt abscesses and infiltrates of monocytes and poly-factor-� (TNF-�) are more elevated in stools of patientsmorphonuclear cells (PMNs) [21–23]. In rabbits, infec-with Stx-producing S. dysenteriae compared to patientstion with E. coli O157:H7 leads to severe inflammatoryinfected with non-Stx-producing S. flexneri [28]. Severalcolitis with PMN infiltrates [24]. In this model, infusion

of monoclonal antibodies directed toward the PMN ad- investigators have found that patients with STEC-induced


Table 1. Blood and urinary cytokines in HUS

Cytokine Levels References

IL-8 Blood Increased [18, 29-33, 35]Urine Increased [33]

TNF-� Blood Increased [36]Unchanged or infrequent [18, 34, 35, 37]

Urine No or scant dataIL-1� Blood Increased [35]

Unchanged or infrequent [37]Urine No or scant data

IL-6 Blood Increased [30-32, 34, 37, 38]Unchanged or infrequent [35]

Urine Increased [34]IL-10 Blood Increased [30-32]

Decreased [29]Fig. 2. Indirect immunofluorescence assay of Stx-2 bound to polymor-Urine No or scant dataphonuclear cells (PMNs) in a patient presenting with hemolytic-uremicsyndrome (HUS). Reprinted with permission from the American Soci-ety of Nephrology [45].

HUS have elevated circulating levels and urinary excre-tion of IL-8 (Table 1). The magnitude of IL-8 elevation

ated with high circulating levels of IL-1 and TNF-�.correlates with markers of PMN degranulation and withIncreases in circulating inflammatory cytokines likely arepatient outcome [18]. Other potential intestinal sourcestriggered by the colonic infection and possibly by theof chemokine production are the resident and influxinglow-grade endotoxemia (Table 1). When assessing theleukocytes in the lamina propria. Murine macrophagesdegree of cytokine elevation, note that renal failure inde-exposed to Stx increase production of both TNF-� andpendently contributes to high levels [49]. Futhermore,IL-6 [39]. Van Setten et al have shown in human mono-patients with uncomplicated STEC also have increasedcytes that Stx increases production of IL-1�, TNF-�, IL-levels. Despite these caveats, it appears that patients with6, and IL-8 and does not inhibit protein synthesis [40].HUS have elevated circulating inflammatory cytokines. IBinding sites for Stx on the monocytes were increasedwill discuss the role of these cytokines as possible “co-with lipopolysaccharide exposure. By contrast, Rameg-conspirators” in vascular injury in a few moments. How-owda and Tesh found that human monocytes and mono-ever, local production of cytokines in the gut wall mightcytic cell lines were quite resistant to Stx unless the cellsbe an important second step in the generation of intesti-were pretreated with phorbol esters [41].nal inflammation because of their ability to activate intes-Several potential consequences of a local inflamma-tinal endothelial cells, thereby increasing adhesion mole-tory response in the lamina propria exist. Transepithelialcule expression and PMN diapedesis.migration of PMNs can lead to transient loss of colonic

barrier function and cause the passage of luminal con- Direct effects of Shiga toxins on inflammatory cellstents into the circulation [42, 43]. In one recent study,

Increasing evidence, both in vitro and in vivo, suggestsmovement of PMNs from the basolateral to the apicalthat Stx binds to PMN, triggers PMN activation, andside of a human intestinal epithelial cell line allows in-promotes adherence of PMNs to target endothelium. Tecreasing translocation of both Stx1 and Stx2 in the oppo-Loo et al recently found that Stx binds to PMNs (notsite direction [44]. Loss of barrier function could contrib-via Gb3) and raised the possibility that these cells areute to direct entry of Stx into the circulation, as well asdirect carriers of the toxin from the gut to target organsendotoxemia that would promote a systemic cytokine[50]. Further, these investigators report that circulatingresponse. To date, no one has been able to detect circu-PMNs and, to a lesser extent monocytes from patientslating Stx in patients with STEC. Recent studies, how-with D�HUS, have Stx bound to the surface (Fig. 2) [45].ever, have revealed Stx bound to circulating PMNs inAs I said before, these data demonstrated unequivocallypatients with D�HUS, thus confirming entry into thethat Stx penetrates the gut in STEC-induced HUS andcirculation [45]. Most cellular responses in vitro, how-suggest that PMNs transport Stx to target cells. Activatedever, are elicited at Stx doses well below the lower limitsPMNs contribute to tissue injury by releasing proteolyticof detection in plasma. Sepsis syndrome is very uncom-enzymes, reactive oxygen intermediates, and phospho-mon in children with STEC, except in the setting oflipid derivatives. Fitzpatrick et al detected in childrenbowel infarction. In this regard, several investigatorswith D�HUS sustained increases in �1-antitrypsin-com-have found specific antibodies to the lipopolysaccharideplexed elastase, a PMN degranulation product thatof Stx-producing E. coli [46–48]. Endotoxemia is muchpeaked approximately 3 to 4 days into admission [18].more prominent in patients with S. dysenteriae, an infec-

tion that is more likely to cause HUS and which is associ- Interestingly, this rise was preceded by an increase in


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

IL-8, a chemokine that also activates PMNs. Others have or to undergo phagocytosis. Noris et al noted an approxi-mate twofold increase in superoxide release of PMNsconfirmed the increase in elastase but were unable to

detect activation of peripheral blood PMNs using chemi- isolated from adults with recurrent TMA compared tocontrols [54]. These investigators also noted higherluminescence to assess for oxidative burst [51].

