Verotoxin (Shiga Toxin) Sensitizes Renal Epithelial Cells to Increased Heme Toxicity: Possible Implications for the Hemolytic Uremic Syndrome
Martin Bitzan,
Brandi B. Bickford and
Gregory H. Foster
Department of Pediatrics, Wake Forest University School of Medicine, Winston-Salem, North Carolina
Correspondence to Dr. Martin M. Bitzan, Pediatric Nephrology, Montreal Children’s Hospital, 2300 Rue Tupper, Room E-222, Montreal, Quebec, H3H 1P3, Canada. Phone: 514-412-4461; Fax: 514-412-4359; E-mail: martin. bitzan{at}muhc.mcgill.ca
Escherichia coli–derived verotoxins (VT; Shiga toxins)are causally related to the pathogenesis of enteropathic hemolyticuremic syndrome (HUS). Profound hemolysis is a defining featureof the disease, but it is not known whether the acute intravascularrelease of heme proteins contributes to HUS pathology. Thisstudy examined the biologic effects of hemin and VT by meansof tubular epithelial-derived ACHN cells. Hemin at concentrations200 µM caused cell rounding, spike formation, and detachmentthat was morphologically distinct from verocytotoxicity. VTcaused apoptosis at concentrations >100 pM, as demonstratedby nuclear segmentation and poly(ADP-ribose) polymerase cleavage,whereas hemin-mediated injury of ACHN cells grown in serum-containingmedium lacked attributes of programmed cell death. Pretreatmentof ACHN monolayers with sublethal concentrations (1 to 10 pM)of VT for 12 to 18 h led to superadditive hemin-mediated cytotoxicity.This effect was not limited to ACHN cells, but was similarlynoted in microvascular endothelial cells. Heme catabolism isregulated by (inducible) heme oxygenase-1 (HO-1). VT abrogatedHO-1 expression in ACHN cells. Stimulation of cells for 6 hwith CdCl2, which markedly increased HO-1 expression beforethe addition of VT, blunted subsequent hemin injury. In conclusion,VT augments hemin-induced toxicity in renal tubular epithelialcells that can be reversed by prior induction of HO-1. It isproposed that VT subverts the physiologic defense against hemeproteins by interfering with the regulated expression of HO-1and that this mechanism contributes to the renal pathology inpatients with Escherichia coli–associated HUS.
Infections by the Escherichia coli O157:H7 serotype typicallypresent as hemorrhagic colitis. About 15% of children with Escherichiacoli O157 colitis develop renal and hematologic complications(1), known as enteropathic (or Escherichia coli–mediated)hemolytic uremic syndrome (eHUS). Escherichia coli–mediatedHUS comprises >80% of HUS in children and is a leading causeof acute renal failure in this age group (2). The acute lethalityis 1% to 5%, often due neurologic or cerebrovascular complications(2,3). About 40% of children with eHUS require dialysis (4),and >10% develop severe long-term complications, includingchronic renal failure, hypertension, or diabetes mellitus (2,3).However, eHUS is not limited to children: several outbreaksin nursing homes and the community have identified the elderlypopulation as another vulnerable group, with a lethality of>10% (5).
Evidence strongly implicates Escherichia coli–derivedverotoxins (Shiga toxins) 1 and 2 in the pathogenesis of HUS.Both toxins belong to a family of structurally and functionallyclosely related bipartite protein exotoxins, bind with highaffinity and specificity to the membrane-anchored glycosphingolipidglobotriaosylceramide (Gb3) of mammalian cells and elicit virtuallyidentical cell biologic effects. After intracellular translocation,the A subunit hydrolyzes a specific adenine residue of the 28Sribosomal RNA, which blocks the coordinated function of eukaryoticelongation factors EF-1 and -2 and de novo peptide synthesis(6,7).
Although HUS is commonly viewed as a disease of the vascularendothelium, evidence is emerging that renal injury in eHUSis not limited to the glomerular vessels, but also affects othercompartments such as renal tubules (8,9). Cultured mammaliancells differ markedly in their susceptibility to verotoxin (VT)cytotoxicity (6). For example, in renal tubular epithelial (10)and microvascular endothelial cells (11,12), VT induces programmedcell death, whereas other tissues exhibit toxin-mediated cellactivation (13,14). Ex vivo studies on renal biopsy and autopsysamples from children with acute Escherichia coli–inducedHUS revealed evidence of apoptosis, predominantly in tubules(9,15,16) as well as necrotic injury, in both renal parenchymaand the gut (17). However, whether the observed tissue injuryis due to the direct effect of VT or the consequence of thrombosisand ischemia remains to be elucidated.
A hallmark of the disease is nonimmune, intravascular hemolysis,known as microangiopathic hemolytic anemia, which often requirespacked red blood cell transfusions. The mechanism and biologiceffects of hemolysis in HUS are not well defined. Acute hemolysisgenerally leads to plasma haptoglobin depletion and accumulationof free hemoglobin. The stable hemoglobin-haptoglobin complexis rapidly cleared by mononuclear phagocytes endowed with ahigh affinity CD163 scavenger receptor (18). Unbound hemoglobinis filtered in the renal glomerulus as dimers, and reabsorbedand catabolized in proximal tubule cells. Spontaneous oxidationto met[Fe3+]hemoglobin destabilizes extraerythrocytic hemoglobin,which results in the release of the heme moiety (19–21).Oxidized (ferric) protoporphyrin IX (hemin) is commonly isolatedas a halide (hemin chloride) (18). It is a highly hydrophobicmolecule capable of inducing adverse biologic effects. It intercalateswith the membrane lipid bilayer and potentiates oxidant-mediatedinjury (20,22–24). The exact mechanism or mechanisms ofhemin toxicity and the predominant pathway of cell death remainsto be defined.
