Molecular Basis for High Renal Cell Sensitivity to the Cytotoxic Effects of Shigatoxin-1: Upregulation of Globotriaosylceramide Expression
Alisa K. Hughes,
Zuhal Ergonul,
Peter K. Stricklett and
Donald E. Kohan
Division of Nephrology, University of Utah School of Medicine and Salt Lake Veterans Affairs Medical Center, Salt Lake City, Utah.
Correspondence to Dr. Donald E. Kohan, Division of Nephrology, University of Utah Health Sciences Center, Salt Lake City, UT 84132. Phone: 801-581-6709; Fax: 801-581-4343; E-mail: donald.kohan{at}hsc.utah.edu
ABSTRACT. Cellular injury in post-diarrheal hemolytic-uremicsyndrome (D+HUS) is related to shigatoxin (Stx) binding to globotriaosylceramide(Gb3). High renal Gb3 expression may determine renal susceptibilityin D+HUS; however, the molecular mechanism(s) responsible forsuch relatively abundant Gb3 levels are unknown. Consequently,kidney cells expressing high Gb3 (cultured human proximal tubulecells [HPT]) were compared with non-kidney cells with low Gb3content (cultured human brain microvascular endothelial cells[HBEC]). HPT were much more sensitive to the cytotoxic and proteinsynthesis inhibitory effects of Stx-1; this correlated withGb3 content and 125I-Stx-1 binding. HPT had greater Gb3 synthase(GalT6) and lower -galactosidase activities than HBEC, whereaslactosylceramide synthase (GalT2) activity was higher in HBEC.Ceramide glucosyltransferase (CGT) activity was similar betweenthe two cell types. The higher HPT GalT6 activity was associatedwith increased GalT6 mRNA steady-state levels, but no differencein GalT6 mRNA half-life. The lower HPT -galactosidase activitywas associated with reduced -galactosidase mRNA steady-statelevels but no difference in -galactosidase mRNA half-life. HigherHBEC GalT2 activity was associated with increased steady-stateGalT2 mRNA levels. These studies suggest that high renal Gb3expression is due to enhanced GalT6 gene transcription and reduced-galactosidase gene transcription and occur despite relativelylow GalT2 activity.
Post-diarrheal hemolytic-uremic syndrome (D+HUS) is the leadingcause of acute renal failure in children. The disorder is characterizedby thrombocytopenia, microangiopathic hemolytic anemia, andacute renal insufficiency (1). Although other organ systemsmay become involved in D+HUS, renal injury is prominent andoccurs early in the course of the disease. Renal cell damagein D+HUS has been related to enteric infection by shigatoxin(Stx)-producing organisms (typically Escherichia coli 0157:H7)(2). Stx binds to glycosphingolipids with a galactose--1,4,galactose moiety, the most abundant form being galactose--1,4,galactose--1,4, glucose-ceramide (Gb3). Once bound, the toxinis internalized, and ultimately is cytotoxic, at least in part,by directly inhibiting peptide elongation. The observationsthat intact kidney and isolated renal cells express relativelyhigh amounts of Gb3 on the cell surface has been used to explainrenal targeting in D+HUS (2–6). Although this scenariois generally accepted, it remains unknown as to why the kidneycontains such large amounts of Gb3.
Several components of the Gb3 metabolic pathway could be involvedin high renal Gb3 expression (Figure 1). Gb3 is synthesizedfrom lactosylceramide (LacCer) and UDP-galactose by Gb3 synthase(GalT6) (7) and is metabolized to LacCer by -galactosidase (8).LacCer is derived, in turn, from glucosylceramide (GlcCer) andUDP-galactose by LacCer synthase (GalT2) (9), whereas GlcCeris synthesized from ceramide and UDP-glucose by ceramide glucosyltransferase(CGT) (10). Consequently, increased renal Gb3 content couldconceivably be due to increased activity of GalT6, GalT2, orCGT, and/or it could be due to decreased activity of -galactosidase.Currently, there is no information on the activities or regulationof these enzymes in the kidney. The current study was thereforeundertaken to define how these enzymes are involved in controllingthe levels of renal cellular Gb3 expression.