Evidence exists that children with HUS are under oxi- plasma nitrite/nitrate levels, a rise indicative of increasedproduction of another reactive oxygen species, nitric ox-dative stress. Malonyldialdehyde blood levels are in-

creased and there is oxidative damage to erythrocytes ide (NO). The source of NO remained speculative, al-though it was observed that the serum of patients with[52–55]. Patients with acute renal failure from other

causes, however, also are under oxidative stress, and TMA induced higher levels of NO production from re-porter human umbilical vein endothelial cells (HUVEC)the clinical literature on HUS in general rarely includes

appropriate controls. We have shown that Stx induces compared to control serum [54].The question remains open as to whether toxin-stimu-release of superoxide from PMNs in vitro in a dose-

dependent fashion and suggested that the toxin directly lated PMNs actually contribute to tissue injury, eitherin the gut or in remote target organs. The first evidencetriggers oxidative burst in PMNs (Fig. 3) [56]. These cells

were subsequently less able to respond to phorbol esters supporting this possibility came from Forsyth et al, who


noted that PMNs isolated from children with diarrhea- Potential role of inflammation on target organ injuryinduced HUS are more adherent to reporter HUVEC Whereas inflammatory cells are abundant in the gutthan are control PMNs [57]. Further, these PMNs in- of patients with STEC-induced HUS, their presence isduced morphologic evidence of endothelial cell injury far more subtle in the primary target organs, the centralin the form of fibronectin breakdown, an effect that was nervous system, and the kidney. In a case control studyblocked in a subset of the patients by a CD-18 antibody. of 15 autopsy specimens from children who died duringOur laboratory has examined this issue in vitro using a the acute phase of D�HUS, however, the median num-co-culture model, whereby normal PMNs are stimulated ber of PMNs per 100 glomeruli was 929 compared toby Stx and placed in co-culture with reporter human 156 for controls (children who died with normal renalsaphenous vein endothelial cells (SVEC) (abstract, Mur- function) [61]. The median peripheral neutrophil count

in this autopsy series was 21 � 109/L. Although animalray SL et al, J Am Soc Nephrol 10:533A, 1999). We dem-models have been problematic in the study of this dis-onstrated that Stx-stimulated PMNs induced apoptosisease, a naturally occurring disease in the greyhound dog,in the co-cultured SVEC, whereas untreated PMNs or Stx“Greentrack disease” (a.k.a. “Alabama rot”), bears aalone did not. Other investigators have confirmed humanstriking resemblance to the human disease; in this model,microvascular endothelial cell cytotoxicity of PMNs fromPMNs are present in the glomerulus [62]. In a recentlypatients with TMA, noting significant protection by adescribed rat model of HUS induced by ricin, a toxinmetalloprotease inhibitor (BB3103), and several reac-with similar ribotoxic effects to Stx, all the features seentive oxygen species inhibitors (abstract, Ruiz-Torres et al,in human HUS exist, including glomerular thrombosisJ Am Soc Nephrol 11:506A, 2000). Taken together, theseand injury [63]. The glomeruli of these rats had a substan-results suggest that Stx-stimulated PMNs in the laminatial increase in glomerular macrophages within eightpropria (resident and recruited) contribute to endothe-hours of ricin infusion and increased circulating mono-lial cell damage in the intestine and lead to hemorrhage,cyte chemotactic protein-1, TNF-�, IL-1, and IL-6. Theloss of barrier function, and local microangiopathy. Thorperecent finding by te Loo et al increases the likelihoodet al also have shown that human intestinal microvascu-that PMNs are an important carrier of Stx to the targetlar endothelial cells express high levels of Gb3 and areendothelium [45]. The number of Stx-positive cellsexquisitely sensitive to the ribotoxic effects of Stx1 anddropped off dramatically after 3 to 5 days in the patients2 [27]. Whether Stx-activated PMNs act on target organswith sequential measurements. Although the fate ofis an intriguing question that I will discuss later.those Stx-laden PMNs is unclear, it was shown convinc-The findings of Forsyth et al that PMNs from patientsingly that transfer of toxin occurs from the as-yet-unde-with TMA are more adherent to endothelial cells raisesfined PMN binding site to Gb3 on glomerular endothelialthe possibility that Stx might alter adhesion molecularcells in vitro [50].

expression [57]. Indeed, Morigi and colleagues foundBy definition, the kidney is the primary target organ

that Stx1 pretreatment of HUVEC increased the bindingin STEC-induced HUS. This vulnerability is in large part

of leukocytes flowing across the endothelial monolayer due to the high density of Gb3 receptors in the renalin a dose-dependent fashion [58]. The magnitude of ef- microvasculature and renal tubular epithelial cells [64].fect was comparable to IL-1� and was inhibited by mono- Indeed, renal tubular epithelial cells are perhaps theclonal antibodies to various adhesion molecules (E-selec- richest cellular source of Gb3 yet identified, equaling ortin, ICAM-1, and VCAM). Further, TNF-� together with perhaps even exceeding that of Vero cells. The openStx led to greater binding than with either TNF-� or Stx question is how much this vulnerability is due to ambientalone. In preliminary studies, we have observed that cytokine effects (Table 2). The laboratory of Hughes etflowing PMNs treated with Stx bind better to untreated al [67] and our laboratory have shown that these cellsHUVEC than do vehicle-treated PMNs, an effect that are exquisitely sensitive to the ribotoxic effects of Stx;is magnified by pretreatment of HUVEC with IL-1 (per- this work raises the possibility that acute tubular necrosissonal observations). In the acute phase of D�HUS, solu- plays a role in the renal failure (abstract, King et al,ble VCAM levels are increased [59, 60]. Once again, the J Am Soc Nephrol 9:580A, 1998). Human glomerulareffects of acute renal failure per se might play a role, as microvascular endothelial cells (GMEC) express lesserVCAM levels in the HUS patients were lower than in amounts of Gb3, and in our studies and those of vandialyzed controls without HUS [59]. Our ability to detect Setten et al [72], GMEC were less sensitive to the ribo-endothelial cell activation in vivo is limited. Neverthe- toxic effects of Stx than were other microvascular endo-less, these in vitro studies do support the possibility that thelial cells, such as human intestinal or foreskin endo-increased levels of pro-inflammatory cytokines have an thelial cells. The variation in the literature with respectaccelerant effect on the binding of toxin-laden PMNs to to human GMEC sensitivity is likely due to differences

in source of cells, method of isolation, culture conditions,target endothelium.