To protect the integrity of the surrounding tissue, excess freeheme is degraded, largely by two (microsomal) heme oxygenase(HO) isoforms, HO-1 (inducible), and HO-2 (constitutive). HOcatalyzes the initial and rate-limiting step in heme catabolism,the oxidative cleavage of the protoporphyrin ring by three sequentialmonooxygenase reactions (25), thereby generating several biologicallyimportant products: ferrous iron, which stimulates the inductionof ferritin, a high capacity intracellular iron store; biliverdin,the precursor of the (antioxidant) bilirubin; and the gas carbonmonoxide (CO) (26). CO is increasingly recognized as an importantmessenger molecule with vasorelaxant, antithrombotic, antiinflammatoryand antiapoptotic properties (24,26,27). Potent inducers ofHO-1, both in vitro and in vivo, are heavy metals, oxidativestress, and its substrate, heme (28,29).
Acute kidney failure due to massive release of hemoglobin ormyoglobin, as in blood transfusion accidents and rhabdomyolysis,demonstrates the potential of heme proteins to cause renal injury(29–31). We reasoned that because of the profound hemolysisobserved in patients with HUS, the kidneys would be burdenedwith large amounts of hemoglobin and heme. This study teststhe hypothesis that VT interferes with the regulated responseto, and compromises the cellular response against, excess hemein renal tubular epithelial cells. We further examined if inductionof the heme-degrading enzyme, HO-1 modifies the cell responseto hemin or VT.
Materials
Rabbit anti–heme oxygenase-1 specific for the N-terminalportion of human HO-1 was purchased from StressGen (Vancouver,BC, Canada); polyclonal anti-poly(ADP-ribose) polymerase (PARP)antibody were from Cell Signaling Technology (Beverly, MA);monoclonal anti–-actin (N-terminus) antibody (mouse IgG1),goat anti-rabbit and rabbit anti-mouse horseradish peroxidase–conjugatedantibodies were purchased from Sigma Chemicals (St. Louis, MO).Hemin (Fe[III] protoporphyrin IX chloride) was from PorphyrinProducts (Logan, UT). Human recombinant TNF- was from ICN Biomedicals(Irvine, CA). Cadmium chloride hemi(pentahydrate), cycloheximide,and DAPI (4',6-diamidino-2'-phenylindole dihydrochloride) andstaurosporine (S 5921) were from Sigma. Crystalline protease-and Ig-free BSA was from Jackson Immuno Research Laboratories(West Grove, PA). All other chemical reagents were of analyticalgrade and were purchased from Fisher, unless indicated otherwise.Cell culture media and reagents were from Life Technologies/LifeTechnologies BRL (Grand Island, NY), unless indicated otherwise.FBS was from Cellgro Mediatech (Hendon, VA). Cell culture disheswere from TTP (The Technology Partnership, Royston, UK) andfrom Corning (Corning, NY). Highly purified verotoxins 1 and2 were gifts from Dr. M. A. Karmali (Guelph, Ontario). Bothtoxins bind to the same globotriaosylceramide (Gb3) receptorand demonstrate similar biologic effects and cytotoxic activitiesby means of standardized Vero cell assays (6,12,13)
Cell Culture
Human renal carcinoma–derived tubular epithelial (ACHN)cells were obtained from the American Tissue Culture Collection(ATCC CRL-1611) and propagated in MEM with Earle salts and L-glutamine,supplemented with penicillin-streptomycin and 10% FBS. Cellswere serially passaged every 5 to 7 d on 100-mm dishes with0.05% trypsin-EDTA. To assess the effects of agents on ACHNcells, 100 mm dishes were seeded at a density of approximately2 x 105 cells and 96-well dishes at approximately 3 x 103 cells/welland grown to confluence. Primary, pooled dermal human microvascularendothelial cells (HMVEC) were obtained at passage 3 from Clonetics/BioWhittacker(Walkersville, MD) and propagated (50 x 104 cells per 35-mmdish) in endothelial growth medium containing a mixture of growthfactors, hydrocortisone, 5% FBS, heparin, and gentamicin-amphotericin(EGM2-MV, Clonetics). Cells were detached with 0.025% trypsin-EDTAand used through passages 5 to 7. ACHN (8,10,32) and HMVEC (11,12,33)have been previously shown to express Gb3 on their surface andto be sensitive to VT-induced cytotoxicity.
VT was added after dilution in tissue culture medium. Heminwas dissolved in 0.1 M NaOH to a concentration of 20 mM andfurther diluted in tissue culture medium under subdued lightand immediately added to the cell culture. CdCl2 was dissolvedin H2O to a concentration of 10 mM and further diluted in tissueculture medium. Because CdCl2 caused ACHN cell toxicity at concentrations100 µM, experiments were performed with 50 µM. Mediumwas replenished 6 h after the addition of CdCl2 to further minimizeits cytotoxic effect. All other agents remained in the dishthroughout the experiment. Morphologic changes of (live) cellcultures were assessed with an inverted phase microscope.
Cytotoxicity Assays
Monolayer disruption was quantitated spectrophotometricallyas described previously (12). Briefly, monolayers were fixedwith 2% formalin in PBS and stained with 0.13% crystal violet,rinsed thoroughly, and dried. Cell-associated dye was elutedwith 50% ethanol in water and monitored at 490 nm with an automatedplate reader. The optical density directly correlates with thenumber of residual cells. The 50% cytotoxic dose (CD50) wascalculated as described previously (12). When appropriate, A490values were normalized to vehicle-treated controls.
Apoptosis Assays
Cell monolayers were grown in tissue culture dishes, treatedwith agents, and stained with DAPI (34). Briefly, nonadherent(floating) cells were collected and washed in ice-cold PBS at200 x g and 4°C for 5 min, settled for 60 min on 12-mm glasscover slips previously coated with poly-L-lysine (1 mg/ml PBS)(Fisher) and fixed with methanol for 1 min, followed by 0.1µg/ml DAPI in methanol (2 min), rinsed sequentially inmethanol and PBS, and mounted on glass slides with 2 µlVectashield mounting medium (Vector Laboratories, Burlingame,CA). Adherent (nondetached) cells were fixed and stained insitu. Cell preparations were viewed under incident ultraviolet(UV) light with a Zeiss Axiovert 10 microscope equipped witha filter for UV excitation (DAPI) and digital imaging fluorescencemicroscopy system (Zeiss Axioplan 2). Apoptotic cells were identifiedby segmented morphology of DAPI-stained cell nuclei. At leastfive different fields per slide were evaluated.