On light microscopy, glomerular endothelial cell injury is aprominent feature of D+HUS (2). This led to the hypothesis thatStx binds directly to these cells by virtue of their high basalGb3 expression. This hypothesis was supported by the observationthat cultured human microvascular endothelial cells were muchmore sensitive to the cytotoxic effects of Stx compared withhuman umbilical vein endothelial cells (3). Subsequent studieshave found, however, that cultured human glomerular endothelialcells are quite insensitive to Stx (6,11). In contrast, humanproximal tubular cells (HPT) have been found to be exquisitelysensitive to Stx cytotoxicity (12). Indeed, one group has notedthat in a side-by-side comparison, Stx-1 potently inhibitedprotein synthesis in HPT, but had minimal effect on glomerularendothelial cells (11). In addition, in Fabry disease, a conditiondue to -galactosidase deficiency, proximal tubule Gb3 accumulationis a major feature of the renal damage (13). Consequently, HPTwere chosen for this study as the model of renal cells expressinghigh basal Gb3 levels. Human brain endothelial cells (HBEC),which contain very low Gb3 and are resistant to the cytotoxiceffects of Stx-1, were chosen for the purpose of comparison(14,15). Several other cell types with low Gb3 expression couldhave been used, but HBEC have been well documented to containlittle Gb3 under basal conditions (14,15).
Cell Culture
HBEC were obtained at primary culture from Cell Systems (Kirkland,WA) and studied at passages 7 to 9. Cells were grown to confluencein EGM2-MV media (Clonetics, San Diego, CA) and switched toserum-free Maintenance Formula media (Cell Systems) 24 h beforeall studies were initiated. In addition to characterizationby Cell Systems, we determined that these cells had uniformlypositive immunofluorescence for von Willebrand factor and plateletendothelial cell adhesion molecule but were negative for cytokeratin.HPT were obtained from Clonetics and studied at the third passage.HPT were maintained in 1:1 Dulbecco’s Modified Eagle Media:Ham’sF12 containing 25 mM HEPES, 2 mM L-glutamine, 100 U/ml penicillin/streptomycin,250 µg/L amphotericin B, 1 µg/ml hydrocortisone,10 µg/ml insulin, 5.5 mg/ml transferrin, 6.7 µg/mlselenium, 0.2 g/L ethanolamine, 6.5 ng/ml L-thyroxine, 10 ng/mlepidermal growth factor, and 10% fetal bovine serum (FBS). Cellswere switched to serum-free Maintenance Formula media 24 h beforeall studies were initiated. The identity of these cultures wasestablished as described previously (12).
Cytotoxicity
Cells grown in 96-well plates were analyzed for neutral reduptake as described previously (12) after 24-h exposure to varyingconcentrations of Stx-1. Cells were incubated in 50 µg/mlneutral red in M199 + 5% FBS for 3 h at 37°C, rinsed in1% formaldehyde and 1% CaCl2, followed by addition of 50% ethanoland 1% acetic acid. Absorbance was determined at OD450.
Leucine Incorporation 3H-incorporation was determined as described previously (12)after 4-hr exposure to varying concentrations of Stx-1. Cellsgrown in 24-well plates were incubated with 1 µCi/ml 3H-leucinefor 20 min at room temperature, rinsed, and solubilized in 0.1%sodium dodecyl sulfate. Protein was precipitated with 10% tricarboxylicacid, collected on GF/C filters (Whatman, Kent, England), andcpm was determined.
Gb3 Content
Gb3 content was determined as described previously (12). Cellswere extracted in chloroform:methanol:water and separated onhigh-performance thin-layer chromatography-silica plates (MallinckrodtBaker Inc., Paris, KY). The plates were dried, immersed in 0.5%polyisobutylmethacrylate in acetone, and sequentially incubatedwith Stx-1, anti-Stx-1 monoclonal antibody (purified from ahybridoma cell line, 13C4 [ATCC, Rockville, MD]), and 125I-goatanti-mouse IgG (DuPont NEN, Boston, MA). Gb3 concentrationswere calculated by densitometry and standardized to total protein.Before centrifugation, a cell aliquot was solubilized in 0.1N NaOH and mixed with Bradford reagent (Bio-Rad, Richmond, CA),and protein concentration was determined by measuring absorbanceat 590 nm.