Table 2. Effects of cytokines on human kidney cell sensitivity to Stx

Cell Cytokine Gb3 density/binding Ribotoxicity Apoptosis Reference

Glomerular endotheliumTNF-� Increased density Increased ND [40]TNF-� ND ND Induceda [65]TNF-� No change No change ND [64]LPS No change No change ND

Glomerular epitheliumIL-1 Increase Increase ND [66]TNF-� Increase Increase NDIL-6 No change No change NDLPS Increase Increase ND

MesangiumIL-1 ND Increase ND [68]TNF-� ND Increase ND

Renal tubular epitheliumProximal tubules IL-1 Increased binding Increase ND [67]

TNF-� No change No change NDIL-6 No change No change NDLPS No change Increase ND

Proximal tubules TNF-� ND ND Increased [69]Cortical cells TNF-� ND ND Increased [70]ACHN cell line TNF-� ND ND Increased [71]

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

and inter-human variability (perhaps related to Gb3 fatty vulnerability to undergo apoptosis. If the finding that Stxincreases adhesion molecule expression holds true foracid chain variance). Pretreatment of GMEC with TNF-�

substantially increased both receptor density and sensi- human GMEC, this mechanism could serve as an acceler-ant to glomerular injury by increasing Stx delivery.tivity to Stx [72]. In keeping with these findings, recent

data suggest that Stx induces apoptosis of GMEC onlyConclusionif cells are pretreated with TNF-� [65]. Taken together,

these data suggest that the magnitude of cytokine re- Shiga toxin-producing Escherichia coli is an “emergingsponse determines the expression of the glomerular dis- pathogen” that causes a devastating systemic disease,ease. However, investigators who have carefully mea- hemolytic-uremic syndrome. The gastrointestinal dis-sured levels of inflammatory cytokines have shown an ease is associated with fecal leukocytes and a robustinconsistent relationship with disease severity, particu- inflammatory response in the intestinal lamina propria.larly when appropriate controls are included, such as Shiga toxin itself likely triggers chemokine production bypatients with uncomplicated STEC or other causes of intestinal epithelial cells and/or resident macrophages,renal failure (Table 1). Other possible sources of in- which leads to the influx of circulating inflammatory cellsflammatory cytokines include activated monocytes ad- into the lamina propria. Once established, the local in-herent to glomerular cells or kidney cells. In support of flammation could promote injury both to the intestinalthe latter possibility, Stx infusion into transgenic mice epithelium and endothelium and lead to intestinal hem-bearing a chloramphenacol acetyl transferase (CAT) orrhage, loss of barrier function, and generation of in-gene indicative of TNF-� synthesis revealed increased flammatory cytokines. Loss of barrier function allowsCAT activity exclusively in the kidney [73]. Furthermore, Stx to penetrate into the circulation, where it circulatesrenal levels of TNF-�, IL-1, and IL-6 are increased in freely or bound to PMN and is delivered to target organs.gnotobiotic mice inoculated with E. coli O157:H7 [74]. Injury of target cells is promoted by increased ambientHughes and colleagues make a strong case for renal inflammatory cytokines, possibly as a result of increasedtubular epithelial cells as a possible source of local cyto- receptor density, and it culminates in organ dysfunction.kine production [67]. They demonstrated potent induc- The excessive vulnerability of the glomerulus likely re-

lates to direct contact with toxin-loaded PMNs and istion and release of TNF-� and IL-1 from non-trans-formed human proximal tubule cells exposed to Stx, potentially increased by high local concentrations of pro-

inflammatory cytokines generated by renal tubular epi-effects that were magnified by lipopolysaccharide. Takentogether, these data suggest that inflammatory cytokines thelial cells. The fact that toxin-loaded PMNs are present

when the patient is admitted to the hospital raises thecontribute to glomerular injury by increasing Gb3 ex-pression on GMEC, thereby heightening the sensitivity possibility that leukopheresis might be a therapeutic op-

tion that will reduce target organ injury.of these cells to the ribotoxic effects of Stx and their


QUESTIONS AND ANSWERS recurrent HUS who would become anuric whenever sheacquired an infection with a fever greater than 101�F.Dr. Nicolaos E. Madias (Executive Academic Dean,Whatever the insult is to the endothelium, three manifes-Tufts University School of Medicine, Boston, Massachu-tations of this injury are the loss of the non-thrombogenicsetts): You mentioned that approximately 5% to 10% ofproperties of the monolayer, exposure of the basementchildren with documented Shiga toxin producing E. colimembrane, and platelet aggregation, all of which leaddevelop HUS. In view of the importance of pro-inflam-to a characteristic pathologic finding.matory cytokines in the pathogenesis of the syndrome,

Dr. Andrew S. Levey (Chief, Division of Nephrology,have polymorphisms in cytokine genes been examinedNew England Medical Center, Boston): You alluded toregarding susceptibility to HUS?a therapy that could “clean up” the toxins from the whiteDr. King: That is an interesting question, given thecells. Do you have any speculations on ways that thatpotential impact of these polymorphisms on the acutemight be done?inflammatory response. However, there are no data in