Protein Extraction and Western Blot Analysis
Nonadherent cells were collected on ice. Adherent cells weredetached by gentle scraping in ice-cold PBS, pooled with thenonadherent cells, pelleted, and lysed with modified Laemmlibuffer in the presence of protease inhibitors (Protease InhibitorCocktail Set II; Calbiochem, San Diego CA). Cell lysates werecleared by centrifugation at 14,000 x g for 5 min at 4°Cand stored at –70°C. Protein concentrations were measuredwith a modified Lowry (Bio-Rad detergent compatible) proteinassay (Bio-Rad, Hercules, CA) with BSA as standard. Proteinextracts were resolved by SDS-PAGE (12%) under reducing (0.02M DTT) and denaturing conditions and transferred onto a polyvinyldifluoridemembrane (Immun-Blot; Bio-Rad) by electroblotting. Membraneswere blocked for 2 h in 50 mM Tris-buffered saline (TBS) containing5% BSA and 0.1% Triton X-100 (TBS-T-BSA5%), followed by incubationwith a polyclonal anti-human HO-1 antibody (1:2000 in TBS-BSA1%)for 1 h at room temperature, or anti-PARP antibody (1:8000 inTBS-BSA4%) overnight at 4°C. Antigen-antibody complexeswere detected with horseradish peroxidase–conjugated secondarygoat anti-rabbit antibody for 1 h at 25°C with luminol/peroxide-basedchemiluminescence (SuperSignal West Pico; Pierce, Rockford IL)and brief exposure to autoradiography films. To ascertain equalloading, blots were stripped with 0.2 M glycine (pH 2.3) for20 min at 25°C and reprobed with anti–-actin 1:5000in TBS-BSA1% in for 1 h. Alternatively, gels were stained withGelCode Blue (Pierce) after transfer.
Statistical Analyses
Data are expressed as mean ± SD. Experiments were performedat least three times, unless indicated otherwise. In experimentsinvolving more than one potentially cytotoxic agent, a synergisticeffect was assumed when the combined effect of the two agentswas greater then the sum of both agents used alone. ANOVA wasapplied when comparisons involved multiple samples and treatments.Differences between means were determined by the two-samplet test. The level of significance was defined as P < 0.05.
Hemin- and Verotoxin Cytotoxicity of Tubular Epithelial Cells
To determine the ability of heme proteins to cause renal tubularinjury, we used human kidney-derived ACHN cells and Fe[III]protoporphyrinIX chloride (hemin) as a model system. In the first series ofexperiments, confluent cell monolayers were exposed to varioushemin concentrations over 6 to 48 h. At concentrations >100µM, hemin treatment caused a distinct cytopathic effect,consisting of cell rounding and the appearance of spindle-formextensions or cell bridges not observed in vehicle-treated controls,and subsequent cell detachment (Figure 1A). Hemin-mediated changesin cell morphology and monolayer disruption were time and concentrationdependent. At concentrations 200 µM, morphologic changeswere observed as early as 6 h after the addition of hemin (Figure 1, A and B).In contrast, VT-induced cytotoxicity became apparentonly over 24 to 72 h (Figure 1, A and C). Morphologic changesinduced by VT1 and VT2 were identical and clearly distinguishablefrom changes induced by hemin (Figure 1A). CD50 was approximately200 µM for hemin at 24 to 48 h (Figure 1B). The CD50 forVT1 and VT2 was approximately 0.1 nM and 0.5 nM, respectively,at 48 to 72 h (Figure 1C).
Figure 1. Hemin- and verotoxin (VT)-induced cytotoxicity of ACHN cells. (A) Cell monolayers were treated with vehicle, hemin (100 or 200 µM), or VT1 for the indicated time intervals. Images were acquired by phase-contrast microscopy (original magnification, x1:10). (B and C) Monolayers were treated with hemin (B) or VT1 (C) as indicated. Residual cells were stained with crystal violet and monitored by spectrophotometry as described in Materials and Methods. Data points are means and SD of triplicate wells. Where error bars are not present, the SD was too small to appear. Figures are representative of at least three independent experiments.
Experiments were designed to further characterize the cytopathiceffect induced by hemin and VT in ACHN cells. Cell monolayerswere treated with 100 or 200 µM hemin for up to 30 h,or with 0.1 to 5 nM VT2 for 48 h, respectively. Replicate monolayerswere treated with TNF- (10 ng/ml) and cycloheximide (10 µg/ml)for control purposes. Cells were stained with DAPI and monitoredby microscopy for the presence of morphologic changes via phasecontrast and incident UV light. Nonadherent (floating) cellsand adherent cells were evaluated separately. Apoptosis, definedby nuclear segmentation and visualized by DAPI staining, wasevident in approximately 30% (range, 10% to 50%) of the detachedcells from VT1 and VT2-treated dishes (Figure 2A). Of the adherentcells <1% were apoptotic. In contrast, apoptotic cells wereconsistently absent among vehicle- and hemin-treated monolayers,both in preparations from detached and residual (adherent) cells(Figure 2A). In another experiment, ACHN cell monolayers weretreated with vehicle, hemin (100 to 200 µM; 24 h), orVT2 (1 nM; 72 h) and monitored for PARP cleavage by Westernblot test. As shown in Figure 2B, the large 89 kD fragment ofthe intact 116 kD PARP protein was noted in the extracts fromVT-treated cells, whereas PARP cleavage was consistently lackingin hemin- and vehicle-treated cells.