Stx-1 Binding
Cells grown in 96-well plates were used for 125I-Stx-1 bindingassays as described previously (12). 125I-Stx-1 (17,000 cpm,Stx-1 iodinated according to the Iodobead manufacturer’sprotocol [Pierce, Rockford, IL]) in 100 µl of M199 containing5% FBS and 25 mM HEPES plus varying concentrations of unlabeledStx-1 was added for 24 h at 4°C. Cells were rinsed withice-cold HBSS and solubilized in 0.1 N NaOH, and cpm was determined.
Biosynthetic Enzyme Activity
GalT6 activity was determined as described previously (12).Briefly, cells were homogenized in 500 µl of 50 mM MESpH 6.5. Dried LacCer (25 nmol; Matreya, State College, PA) wasadded to sodium cholate in water (250 µg) and dried undervacuum, and the dried mixture incubated for 60 min at 4°C.A total volume of 100 µl of 50 mM MES pH 6.5 containing10 mM MnCl2, 100 µM 5'-adenylimidodiphosphate (p(NH)ppA),250 µM cold UDP-galactose, 44 µM UDP-14C galactose(150,000 to 400,000 cpm; Amersham Biosciences, Piscataway, NJ),and 125 µg of total cellular protein was added to thedried LacCer/sodium cholate, the samples vortexed, incubatedat 37° for 1 hr, and the reaction stopped by adding 1 mlof 2:1 chloroform:methanol. A Folch partition was establishedby adding 200 µl of 0.1 M KCl, and the upper phase re-extractedby adding 500 µl of 2:1 chloroform:methanol. The lowerphase was reextracted by adding 500 µl of 1:1 methanol:0.1M KCl. The lower phases (containing the neutral lipids) werecombined, dried under vacuum, and chromatographed as describedfor Gb3 above. For GalT2 and GlcCer synthase activity, cellswere treated as for GalT6. Either dried GlcCer (GalT2) or ceramide(GlcCer synthase) (both reagents from Matreya) were vortexedwith 2.5 µl of 10% TritonX-100 and incubated at room temperaturefor 30 min. A total volume of 100 µl of 50 mM HEPES (pH6.8) containing 5 mM MgCl2, 5 mM MnCl2, 55.7 µM UDP-galactose,44 µM UDP-14C galactose (150,000 to 400,000 cpm), and150 µg of cell sonicate were added to the GlcCer or ceramide(100 nM final concentration)/detergent mixture, vortexed, andincubated at 37°C for 1.5 hr. A Folch partition was established,and the neutral lipids chromatographed as described for GalT6.
-Galactosidase Activity -Galactosidase activity was determined as described previously(16). Cells were suspended in 3 mg/ml sodium taurocholate and28 mM citric acid/44 mM Na2HPO4 (pH 4.4), sonicated, and centrifuged,and the supernatant was analyzed. For the reaction, supernatantwas incubated with 5 mM p-nitrophenyl--galactopyranoside for10 to 90 min. followed by addition of 0.2 M Na2CO3 and absorbancemeasured at 400 nm.