Dr. King: Leukopheresis would directly remove thepatients with HUS. This raises the larger question as totoxin-coated white cells. Potentially, you could create anwhy only a minority of patients infected with STEC getaffinity column using Gb3 or another molecule with aHUS. Lingwood et al have raised the possibility thatGal-Gal disaccharide. In most forms of acute renal fail-differences among individuals in the fatty acid composi-ure in outpatients, we first see the patient two or threetion of the glycolipid receptor alter binding characteris-days into the disease, when many cellular responsestics of Stx and thus susceptibility to its toxic effects [75].might have occurred. Perhaps in the first 24 hours ofSupport for this hypothesis remains limited. I think it isSTEC infection there is a robust influx of polymorphonu-more compelling that the children with the most robustclear cells in the glomerulus. What I find intriguing aboutevidence of systemic inflammation appear to be thosethe report of te Loo et al was that even days after presen-at greater risk. I suspect those are the children who havetation there was evidence of toxin-bearing polymorpho-greater movement of toxin into the circulation (a highernuclear cells in the blood [45]. This finding would imply

dosage of Stx) and higher concentrations of pro-inflam-ongoing endothelial cell injury early in the hospitaliza-

matory cytokines making target organs more vulnerable. tion and a potential therapeutic window of opportunityDr. Madias: What do we know about the mechanism for us to lessen the toxin exposure.

of binding of Shiga toxin to polymorphonuclear cells? Dr. Madias: Could you please address other therapeu-Dr. King: These recent data are based on two reports tic approaches, such as Shiga toxin binding residue to

[45, 50]. The nature of the receptor is unclear; however, prevent absorption of toxin, hyperimmune antitoxin an-it is not Gb3. tiserum, and Shiga toxin receptor antagonists? Also,

Dr. John T. Harrington (Dean, Tufts University would you comment on the utility of antibiotic treatmentSchool of Medicine, Boston): You told us that 70% of of enterohemorrhagic E. coli?the cases of HUS are related to O157:H7. Do we know Dr. King: A large clinical trial using a product namedwhether the residual 30% of E. coli share the same char- “Synsorb” was recently halted. It is a form of dirt thatacteristics? bears the receptor for toxin that is ingested, binds the

Dr. King: A large number of serotypes of E. coli are toxin, and lowers the concentration of toxin within thecapable of producing Stx1 and Stx2. To my knowledge, gut. The jury is out on the specifics of the trial. Also,the serotype is not a major factor in determining the newer carbohydrate ligands have been reported that ef-relative risk of acquiring HUS. However, it appears that fectively bind the toxin and could potentially be usedthe majority of patients with HUS have STEC that pro- therapeutically in the future. Antibiotic therapy remainsduce Stx2. Thus, this subtype probably is more likely to somewhat controversial, yet some very important datainduce systemic disease. suggest that ciprofloxacin actually increases production

Dr. Harrington: Andy, you said that “cleaving prote- of the Shiga toxin. I think at this time the use of antibiot-ase activity” was normal in HUS but absent in TTP, yet ics is best avoided.we see the same pathologic findings in both diseases. Dr. Lawrence Milner (Chief, Division of PediatricHow do you account for that discrepancy? Nephrology, Floating Hospital for Infants and Children,

Dr. King: I think that thrombotic microangiopathies New England Medical Center): You mentioned that pa-are diseases associated with endothelial cell injury. Per- tients with E. coli O15:H7 with a higher white cell countsons with rare forms of recurrent TTP are not in a state have a worse prognosis. You also demonstrated a mecha-of unrelenting thrombotic microangiography. Thus there nism of transportation of toxin in the body by white cells.must be an additional provocateur triggering the re- Are children with the higher white cell count at greatersponse. The deficiency in cleaving protease activity in risk because they have more toxin being delivered tothese patients makes them susceptible to the full expres- the kidneys, or is it the high white cell numbers arriving

at the kidney that are causing the damage?sion of microvascular disease. I managed a woman with


Dr. King: My answer has to be purely speculative; serum creatinine level, proteinuria, or hematuria. I sus-pect that the majority of these children have subclinicalhowever, the association of leukocytosis with a poor

outcome is well documented. I would guess that the neurologic involvement. The small number of childrenwho die are frequently those with severe neurologicdegree of leukocytes is a reflection of the magnitude of

bowel disease, and that the greater the gut inflammation, involvement.Dr. Modi: Besides E. coli, what are the other infec-the greater the influx of Shiga toxin. The parent com-

pound of this family of toxins comes from Shigella dysent- tious causes of HUS? Does the pathogenesis in thesecases involve Shiga-toxin-like substances?eriae. Although we in the United States don’t see Shigella

dysenteriae very often, it is a violent disease that has Dr. King: A variety of infections are associated withthrombotic microangiopathies other than E. coli and Shi-much more hemorrhagic colitis and bowel wall inflam-

mation. gella, most notably HIV. Furthermore, rarely other en-teric gram-negative rods can express Shiga toxins.Dr. Dana Miskulin (Nephrology Fellow, New En-

gland Medical Center): Have any in-vitro studies shown Dr. Ronald D. Perrone (Division of Nephrology,New England Medical Center): I found the hypothesisthat the Shiga toxin itself causes lymphocyte degranu-

lation? that polymorphonuclear cells carry the Shiga toxin tothe glomerulus very interesting. Are there experimentalDr. King: Little is known of the effects of Stx on

lymphocytes. It is noteworthy that the B subunit can models with cytokines in animals in situ?Dr. King: Shiga toxin has been infused into a varietyinduce apoptosis in B lymphocytes. Our finding that Stx

induces oxidative burst in PMNs was the first in-vitro of species with differing responses. Perhaps the baboonis one of the most relevant models, and on a recent visitevidence of PMN oxidative burst triggered by Stx. As I

alluded to earlier, data are conflicting regarding the di- to Utah I suggested that Dr. Seigler admix blood andStx prior to infusion. Potentially, the carrier cell mightrect effects of toxin on monocytes/macrophages. Tesh

et al clearly found that there was an increase in IL-1 be a critical determinant of target organ injury.Dr. V. Balakrishnan (Division of Nephrology, Newproduction [39].