Figure 2. Hemin- and verotoxin (VT)-induced cell death. (A) ACHN cell monolayers were treated with vehicle (NaOH), hemin (100 µM, for 16 and 28 h), VT2 (5 nM, for 48 h), or TNF- (10 ng/ml) plus cycloheximide (10 µg/ml) (1 h), as apoptosis control. Floating cells were collected, fixed, and stained with DAPI as described in Materials and Methods. The number of floating cells from the control monolayers was too low to be evaluated, and the residual cells were fixed and stained in situ. Arrows point at examples of apoptotic cells, defined by nuclear segmentation in the DAPI-stained preparations. (B) ACHN monolayers were treated with vehicle (Veh), VT2 (1 nM, 3 d), or hemin (200 µM, 24 h). Total cellular protein extracts were subjected to Western blot analysis with anti-poly(ADP-ribose) polymerase (PARP) antibody. Gels were stained for proteins and photographed for transfer as loading control.
Combined Effects of VT and Hemin on ACHN Cells
VT is central to the development of human HUS and VT-producingEscherichia coli-induced systemic disease in experimental animals(35,36). We therefore investigated the combined effect of VTand hemin on tubular epithelial cells and asked whether VT altersthe susceptibility of these cells to hemin toxicity. ACHN cellmonolayers were treated for various intervals with VT beforehemin was added. In a typical experiment, depicted in Figure 3,pretreatment of cells with VT 18 h before the addition ofhemin caused increased monolayer disruption and cell loss, whereastreatment with hemin or VT alone had only minor effects on monolayerintegrity. Comparable effects were observed when replicate monolayerswere pretreated with cycloheximide instead of VT (Figure 3A).In a similar experiment, cells were grown in 96-well platesand residual monolayers stained with crystal violet for quantitativespectrophotometry. Results, depicted in Figure 3B, showed thatpreincubation of cell monolayers with VT increased the cytopathiceffect of hemin. Further analysis of the A490 values revealedthat sequentially added VT and hemin exerted, at least in part,supperadditive or synergistic cytotoxic effects (Figure 3B).
Figure 3. Combined effects of verotoxin and hemin. (A) ACHN cell monolayers were treated with vehicle or VT1 10 pM for 18 h, followed by the addition of 100 or 200 µM hemin (or vehicle) for another 12 h. Additional dishes were treated with cycloheximide (CHX) 10 µg/ml for 30 min instead of VT. Monolayers were photographed in situ with an inverted Zeiss microscope with a x10 objective. (B) ACHN monolayers were treated with diluent or VT1 (1 or 10 pM) for 18 h in 96-well dishes, followed by vehicle or hemin (100 or 200 µM) for another 6 or 48 h. Residual cells were quantitated by crystal violet staining. Shown are means and SD of triplicate determinations. * P < 0.05; ** P < 0.01; *** P < 0.005. Results represent at least three similar experiments. (C) Human microvascular endothelial cell (HMVEC) monolayers were incubated with 0.1 pM VT2 (top), or sequentially with VT2 and hemin (50 µM) after 18 h (bottom). Hemin failed to induce microscopically discernible monolayer changes at the used concentration when added alone (data not shown). Images were acquired via time-lapse video microscopy. The number of hours elapsed from the addition of VT and of hemin is indicated. For comparison, the cytotoxic effect of a 100-fold higher concentration of VT2 on HMVEC is shown on the right.
In separate experiments, endothelial cell monolayers (HMVEC)were exposed to VT and/or hemin and monitored by time-lapsevideo microscopy. Sequential treatment of HMVEC with VT2 andhemin at concentrations that failed to induce microscopicallydiscernible cytotoxicity when added alone resulted in increasedcell injury (Figure 3C).
HO-1 Induction Modulates Hemin and VT Toxicity
HO-1 induction is critical for the effective degradation ofexcess heme. We therefore tested to see whether the stimulationof ACHN cells with an inducer of HO-1, such as CdCl2, (26,37),will enhance HO-1 expression and enzymatic activity and protectthese cells from hemin-mediated toxicity. CdCl2 increased HO-1expression in a time and concentration-dependent manner up toabout 50 µM. HO-1 expression diminished with greater CdCl2concentrations, possibly as a result of increasing toxicityof the compound (Figure 4A). As shown in Figure 4B, CdCl2-inducedHO-1 protein was evident after 6 h of incubation. HO-1 levelspeaked after 9 h, were sustained for at least 12 h more, anddecreased to baseline >36 h after stimulation (Figure 5,lanes 5 to 7).
Figure 4. Induction of heme oxygenase-1 (HO-1) by cadmium chloride (CdCl2) in ACHN cells. Confluent monolayers were treated (A) for 6 h with the indicated CdCl2 concentrations, or (B) with 50 µM CdCl2 for the indicated time intervals. Total cellular extracts were resolved by SDS-PAGE, blotted, and developed with a polyclonal antibody to HO-1. Gels were stained after transfer to assess relative protein loading as described in Materials and Methods (A). Alternatively, blots were stripped and reprobed with a monoclonal antibody to -actin as loading control (B).
Figure 5. Verotoxin (VT) abrogates induction of HO-1 protein expression in ACHN cells. Monolayers were sequentially treated with vehicle (– CdCl2 (50 µM), VT2 (10 or 100 pM), or hemin (200 µM) for the indicated time intervals. Time 0 h corresponds to the addition of VT2. Total cellular protein was extracted at time X. Blots were developed with a polyclonal antibody to HO-1, stripped, and reprobed with a monoclonal antibody to -actin as loading control.
Because VT is known to block nascent peptide synthesis (6),we determined whether VT modifies expression levels HO-1 protein.When cell monolayers were pretreated with VT2 for 9 h and subsequentlystimulated with CdCl2, HO-1 expression was blocked (Figure 5,lanes 10 and 11). VT also diminished the induction of HO-1 expressionby hemin in a time- and concentration-dependent manner (Figure 5,lanes 1 to 4). Inhibition of hemin-induced HO-1 expressionwas complete when VT pretreatment was extended to 18 h (resultsnot shown). In contrast, when cadmium was used to stimulateHO-1 expression, elevated HO-1 protein levels were maintainedfor at least 9 h after the addition of VT (Figure 5, lanes 8and 9).