Northern Analyses
Total RNA was isolated from confluent cells, electrophoresedon a 0.9% formaldehyde gel, transferred to a nylon membrane,and prehybridized for 3 h at 60°C in 50% formamide, 5x SSC,5x Denhardt’s solution, 1% sodium dodecyl sulfate, and100 µg/mL salmon sperm DNA. Fresh solution was added forhybridization along with radioactively labeled probe. For probes,cDNA was made from human proximal tubule cell total RNA usingoligo dT mRNA primer and SuperScript II reverse transcriptase(Invitrogen, Carlsbad, CA). The cDNA was then used as a templatefor PCR amplification of the coding region of the gene usingspecific primers: (1) GalT6 (GeneBank Accession No. AB037883):Forward, 5'- GAT CTG GGG ATA CCA TGT CCA AG -3' and Reverse,5'- CAG TAG CGG GCA TGC AGC TGG -3' , which yields a productsize of 1040 bp; (2) -galactosidase (GeneBank Accession No.XM037096): Forward, 5'- GGC TAG AGC ACT GGA CAA TGG A -3' andReverse, 5'- CTG CGA TGG TAT AAG AGC GAG G -3', which yieldsa product size of 1021 bp; (3) GalT2 (GeneBank Accession No.AF097159): Forward, 5'- AAC GGT ACA GAT TAT CCC GAA GG -3' andReverse, 5'- TGG AGC TAA CTC TGG CAT GAG G -3', which yieldsa product size of 912 bp; and (4) Ceramide glucosyltransferase(GeneBank Accession No. D50840): Forward, 5'- GCT GTG GCT GATGCA TTT CAT GG -3' and Reverse, 5'- CAG TTC TCC AGC TTA TAGTTG GG-3', which yields a product size of 1070 bp. PCR conditionswere the same for each reaction: one eighth of the RT reactionwas amplified per reaction using Taq polymerase in a reactionmixture containing 800 uM dNTP, 1 pmol of each primer, and 50mM MgCl2 (all reagents and enzymes from Invitrogen). PCR conditionswere as follows: 94° for 1.5 min; 94° for 20 s; 65°for 25 s; 72° for 1 min (30 cycles); 72° for 5 min,final extension. All reactions were tested for size before combiningand purifying using Wizard PCR preps (Promega, Madison, WI)as per manufacturer’s protocol. All products were purified,sequenced, and cloned into pGEM-T cloning vector (Promega).The inserts were again sequenced to ensure cloning fidelityand confirm orientation. Probes were digested with appropriaterestriction enzyme to give the antisense strand, and riboprobeswere made using 32P-UTP incorporation with either T7 or SP6RNA polymerase (Invitrogen). The radioactive probes were purifiedover a G-50 column, and specific activity was calculated. Theprobe was added to hybridization solution at 10 ng/ml, witha specific activity 109 dpm/µg and incubated overnightat 60°C. Blots were washed in decreasing SSC concentrationsand increasing temperature until background removed. Labeledblots were subjected to autoradiography and densitometry.
mRNA Half-Life Determination
Confluent cells in 6-well plates were treated with 10 µg/mlactinomycin D for 0 to 24 h followed by determination of GalT6and -galactosidase mRNA levels as described under Northern analysisabove.
Statistical Analyses
Data were analyzed by one-way ANOVA using the Bonferroni correction.Where only two data points were compared, the unpaired t testwas used. P < 0.05 was taken as significant. Data are expressedas mean ± SEM.
The work to be described was conducted using HPT and HBEC. Atthe outset, it is important to note why human glomerular endothelialcells (HGEC), which have been implicated as a primary targetin HUS, were not studied. First, in our hands, pure culturesof HGEC are difficult to obtain—the yield is quite small,and the cells rapidly lose endothelial cell markers with passaging.Additionally, the one group (6) that has published studies usingHGEC reports that these cells are quite insensitive to Stx andcontain relatively little Gb3. We have also obtained severaldifferent batches of commercially available HGEC (Cell Systems,Kirkland, WA) and have found these cells, regardless of donorage, to express little Gb3 and to be resistant to Stx. Thus,in our opinion, it remains unknown whether HGEC really are sensitiveto Stx in vivo; this issue is problematic until a better experimentalsystem can be devised to assess HGEC function in vitro.
To establish that HPT and HBEC serve as models of high and lowGb3-expressing cells, respectively, Stx-1 cytotoxicity, Stx-1protein synthesis inhibition, Stx-1 binding, and Gb3 contentwere determined. Stx-1 potently killed HPT cells (LD50, approximately100 pg/ml), whereas HBEC were highly resistant (Figure 2). Similarresults were obtained with Stx-1 inhibition of protein synthesis(Figure 2). This high HPT sensitivity was associated with over100-fold greater Gb3 content and 125I-Stx-1 binding (Figure 3)as compared with HBEC. Hence, these data confirm that HPTand HBEC should serve as valid models.