Dr. Bertrand L. Jaber (Division of Nephrology, New England Medical Center, Boston): I still have some con-ceptual problems with the neutrophils. These cells seemEngland Medical Center): Since Shiga toxin binds to cir-

culating leukocytes, which can be detected by immuno- to be the least suited to be transporters of toxin. Yourlab has shown that Shiga toxin is a potent activator offluorescence, do you envision the commercialization of

this assay for early detection of the hemolytic-uremic neutrophils. I find it difficult to accept that these acti-vated neutrophils can migrate beyond the microvascula-syndrome?

Dr. King: I think detection of Stx bound to PMNs ture, as they are likely to be adherent to the endothelium.You showed some autopsy data that reported approxi-should be relatively straightforward. This could be

achieved using standard FACS analysis with appropri- mately 900 neutrophils per 100 glomeruli [63]. Given thislow ratio of neutrophils to glomeruli, it is difficult toate immunofluorescent probes.

Dr. Gopesh K. Modi (Nephrology Fellow, New Eng- postulate a role for significant neutrophil-induced endo-thelial injury in this disease.land Medical Center): What is the normal function of Gb3

receptors in the body? The density of these receptors is Dr. King: I suspect that the movement of PMNs intothe lamina propria generates a local inflammatory re-high in the kidneys. What is the density of these receptors

in the central nervous system, and does it correlate with sponse disrupting barrier function of the intestinal walland leading to movement of toxin into the circulationthe relative absence of neurologic involvement in diar-

rhea-associated HUS? and thus, exposure of the circulating cells. I think it isvery unlikely that the cells move in and out of the laminaDr. King: The natural role of Gb3 in cell function is

unknown. The richest source of Gb3 is the cortical tubu- propria. I think the cells are firmly attached to variousadhesion molecules.lar cell. However, the renal vasculature and mesangial

cells also express the glycolipid. The brain is another Regarding the small number of PMN cells in the glo-meruli found in autopsy specimens, there are severalabundant source of Gb3, but I suspect that the kidney

is more vulnerable to injury for several reasons. The potential explanations. First, it is possible that unboundor “free” toxin in the plasma causes direct injury to theglomerular microvasculature is distinctive because of its

high flow rates, hydraulic permeability, and pressure. glomerular cell and the tubular cell, that is, the PMNplays no role in injury. Second, the PMN might functionHigh blood flow rates directly increase local shear stress,

and ultrafiltration increases shear stress by increasing as a carrier for toxin with transfer from the lower affinitybinding site on the PMN to the higher affinity Gb3 onviscosity within the glomerular capillary. Higher shear

stress tends to increase platelet binding and activation. the target cell. This possibility does not implicate thePMN in the tissue injury. The third possibility is that aIt’s a lot harder to quantitate neurologic injury than renal

injury. There’s no obvious quantifiable marker, such as binding event occurs, toxin is delivered to the target,


and the PMN contributes to cell injury either by release mains relatively uncommon, and thus a child might neverbe exposed again. Children with E. coli also generateof reactive oxygen species, proteases, or phospholipids.

When you analyze autopsy specimens, by definition the antibodies to specific components of the bacterial lipo-polysaccharide, which might afford protection from fu-children have died of the disease. The timing of death

is usually about one week into the hospitalization and ture infection. It is clear that there are patients withrecurrent TTP or HUS, the pathogenesis of which hasthus the glomerular disease might have passed through

several phases. been established for some of these patients, for example,those with defective vWF cleaving protease activity orDr. Balakrishnan: Most of the pathologic data indi-

cate endothelial swelling and separation from the base- Factor H deficiency. However, I am unaware of dataconfirming whether children with E. coli-induced HUSment membrane. Most of the data that we have seen

describe apoptosis, a very controlled form of cell death. are protected from subsequent exposures.Dr. Harrington: Has the disease decreased in inci-Dr. King: There is only one report that Shiga toxins

induce apoptosis in glomerular endothelial cells [76]. dence since 1993? Did the new public health regulationsmake any difference?Whether apoptosis plays a dominant role in the glomeru-

lar injury is speculative. As you alluded to, the most Dr. King: The disease is increasing in frequency andhas been labeled as an emerging pathogen by the CDC.common histologic manifestation is endothelial swelling

and uplifting from the basement membrane, not the typi- Dr. Harrington: Why don’t we just cook the ham-burgers?cal findings of apoptosis. I suspect that this is a manifesta-

tion of direct toxicity of Stx on the glomerular endothe- Dr. King: There are no rare hamburgers in our house.Nothing pink!lial cell. It’s interesting to note in the studies by Forsyth

et al that PMNs obtained from children with HUS causedReprint requests to Dr. A. King, Division Head, Scripps Clinic, Divi-

disruption of the fibrillar architecture of reporter endo- sion of Nephrology, 10666 North Torrey Pines, N 239, La Jolla, Califor-nia, 92037, USA.thelial cells [57]. I think these cells are injured in waysE-mail: [email protected] than just apoptosis. Although apoptosis might not

be a dominant feature in glomerular injury, more evi-REFERENCESdence indicates that apoptosis is important in tubular

injury. These cells are exquisitely sensitive to toxin, at 1. Moschowitz E: Hyaline thrombosis of the terminal arterioles andcapillaries: A hitherto undescribed disease. Proc NY Pathol Socleast in part because of the high concentration of Gb3.29:21–24, 1924