We then asked whether induction of HO-1 expression by CdCl2protects cells from the combined cytotoxicity of VT and hemin.On the basis of the HO-1 induction experiments described above(Figure 4 and 5), monolayers were incubated with CdCl2 for 6h, followed by VT2 and hemin. As shown in Figure 6A, CdCl2 treatmentresulted in blunting of hemin-induced cytotoxicity. CdCl2 failedto prevent the cytotoxic effect of VT alone (results not shown).However, the prior induction of HO-1 reversed the combined cytotoxiceffect of hemin and VT in ACHN cells. Results from a typicalexperiments that used various VT2 (10 to 1000 pM) concentrationsand hemin over 6 and 48 h are depicted in Figure 6, B and C.Taken together, results indicate that the induction of HO-1protects ACHN cells from augmented hemin toxicity in the presenceof VT.
Figure 6. Cadmium chloride blunts hemin toxicity in verotoxin (VT)-treated ACHN cells. (A) Monolayers were treated as indicated with vehicle, or VT2 1 nM for 24 h, or CdCl2 50 µM for 6 h, followed by 200 µM hemin for 24 h, or hemin 200 µM alone for 24 h. Monolayers were photographed with an inverted microscope, x10 objective. (B) Monolayers in 96-well format were sequentially treated with vehicle or 50 µM CdCl2 for 6 h, 0.001 nM VT2 (or diluent) for 9 h, and 200 µM hemin (or vehicle) for 24 h. (C) Cells were treated as under (B), except that the VT2 concentrations varied between 0.01 and 1 nM, and hemin treatment was prolonged to 48 h. Residual cells were quantitated by crystal violet staining. Shown are the means ± SD of triplicate wells, normalized for vehicle-treated controls. * P < 0.05; ** P < 0.01; *** P < 0.005. ¶ P < 0.05 compared with hemin, and P < 0.005 with VT2 alone (B). # P < 0.05 compared with VT2 alone at the corresponding concentrations (C). Data represent at least three independent experiments.
Acute intravascular hemolysis is a defining feature of hemolyticuremic syndromes. The etiology of the hemolytic process is unclear,and its contribution to the pathogenesis of HUS is not known.Here we show that hemoglobin-derived hemin imparts enhancedtoxicity toward human renal epithelial (and endothelial) cellsin the presence of verotoxin, the principal causative agentof the classic, enteropathic HUS.
Release of hemoglobin during hemolysis can exceed the rate ofhaptoglobin production in the liver resulting in rising plasmaconcentrations of free (unbound) hemoglobin (38). Free hemoglobinis in part filtered in the glomerulus and reabsorbed by proximalrenal tubule cells, where it accumulates in the apical regionsto be degraded (18,39). Extracellular hemin is released uponoxidation of (ferrous) hemoglobin to (ferric) methemoglobin(19,21,40). Similar to haptoglobin, the plasma glycoproteinhemopexin, which binds heme with high affinity, becomes saturatedand depleted during profound hemolysis (38) or (therapeutic)infusion of hemin (41). Heme-mediated toxicity has been coupledto oxidative and nonoxidative mechanisms. It causes peroxidationof various cellular components, including membrane lipids (40);activates neutrophils and induces monocyte chemoattractant protein-1(24,42); and potentiates cell killing by polymorphonuclear leukocytesand other sources of reactive oxygen intermediates (43).
Induction of the microsomal heme oxygenase system, e.g., byheavy metals, oxidative stress, hemoglobin, and heme (26,37),is an effective and widely conserved mechanism to limit hemeprotein-mediated tissue injury (18). Recent cDNA microarraystudies in human microvascular endothelial cells revealed thatHO-1 overexpression is associated with a decrease in mRNA levelsfor antiproliferative and proapoptotic genes (44). The physiologicand clinical importance of HO-1 has become strikingly evidentin the case of a child with hereditary HO-1 deficiency (45)and in HO-1 knockout mice (46). A prominent feature of HO-1deficiency is chronic oxidative stress and—at least inthe reported child—chronic (intravascular) hemolytic anemiawith profound schistocytosis (45,46). This experiment of naturealso suggests that inducible HO-1 is necessary to effectivelyremove heme (45). Although constitutively expressed HO-2 appearsto play a protective role in lung tissue and neurons, it isnot sufficient to metabolize excess heme (45,47,48).
Our results demonstrate enhanced renal tubule cell injury inthe presence of VT. This effect was mirrored by cycloheximide,an antibiotic with peptide synthesis inhibiting activity (Figure 3A).Moreover, VT also led to exacerbation of hemin-inducedinjury in microvascular endothelial cells (Figure 3C). AlthoughVT caused apoptosis of ACHN cells, in agreement with an earlierreport (10), the mechanism of heme-induced cytotoxicity is notwell understood. Cell death due to oxidative injury as wellas oxidative stress–independent hemin toxicity has beendescribed (21). Hemin produced peculiar morphologic changesin ACHN cells consisting of cell rounding and the appearanceof elongated cell processes, but failed to induce nuclear segmentationor PARP cleavage. We inferred that hemin causes nonapoptotic,likely necrotic death in ACHN cells in our model. Interestingly,after completion of our experiments, Gonzalez-Michaca et al.(48) reported that hemin induced apoptosis in immortalized ratproximal renal tubular cells under serum-deprived, but not serum-replete,culture conditions. Inhibiting HO-1 activity by zinc protoporphyrinfurther enhanced the proapoptotic effect of hemin in serum-deprivedcultures (48). These results are of great interest in the contextof the proposed hypothesis and warrant further scrutiny in anin vivo model of verotoxemia.
Although the HO system has been increasingly recognized as partof an adaptive cell response, specifically to oxidative stress,overexpression of HO-1 can also impart adverse effects becauseof increased release of redox-active iron (23). Observationsin patients and animal models of hemolytic anemia and sicklecell disease (31,38) suggested a priori that HUS causes substrate-inducedupregulation of HO-1 in response to excess hemoglobin. We foundthat VT repressed HO-1 expression in vitro in renal tubularepithelial (ACHN) (Figure 5) by translational inhibition. IfVT represses HO-1 induction in vivo, it is expected that thelack of HO-1 protein induction in the presence of hemoglobinuriaresults in increased renal parenchymal toxicity (29,30). Nathet al. reported that infusion of hemoglobin caused renal failureand death within 1 to 2 wk in HO-1 knockout mice, but not inheterozygote littermates (28). A day after hemoglobin administration,HO-1–deficient mice revealed extensive tubular epithelialcell necrosis. In contrast, HO-1–competent mice maintainednormal renal function and histologic appearance when challengedwith the same dose of hemoglobin (28).