Figure 2. Effect of shigatoxin-1 (Stx-1) on human proximal tubule cell (PT) and human brain microvascular endothelial cell (HBEC) survival (neutral red uptake) and 3H-leucine incorporation. Panel A shows survival (24-h Stx-1 exposure, n = 12 each data point), and panel B shows 3H-leucine incorporation (4-h Stx-1 exposure, n = 3 each data point).
Figure 3. Total cell Gb3 content (panel A, n = 4 each data point) and 125I-Stx-1 binding (panel B, 24-h exposure at 4°C, n = 6 each data point) in human PT and HBEC. *P < 0.001 versus HBEC.
CGT, GalT2, GalT6, and -galactosidase activities were next determinedin HPT and HBEC to compare Gb3 biosynthetic pathways betweenthe two cell types. No differences were detected in CGT activity(Figures 4 and 5). GalT2 activity was lower in HPT than in HBEC(Figures 4 and 5), an unexpected finding that indicated thatGalT2 activity was not rate-limiting for Gb3 formation. GalT6activity was approximately twofold greater in HPT (Figures 4 and 5).In addition, -galactosidase was about fivefold lessin HPT than in HBEC (Figure 5). Thus, when the relative degreesof GalT6 and -galactosidase activities are considered together,it is evident that there is about a tenfold greater drive forGb3 accumulation in HPT than in HBEC.
Figure 4. High-performance thin-layer chromatography of GalT6, GalT2, and CGT activities in human PT and HBEC. Gb3 bands are GalT6 activity, lactosylceramide (LacCer) bands are GalT2 activity, and glucosylceramide (GlcCer) bands are CGT activity. Standards confirming glycosphingolipid location were run on the same gels but are not shown. All data are normalized to total cell protein.
Figure 5. Densitometry of GalT6, GalT2, and CGT activities in human PT and HBEC. Ordinate values are expressed as percent of HBEC enzyme activity. n = 6 each data point; *P < 0.005 versus HBEC; **P < 0.001 versus HBEC.
To determine the mechanism of enhanced Gb3 biosynthetic enzymeactivity, CGT, GalT2, GalT6, and -galactosidase mRNA levelswere assessed in the two cell types. In general, steady-statemRNA levels directly correlated with enzyme activity (Figures 6 and 7).CGT mRNA tended to be higher in HPT, but this didnot achieve statistical significance. GalT2 message was reducedin HPT as compared with HBEC. GalT6 mRNA content was approximatelytwofold greater in HPT, whereas -galactosidase mRNA was aboutthreefold less in PT. Again, as for the enzyme activities, thecombination of relative GalT6 and -galactosidase mRNA levelsindicate that HPT have a substantially greater Gb3 biosyntheticdrive than do HBEC.
Figure 7. Densitometry of GalT6, GalT2, CGT, and -galactosidase mRNA steady-state levels in human PT and HBEC. Ordinate values are expressed as percent of HBEC mRNA levels. n = 3 each data point; *P < 0.025 versus HBEC; **P < 0.01 versus HBEC.
The mRNA half-lives for GalT6 and -galactosidase were evaluatedto determine the mechanism of increased mRNA levels of thesetwo enzymes. The half-lives for both GalT6 and -galactosidase(Figure 8) were similar between HPT and HBEC. Consequently,these data suggest that the differences in GalT6 and -galactosidaseactivities between HPT and HBEC are controlled, at least inpart, at the transcriptional level.
Figure 8. Densitometry of GalT6 (panel A) or -galactosidase (panel B) mRNA half-life in human PT and HBEC. Cells were exposed to actinomycin D (10 µg/ml) for up to 24 h followed by Northern analysis of GalT6 or -galactosidase mRNA levels at each time point. n = 3 each data point.