Dr. Madias: How important is the attachment of E. 2. Gasser C, Gautier E, Steck A, et al: Hamolytisch-uramische Syn-drome: bilaterale Nierenrindennekrosen bei akuten erwobenencoli to intestinal epithelial cells for development of pa-hamolytischen anamien. Schweiz Med Wochenshr 85:905–909, 1955thology, and what do we know about the biology of this

3. Furlan M, Robles R, Galbusera M, et al: von Willebrand factor-process? cleaving protease in thrombotic thrombocytopenic purpura and

Dr. King: There are very specific proteins involved in the hemolytic uremic syndrome. N Engl J Med 339:1578–1584, 19984. Tsai HM, Lian EC: Antibodies to von Willebrand factor-cleavingintestinal attachment, which is essential for the disease

protease in acute thrombotic thrombocytopenic purpura. N Englto occur. I cannot quantify how important that is for the J Med 339:1585–1594, 1998disease. 5. Rock GA, Shumak KH, Buskard NA, et al: Comparison of plasma

exchange with plasma infusion in the treatment of thromboticDr. Klemens B. Meyer (Division of Nephrology, Newthrombocytopenic purpura. N Engl J Med 325:393–397, 1991England Medical Center): The Gb3 receptor is found on 6. Feng P: Escherichia coli serotype O157:H7 infection in humans.

the endothelial cells but also on mesangial and tubular Emerg Infect Dis 1:47–52, 19957. Siegler RL, Pavia AT, Christofferson RD, Milligan MK: Acells. Do we know what its normal function is? Is there

20-year population-based study of postdiarrheal hemolytic uremicany relationship to the receptor density and does that syndrome in Utah. Pediatrics 94:35–40, 1994have anything to do with susceptibility? 8. Koster F, Levin J, Walker L, et al: Hemolytic-uremic syndrome

after shigellosis. Relation to endotoxemia and circulating immuneDr. King: I am unaware of the role that Gb3 plays incomplexes. N Engl J Med 298(17):927–933, 1978the normal function of the cell. Susceptibility does ap- 9. Karmali MA, Petric M, Lim C, et al: The association between

pear to differ among individuals as only 5% of infected idiopathic hemolytic uremic syndrome and infection by verotoxin-producing Escherichia coli. J Infect Dis 151:775–782, 1985individuals acquire HUS. I am unaware of data relating

10. Acheson DWK, Donohue-Rolfe A, Keusch GT: The family ofreceptor density to disease risk. However, the fatty acidShiga and Shiga-like toxins, in Sourcebook of Bacterial Protein

chain length of Gb3 does vary among individuals. Thus, Toxins, edited by Alouf JE, Freer JH, London, Academic Press,1991, pp 415–433this variable might alter susceptibility to the disease.

11. Takeda T, Dohi S, Igarahsi T, et al: Impairment by verotoxin ofDr. Jeanine Carlson (Department of Medicine, Newtubular function contributes to the renal damage seen in haemolytic

England Medical Center): Why don’t people get HUS uraemic syndrome. J Infect 27:339–341, 199312. Paton AW, Paton JC: Detection and characterization of Shigatwice? The Gb3 and Shiga toxin are there; why doesn’t

toxigenic Escherichia coli by using multiplex PCR assays for stx1,this happen again? I’m sure that all these people don’tstx2, eaeA, enterohemorrhagic E. coli hlyA, rfb0111 and rfb0157.

become vegetarians. J Clin Microbiol 36:598–602, 199813. Newburg DS, Chaturvedi P, Lopez EL, et al: Susceptibility ofDr. King: First, this type of infection fortunately re-


hemolytic-uremic syndrome relates to erythrocyte glycosphingo- concentrations in hemolytic uremic syndrome patients and chil-dren with bloody diarrhea in Argentina. Pediatr Infect Dis J 14:594–lipid patterns. J Infect Dis 168:476–479, 1993

14. Endo Y, Tsurugi K, Yutsudo T: Site of action of a Verotoxin 598, 199537. Van de Kar NC, Sauerwein RW, Demacker PN, et al: Plasma(VT2) from Escherichia coli O157:H7 and of Shiga toxin on eukary-

otic ribosomes. Eur J Biochem 171:45–50, 1988 cytokine levels in hemolytic uremic syndrome. Nephron 71:309–313, 199515. Walters MD, Matthel U, Jay R, et al: The polymorphonuclear

count in childhood hemolytic uremic syndrome. Pediatr Nephrol 38. Martin AA, Woolven BL, Harris SJ, et al: Plasminogen activatorinhibitor type-1 and interleukin-6 in haemolytic uraemic syndrome.3:130–134, 1989

16. Coad NAG, Marshall T, Rowe B, Taylor CM: Changes in the J Paediatr Child Health 36:327–331, 200039. Tesh VL, Ramegowda B, Samuel JE: Purified Shiga-toxins inducepostenteropathic form of the hemolytic uremic syndrome in chil-

dren. Clin Nephrol 35:10–16, 1991 expression of proinflammatory cytokines from murine peritonealmacrophages. Infect Immun 62:5085–5094, 199417. Milford DV, Taylor CM, Guttridge B, et al: Haemolytic uremic

syndromes in the British Isles, 1985–8: Association with Verocyto- 40. Van Setten PA, Monnens LA, Verstraten RG, et al: Effectsof verocytotoxin-1 on nonadherent human monocytes: Bindingtoxin-producing Escherichia coli. Part 1: clinical and epidemiologi-

cal aspects. Arch Dis Child 65:716–721, 1990 characteristics, protein synthesis, and induction of cytokine release.Blood 88:174–183, 199718. Fitzpatrick MM, Shah V, Trompeter RS, et al: Interleukin-8 and

polymorphonuclear leukocyte activation in hemolytic uremic syn- 41. Ramegowda B, Tesh V: Differentiation-associated toxin receptormodulation, cytokine production and sensitivity to Shiga-like tox-drome. Kidney Int 42:951–956, 1992