We hypothesize that VT blocks HO-1 expression in patients withEscherichia coli–induced HUS and that lack of an adequateHO-1 response would increase tissue sensitivity to heme proteintoxicity. Our observation that treatment of ACHN cells withCdCl2, a potent inducer of HO-1, affords cell protection whenadded before treatment with VT and hemin, further adds to theplausibility of our hypothesis. We posit that VT causes a stateof vulnerability to heme protein-mediated injury when it isneeded most, i.e. during the acute, hemolytic stage of the HUS.On the basis of the growing understanding of the physiologyof the HO system (49), we further hypothesize that the repressionof HO-1 leads to diminished local production of the gaseoussignaling molecule carbon monoxide with its emergent vasorelaxant,antithrombotic, antiapoptotic, and antiinflammatory activities(24,27,46,49,50). In this scenario, excess hemoglobin (and heme)release is expected to cause or worsen vasoconstriction, plateletaggregation, and acute renal failure (Figure 7). This model,if confirmed, may help understanding the central role of thekidney in enteropathic (Escherichia coli–induced) HUS.Furthermore, lack of adequate HO-1 expression may contributeto the intravascular hemolytic process (45)
Figure 7. Schematic diagram of the proposed pathophysiological model of the effects of heme and verotoxin in Escherichia coli–associated hemolytic uremic syndrome.
In conclusion, our results led us to propose the hypothesisthat hemoglobin and hemoglobin-derived heme may play a previouslyunrecognized role in classical HUS. It is possible that VT-mediatedrepression of vital cell defense systems, particularly HO-1,contributes to the renal pathology of HUS.
Acknowledgments
Part of this work was presented at the Fifth International Symposiumon Shiga Toxin (Verocytotoxin)-Producing Escherichia coli, Edinburgh,Scotland, June 8 to 11, 2003. This work was supported by theDepartment of Pediatrics, the Wake Forest University Schoolof Medicine Physician-Scientist Mentorship Program, and theBrenner Children’s Hospital Foundation. We thank DonnaLeigh Holder for technical assistance, Drs. D. W. Busija, M.C. Willingham, M. O. Lively, and C. E. McCall for advice throughoutthis project, and Dr. Busija for critical reading of the articlein manuscript. The verotoxin preparations were the gift of Dr.M. A. Karmali.
Wong CS, Jelacic S, Habeeb RL, Watkins SL, Tarr PI: The risk of the hemolytic-uremic syndrome after antibiotic treatmnt of Escherichia coli O157:H7 infections. N Engl J Med 342: 1930–1936, 2000[Abstract/Free Full Text]
Siegler RL: The hemolytic uremic syndrome. Pediatr Clin North Am 42: 1505–1529, 1995[Medline]
Brandt JR, Fouser LS, Watkins SL, Zelikovic I, Tarr PI, Nazar-Stewart V, Avner ED: Escherichia coli O157:H7–associated hemolytic-uremic syndrome after ingestion of contaminated hamburgers. J Pediatr 125: 519–526, 1994[CrossRef][Medline]
Trachtman H, Cnaan A, Christen E, Gibbs K, Zhao S, Acheson DW, Weiss R, Kaskel FJ, Spitzer A, Hirschman GH: Effect of an oral Shiga toxin-binding agent on diarrhea-associated hemolytic uremic syndrome in children: A randomized controlled trial. JAMA 290: 1337–1344, 2003[Abstract/Free Full Text]
Dundas S, Todd WT, Stewart AI, Murdoch PS, Chaudhuri AK, Hutchinson SJ: The central Scotland Escherichia coli O157: H7 outbreak: Risk factors for the hemolytic uremic syndrome and death among hospitalized patients. Clin Infect Dis 33: 923–931, 2001[CrossRef][Medline]
O’Loughlin EV, Robins-Browne RM: Effect of Shiga toxin and Shiga-like toxins on eukaryotic cells. Microbes Infect 3: 493–507, 2001[CrossRef][Medline]
Sandvig K, Grimmer S, Lauvrak SU, Torgersen ML, Skretting G, van Deurs B, Iversen TG: Pathways followed by ricin and Shiga toxin into cells. Histochem Cell Biol 117: 131–141, 2002[CrossRef][Medline]
Uchida H, Kiyokawa N, Horie H, Fujimoto J, Takeda T: The detection of Shiga toxins in the kidney of a patient with hemolytic uremic syndrome. Pediatr Res 45: 133–137, 1999[Medline]
Kaneko K, Kiyokawa N, Ohtomo Y, Nagaoka R, Yamashiro Y, Taguchi T, Mori T, Fujimoto J, Takeda T: Apoptosis of renal tubular cells in Shiga-toxin–mediated hemolytic uremic syndrome. Nephron 87: 182–185, 2001[CrossRef][Medline]
Taguchi T, Uchida H, Kiyokawa N, Mori T, Sato N, Horie H, Takeda T, Fujimoto J: Verotoxins induce apoptosis in human renal tubular epithelium derived cells. Kidney Int 53: 1681–1688, 1998[CrossRef][Medline]
Pijpers AH, van Setten PA, van den Heuvel LP, Assmann KJ, Dijkman HB, Pennings AH, Monnens LA, van Hinsbergh VW: Verocytotoxin-induced apoptosis of human microvascular endothelial cells. J Am Soc Nephrol 12: 767–778, 2001[Abstract/Free Full Text]
Bitzan M, te Loo DM: Interaction of Shiga toxin with endothelial cells. Methods Mol Med 73: 243–262, 2003[Medline]
Bitzan MM, Wang Y, Lin J, Marsden PA: Verotoxin and ricin have novel effects on preproendothelin-1 expression but fail to modify nitric oxide synthase (ecNOS) expression and NO production in vascular endothelium. J Clin Invest 101: 372–382, 1998[Medline]
Foster GH, Armstrong CS, Sakiri R, Tesh VL: Shiga toxin-induced tumor necrosis factor alpha expression: Requirement for toxin enzymatic activity and monocyte protein kinase C and protein tyrosine kinases. Infect Immun 68: 5183–5189, 2000[Abstract/Free Full Text]
Karpman D, Hakansson A, Perez MT, Isaksson C, Carlemalm E, Caprioli A, Svanborg C: Apoptosis of renal cortical cells in the hemolytic-uremic syndrome: In vivo and in vitro studies. Infect Immun 66: 636–644, 1998[Abstract/Free Full Text]
Te Loo DM, Monnens LA, van den Heuvel LP, Gubler MC, Kockx MM: Detection of apoptosis in kidney biopsies of patients with D+ hemolytic uremic syndrome. Pediatr Res 49: 413–416, 2001[Medline]
Richardson SE, Karmali MA, Becker LE, Smith CR: The histopathology of the hemolytic uremic syndrome associated with verocytotoxin-producing Escherichia coli infections. Hum Pathol 19: 1102–1108, 1988[CrossRef][Medline]
Sassa S, Shibahara S: Disorders of heme production and catabolism. In: Blood. Principles and Practice of Hematology, edited by Handing RI, Lux IV SE, Stossel TP, Philadelphia, Lippincott Williams & Wilkins, 2003, pp 1435–1501
Bunn HF, Jandl JH: Exchange of heme among hemoglobins and between hemoglobin and albumin. J Biol Chem 243: 465–475, 1968[Abstract/Free Full Text]
Jeney V, Balla J, Yachie A, Varga Z, Vercellotti GM, Eaton JW, Balla G: Pro-oxidant and cytotoxic effects of circulating heme. Blood 100: 879–887, 2002[Abstract/Free Full Text]
Goldstein L, Teng ZP, Zeserson E, Patel M, Regan RF: Hemin induces an iron-dependent, oxidative injury to human neuron-like cells. J Neurosci Res 73: 113–121, 2003[CrossRef][Medline]
Hebbel RP, Eaton JW: Pathobiology of heme interaction with the erythrocyte membrane. Semin Hematol 26: 136–149, 1989[Medline]
Balla J, Balla G, Jeney V, Kakuk G, Jacob HS, Vercellotti GM: Ferriporphyrins and endothelium: A 2-edged sword—Promotion of oxidation and induction of cytoprotectants. Blood 95: 3442–3450, 2000[Abstract/Free Full Text]
Wagener FA, Eggert A, Boerman OC, Oyen WJ, Verhofstad A, Abraham NG, Adema G, van Kooyk Y, de Witte T, Figdor CG: Heme is a potent inducer of inflammation in mice and is counteracted by heme oxygenase. Blood 98: 1802–1811, 2001[Abstract/Free Full Text]
Zhang X, Fujii H, Matera KM, Migita CT, Sun D, Sato M, Ikeda-Saito M, Yoshida T: Stereoselectivity of each of the three steps of the heme oxygenase reaction: Hemin to meso-hydroxyhemin, meso-hydroxyhemin to verdoheme, and verdoheme to biliverdin. Biochemistry 42: 7418–7426, 2003[CrossRef][Medline]
Maines MD: The heme oxygenase system: A regulator of second messenger gases. Annu Rev Pharmacol Toxicol 37: 517–554, 1997[CrossRef][Medline]
Soares MP, Usheva A, Brouard S, Berberat PO, Gunther L, Tobiasch E, Bach FH: Modulation of endothelial cell apoptosis by heme oxygenase-1–derived carbon monoxide. Antioxid Redox Signal 4: 321–329, 2002[CrossRef][Medline]
Nath KA, Haggard JJ, Croatt AJ, Grande JP, Poss KD, Alam J: The indispensability of heme oxygenase-1 in protecting against acute heme protein-induced toxicity in vivo. Am J Pathol 156: 1527–1535, 2000[Abstract/Free Full Text]
Hill-Kapturczak N, Chang SH, Agarwal A: Heme oxygenase and the kidney. DNA Cell Biol 21: 307–321, 2002[CrossRef][Medline]
Zager RA: Rhabdomyolysis and myohemoglobinuric acute renal failure. Kidney Int 49: 314–326, 1996[Medline]
Nath KA, Grande JP, Haggard JJ, Croatt AJ, Katusic ZS, Solovey A, Hebbel RP: Oxidative stress and induction of heme oxygenase-1 in the kidney in sickle cell disease. Am J Pathol 158: 893–903, 2001[Abstract/Free Full Text]
Ishitoya S, Kurazono H, Nishiyama H, Nakamura E, Kamoto T, Habuchi T, Terai A, Ogawa O, Yamamoto S: Verotoxin induces rapid elimination of human renal tumor xenografts in SCID mice. J Urol 171: 1309–1313, 2004[CrossRef][Medline]
Eisenhauer PB, Chaturvedi P, Fine RE, Ritchie AJ, Pober JS, Cleary TG, Newburg DS: Tumor necrosis factor alpha increases human cerebral endothelial cell Gb3 and sensitivity to Shiga toxin. Infect Immun 69: 1889–1894, 2001[Abstract/Free Full Text]
Willingham MC: Cytochemical methods for the detection of apoptosis. J Histochem Cytochem 47: 1101–1110, 1999[Abstract/Free Full Text]
Richardson SE, Rotman TA, Jay V, Smith CR, Becker LE, Petric M, Olivieri NF, Karmali MA: Experimental verocytotoxemia in rabbits. Infect Immun 60: 4154–4167, 1992[Abstract/Free Full Text]
Siegler RL, Pysher TJ, Tesh VL, Taylor FB, Jr.: Response to single and divided doses of Shiga toxin-1 in a primate model of hemolytic uremic syndrome. J Am Soc Nephrol 12: 1458–1467, 2001[Abstract/Free Full Text]
Hill-Kapturczak N, Sikorski E, Voakes C, Garcia J, Nick HS, Agarwal A: An internal enhancer regulates heme- and cadmium-mediated induction of human heme oxygenase-1. Am J Physiol Renal Physiol 285: F515–523, 2003[Abstract/Free Full Text]
Beutler E: Production and destruction of erythrocytes. In: Williams Hematology, edited by Beutler E, Coller BS, Lichtman MA, Kipps TJ, Seligsohn U. New York, McGraw-Hill, 2001, pp 355–368