19. Ikeda K, Ida O, Kimoto K, et al: Predictors for the development ins in human monocytes and monocytic cell lines. Infect Immun64:1173–1180, 1996of haemolytic uraemic syndrome with Escherichia coli O157:H7

infections: With focus on the day of illness. Epidemiol Infect 124: 42. Nash S, Stafford J, Madara JL: Effects of polymorphonuclearleukocyte transmigration on the barrier function of the cultured343–349, 2000

20. Slutsker L, Ries AA, Greene KD, et al: Escherichia coli O157:H7 intestinal epithelial monolayers. J Clin Invest 80:1104–1113, 198743. Parsons PE, Sugahara K, Cott GR, et al: The effect of neutrophildiarrhea in the United States: clinical and epidemiologic factors.

Ann Intern Med 126:505–513, 1997 migration and prolonged neutrophil contact on epithelial perme-ability. Am J Pathol 129:302–312, 198721. Remis R, MacDonald KL, Riley LW, et al: Sporadic cases of

hemorrhagic colitis associated with Escherichia coli O157:H7. Ann 44. Hurley BP, Thorpe C, Acheson DWK: Neutrophil translocationacross intestinal epithelial cells is enhanced by neutrophil transmi-Intern Med 101:624–626, 1984

22. Kelly J, Oryshak A, Wenetsek M: The colonic pathological vari- gration. Infect Immun 69:6148–6155, 200145. Te Loo DM, Hinsbergh VW, Heuvel LP, Monnens LA: Detec-ability of infection by O157:H7 infection. Am J Surg Pathol 14:87–

92, 1990 tion of verocytotoxin bound to circulating polymorphonuclear leu-kocytes of patients with hemolytic uremic syndrome. J Am Soc23. Hunt CM, Harvey JA, Youngs ER, Reid TM: Clinical and patho-

logical variability of infection by enterohaemorrhagic (Verocyto- Nephrol 12:800–806, 200146. Chart H, Scotland SM, Smith HR, Rowe B: Antibodies to Esche-toxin producing) Escherichia coli. J Clin Pathol 42:847–852, 1989

24. Zhe L, Bell C, Buret A, et al: The effect of enterohemorrhagic richia coli O157 in patients with haemorrhagic colitis and haemo-lytic uraemic syndrome. J Clin Pathol 42:973–976, 1989Escherichia coli O157:H7 on intestinal structure and solute trans-

port in rabbits. Gastroenterology 104:467–474, 1993 47. Bitzan M, Moebius E, Ludwig K, et al: High incidence of serumantibodies to Escherichia coli O157 lipopolysaccharide in children25. Elliott E, Li Z, Bell C, et al: Modulation of host response to

Escherichia coli O157:H7 infection by anti-CD18 antibody in rab- with hemolytic-uremic syndrome. J Pediatr 119:380–385, 199148. Caprioli A, Luzzi I, Rosmini F, et al: Hemolytic-uremic syndromebits. Gastroenterology 106:1554–1561, 1994

26. Hurley BP, Jacewicz M, Thorpe CM, et al: Shiga toxins 1 and 2 and Vero cytotoxin-producing Escherichia coli infection in Italy.The HUS Italian Study Group. J Infect Dis 166:154–158, 1992translocate differently across polarized intestinal epithelial cells.

Infect Immun 67:6670–6677, 1999 49. Pereira BJG, Shapiro L, King AJ, et al: Plasma levels of IL-1�and TNF� and their specific inhibitors in undialyzed chronic re-27. Thorpe CM, Hurley BP, Lincicome LL, et al: Shiga toxins stimu-

late secretion of interleukin-8 from intestinal epithelial cells. Infect nal failure, CAPD and hemodialysis patients. Kidney Int 45:8490–8896, 1994Immun 67:5985–5993, 1999

28. Raqib R, Wretlind B, Anderson J, Lindberg AA: Cytokine 50. Te Loo DM, Monnens LA, van Der Velden TJ, et al: Bindingand transfer of verocytotoxin by polymorphonuclear leukocytessecretion in acute shigellosis is correlated to disease activity and di-

rected more to stool than to plasma. J Infect Dis 171:376–384, 1995 in hemolytic uremic syndrome. Blood 95:3396–3402, 200051. Hughes DA, Smith GC, Davidson JE, et al: The neutrophil oxida-29. Westerholt S, Hartung T, Tollens M, et al: Inflammatory and

immunological parameters in children with haemolytic uremic syn- tive burst in diarrhoea-associated haemolytic uraemic syndrome.Pediatr Nephrol 10:445–447, 1996drome (HUS) and gastroenteritis–pathophysiological and diagnos-

tic clues. Cytokine 12:822–827, 2000 52. Brown RE, Alade SL, Knight JA, Evans BJ: Serum lipoperoxida-tion products in an infant with hemolytic uremic syndrome. Clin30. Yamamoto T, Nagayama K, Satomura K, et al: Increased serum

IL-10 and endothelin levels in hemolytic uremic syndrome caused Chem 34:2382–2384, 198853. Turi S, Nemeth I, Vargha I, Matkovics B: Oxidative damage ofby Escherichia coli O157. Nephron 84:326–332, 2000

31. Litalien C, Proulx F, Mariscalco MM, et al: Circulating inflam- red blood cells in haemolytic-uraemic syndrome. Pediatr Nephrol8:26–29, 1994matory cytokine levels in hemolytic uremic syndrome. Pediatr Ne-

phrol 13:840–845, 1999 54. Noris M, Ruggenenti P, Todeschini M, et al: Increased nitric oxideformation in recurrent thrombotic microangiopathies: A possible32. Proulx F, Turgeon JP, Litalien C, et al: Inflammatory mediators

in Escherichia coli O157:H7 hemorrhagic colitis and hemolytic- mediator of microvascular injury. Am J Kidney Dis 27:790–796, 199655. Facorro G, Aguirre F, Florentin L, et al: Oxidative stress anduremic syndrome. Pediatr Infect Dis J 17:899–904, 1998

33. Van Setten PA, van Hinsbergh VW, van den Heuvel LP, et al: membrane fluidity in erythrocytes from patients with hemolyticuremic syndrome. Acta Physiol Pharmacol Ther Latinoam 47:137–Monocyte chemoattractant protein-1 and interleukin-8 levels in

urine and serum of patients with hemolytic uremic syndrome. Pedi- 146, 199756. King AJ, Sundaram S, Cendoroglo M, et al: Shiga toxin inducesatr Res 43:759–767, 1998

34. Karpman D, Andreasson A, Thysell H, et al: Cytokines in child- superoxide production in polymorphonuclear cells with subsequentimpairment of phagocytosis and responsiveness to phorbol esters.hood hemolytic uremic syndrome and thrombotic thrombocytope-

nic purpura. Pediatr Nephrol 9:694–699, 1995 J Infect Dis 179:503–507, 199957. Forsyth SD, Simpleson AC, Fitzpatrick MM, et al: Neutrophil-35. Inward CD, Varagunam M, Adu D, et al: Cytokines in haemolytic

uraemic syndrome associated with verocytotoxin-producing Esche- mediated endothelial injury in haemolytic uraemic syndrome. Lan-cet 2:411–414, 1989richia coli infection. Arch Dis Child 77:145–147, 1997

36. Lopez EL, Contrini MM, Devotos S, et al: Tumor necrosis factor 58. Morigi M, Micheletti G, Figliuzzi M: Verotoxin-1 promotes


leukocyte adhesion to cultured endothelial cells under physiologic of cytokine production by human proximal tubule cells. KidneyInt 54:1093–1106, 1998flow conditions. Blood 86:4553–4558, 1995

68. Simon M, Cleary TG, Hernandez JD, Abboud HE: Shiga toxin59. Nevard CH, Blann AD, Jurd KM, et al: Markers of endothelial1 elicits diverse biologic responses in mesangial cells. Kidney Intcell activation and injury in childhood haemolytic uraemic syn-54:1117–1127, 1998drome. Pediatr Nephrol 13:487–492, 1999

69. Kodama T, Nagayama K, Yamada K, et al: Induction of apoptosis60. Inward CD, Pall AA, Adu D, et al: Soluble circulating cell adhe-in human renal proximal epithelial cells by Escherichia coli verocy-sion molecules in haemolytic uraemic syndrome. Pediatr Nephroltotoxin 1 in vitro. Med Microbiol Immunol (Berl) 188:73–78, 19999:574–578, 1995 70. Karpman D, Hakansson A, Perez MT, et al: Apoptosis of renal

61. Inward CD, Howie AJ, Fitzpatrick MM, et al: Renal histopathol- cortical cells in the hemolytic-uremic syndrome: In vivo and inogy in fatal cases of diarrhoea-associated haemolytic uraemic syn- vitro studies. Infect Immun 66:636–644, 1998drome. Pediatr Nephrol 11:556–559, 1997 71. Taguchi T, Uchida H, Kiyokawa N, et al: Verotoxins induce

62. Hertzke DM, Cowan LA, Schoning P, Fenwick BW: Glomerular apoptosis in human renal tubular epithelium derived cells. Kidneyultrastructural lesions of idiopathic cutaneous and renal glomerular Int 53:1681–1688, 1998vasculopathy of greyhounds. Vet Pathol 32:451–459, 1995 72. Van Setten PA, van Hinsbergh VWM, van der Velden TJAN, et

63. Taylor CM, Williams JM, Lote CJ, et al: A laboratory model of al: Effects of TNF� on verocytotoxin in purified human glomerularmicrovascular endothelial cells. Kidney Int 51:1245–1256, 1997toxin-induced hemolytic uremic syndrome. Kidney Int 55:1367–

73. Harel Y, Silva M, Giroir B, et al: A reporter transgene indicates1374, 1999renal-specific induction of tumor necrosis factor (TNF) by Shiga-64. Obrig TG, Louise CB, Lingwood CA, et al: Endothelial heteroge-like toxin. Possible involvement of TNF in hemolytic uremic syn-neity in Shiga toxin receptors and responses. J Biol Chem 268:drome. J Clin Invest 92:2110–2116, 199315484–15488, 1993

74. Isogai E, Isogai H, Kimura K, et al: Role of tumor necrosis factor65. Pijpers AH, van Setten PA, van den Heuvel LP, et al: Verocyto-alpha in gnotobiotic mice infested with an Escherichia coli O157:H7toxin-induced apoptosis of human microvascular endothelial cells. strain. Infect Immun 66:197–202, 1998

J Am Soc Nephrol 12:767–778, 2001 75. Lingwood CA: Role of verotoxin receptors in pathogenesis.66. Hughes AK, Stricklett PK, Schmid D, Kohan DE: Cytotoxic Trends Microbiol 4:147–153, 1996

effect of Shiga toxin-1 on human glomerular epithelial cells. Kidney 76. Pijpers AH, van Setten PA, van den Heuvel LP, et al: Verocyto-Int 57:2350–2359, 2000 toxin-induced apoptosis of human microvascular endothelial cells.

J Am Soc Nephrol 12:767–778, 200167. Hughes AK, Stricklett PK, Kohan DE: Shiga toxin-1 regulation

Documents

Acute inflammation in the pathogenesis of hemolytic-uremic syndrome