Bunn HF, Jandl JH: The renal handling of hemoglobin. II. Catabolism. J Exp Med 129: 925–934, 1969[Abstract]
Grinshtein N, Bamm VV, Tsemakhovich VA, Shaklai N: Mechanism of low-density lipoprotein oxidation by hemoglobin-derived iron. Biochemistry 42: 6977–6985, 2003[CrossRef][Medline]
Dhar GJ, Bossenmaier I, Cardinal R, Petryka ZJ, Watson CJ: Transitory renal failure following rapid administration of a relatively large amount of hematin in a patient with acute intermittent porphyria in clinical remission. Acta Med Scand 203: 437–443, 1978[Medline]
Graca-Souza AV, Arruda MA, de Freitas MS, Barja-Fidalgo C, Oliveira PL: Neutrophil activation by heme: Implications for inflammatory processes. Blood 99: 4160–4165, 2002[Abstract/Free Full Text]
Balla J, Vercellotti GM, Nath K, Yachie A, Nagy E, Eaton JW, Balla G: Haem, haem oxygenase and ferritin in vascular endothelial cell injury. Nephrol Dial Transplant 18 [Suppl 5]: v8–v12, 2003[Abstract/Free Full Text]
Abraham NG, Scapagnini G, Kappas A: Human heme oxygenase: Cell cycle-dependent expression and DNA microarray identification of multiple gene responses after transduction of endothelial cells. J Cell Biochem 90: 1098–1111, 2003[CrossRef][Medline]
Yachie A, Niida Y, Wada T, Igarashi N, Kaneda H, Toma T, Ohta K, Kasahara Y, Koizumi S: Oxidative stress causes enhanced endothelial cell injury in human heme oxygenase-1 deficiency. J Clin Invest 103: 129–135, 1999[Medline]
Poss KD, Tonegawa S: Reduced stress defense in heme oxygenase 1–deficient cells. Proc Natl Acad Sci U S A 94: 10925–10930, 1997[Abstract/Free Full Text]
Dennery PA, Spitz DR, Yang G, Tatarov A, Lee CS, Shegog ML, Poss KD: Oxygen toxicity and iron accumulation in the lungs of mice lacking heme oxygenase-2. J Clin Invest 101: 1001–1011, 1998[Medline]
Gonzalez-Michaca L, Farrugia G, Croatt AJ, Alam J, Nath KA: Heme: A determinant of life and death in renal tubular epithelial cells. Am J Physiol Renal Physiol 286: F370–F377, 2004[Abstract/Free Full Text]
Perrella MA, Yet SF: Role of heme oxygenase-1 in cardiovascular function. Curr Pharm Des 9: 2479–2487, 2003[CrossRef][Medline]
Wagner M, Cadetg P, Ruf R, Mazzucchelli L, Ferrari P, Redaelli CA: Heme oxygenase-1 attenuates ischemia/reperfusion-induced apoptosis and improves survival in rat renal allografts. Kidney Int 63: 1564–1573, 2003[CrossRef][Medline]
Received for publication January 31, 2004.
Accepted for publication June 15, 2004.
Top
Abstract
Introduction
200 µM caused cell rounding, spike formation, and detachment that was morphologically distinct from verocytotoxicity. VT caused apoptosis at concentrations >100 pM, as demonstrated by nuclear segmentation and poly(ADP-ribose) polymerase cleavage, whereas hemin-mediated injury of ACHN cells grown in serum-containing medium lacked attributes of programmed cell death. Pretreatment of ACHN monolayers with sublethal concentrations (1 to 10 pM) of VT for 12 to 18 h led to superadditive hemin-mediated cytotoxicity. This effect was not limited to ACHN cells, but was similarly noted in microvascular endothelial cells. Heme catabolism is regulated by (inducible) heme oxygenase-1 (HO-1). VT abrogated HO-1 expression in ACHN cells. Stimulation of cells for 6 h with CdCl2, which markedly increased HO-1 expression before the addition of VT, blunted subsequent hemin injury. In conclusion, VT augments hemin-induced toxicity in renal tubular epithelial cells that can be reversed by prior induction of HO-1. It is proposed that VT subverts the physiologic defense against heme proteins by interfering with the regulated expression of HO-1 and that this mechanism contributes to the renal pathology in patients with Escherichia coli–associated HUS. 
dimers, and reabsorbed and catabolized in proximal tubule cells. Spontaneous oxidation to met[Fe3+]hemoglobin destabilizes extraerythrocytic hemoglobin, which results in the release of the heme moiety (19–21). Oxidized (ferric) protoporphyrin IX (hemin) is commonly isolated as a halide (hemin chloride) (18). It is a highly hydrophobic molecule capable of inducing adverse biologic effects. It intercalates with the membrane lipid bilayer and potentiates oxidant-mediated injury (20,22–24). The exact mechanism or mechanisms of hemin toxicity and the predominant pathway of cell death remains to be defined. 
), monolayers were incubated with CdCl2 for 6 h, followed by VT2 and hemin. As shown in Figure 6A, CdCl2 treatment resulted in blunting of hemin-induced cytotoxicity. CdCl2 failed to prevent the cytotoxic effect of VT alone (results not shown). However, the prior induction of HO-1 reversed the combined cytotoxic effect of hemin and VT in ACHN cells. Results from a typical experiments that used various VT2 (10 to 1000 pM) concentrations and hemin over 6 and 48 h are depicted in Figure 6, B and C. Taken together, results indicate that the induction of HO-1 protects ACHN cells from augmented hemin toxicity in the presence of VT. 