High renal Gb3 expression is likely to play an important rolein kidney damage in at least two diseases, namely D+HUS andFabry’s disease. In D+HUS, Gb3 serves as the major receptorfor the putative pathogenic toxin, shigatoxin (2). In Fabrydisease (-galactosidase deficiency), Gb3 accumulates in lysosomesand eventually causes severe tubulointerstitial damage and renalfailure (13). Despite the central importance of Gb3, to datethere has been no information on the mechanisms responsiblefor predominant renal expression of this glycosphingolipid.Such information would potentially be of benefit in designingtherapeutic strategies for these disorders. This is particularlyimportant in that current treatments have not proven highlyeffective in ameliorating D+HUS (enteral toxin-absorbing resins)(2) or Fabry disease (intravenous recombinant -galactosidase)(13). Indeed, efforts are now underway toward developing and/orusing inhibitors of glycosphingolipid biosynthesis in thesediseases (17). Clearly, an understanding of which enzymes areresponsible for high renal Gb3 levels, as well as an understandingof how these enzyme levels are regulated, would be crucial inthe development of rational therapies.
The present study constitutes the first description of factorsinvolved in high renal Gb3 expression. First, it was not surprisingthat CGT activity and mRNA were not upregulated in HPT in thatthis constitutes the earliest step in synthesis of all glycosphingolipidsbased on glucosylceramide. However, the findings that GalT2activity and mRNA content were reduced in HPT were unexpected.This clearly suggests that GalT2 activity and resultant LacCerlevels are not rate-limiting for Gb3 accumulation in HPT. Thereasons why GalT2 activity is lower in HPT are unknown becausethe factors responsible for directly or indirectly modifyingGalT2 activity and/or mRNA levels have not been determined.One might speculate that downstream glycosphingolipid productsderived from LacCer could modulate GalT2 activity, but the biologicsignificance of renal glycosphingolipids is also poorly understood.This issue will therefore need to await additional studies.
In contrast to GalT2, GalT6 activity was increased in HPT. Thisobservation is as predicted and serves to explain, at leastin part, elevated HPT Gb3 levels. Furthermore, the combinationof increased GalT6 mRNA steady-state levels, together with nodifference in GalT6 mRNA half-life, suggests that GalT6 genetranscription is upregulated in HPT. The GalT6 gene has onlyrecently been cloned (7), and relatively little is known aboutits regulation. Tumor necrosis factor and protein kinase Cinduce GalT6 enzyme activity in endothelial cells (5). Interestingly,the GalT6 promoter contains an AP-1 domain 200 bp 5' to thetranscription start site; whether this is relevant to heightenedrenal Gb3 expression remains to be determined. Studies are currentlyunderway involving transfection of HPT with GalT6 promoter-reporterconstructs to begin to identify promoter domains critical forgene expression.
-Galactosidase activity was markedly reduced in HPT comparedwith HBEC. -Galactosidase steady-state mRNA levels were alsogreatly reduced in HPT and, similar to GalT6, -galactosidasemRNA half-life was not different between the two cell types.These data suggest that decreased -galactosidase activity inHPT is a result of decreased gene transcription. As for GalT6,relatively little is known about -galactosidase gene regulation.There are CRE-BP1/c-Jun and Pax-2 (the latter may be involvedin renal development and Wilms tumor formation [18]) bindingdomains in the first 200 bp of the -galactosidase promoter immediately5' to the transcription start site. However, whether these areinvolved in renal-specific downregulation of Gb3 accumulationremains to be determined.
The combination of reduced -galactosidase and enhanced GalT6activities clearly provides a strong net driving force towardGb3 accumulation in HPT. Such a scenario is not without precedent.For example, the increase in rabbit intestinal microvillus Gb3expression that occurs with aging is associated with coordinateincreases in GalT6 and decreases in -galactosidase activities(19). These investigators did not, however, determine mRNA levelsof these enzymes; it is therefore unclear if alterations ingene transcription similar to those seen in the present studyare involved. It is nonetheless tempting to speculate that coordinatedand inverse regulation of GalT6 and -galactosidase gene expressionmay involve common regulatory factors.
In summary, using HPT as a model of high Gb3-containing renalcells, these data indicate that the combination of increasedGalT6 and decreased -galactosidase enzyme activities and geneexpression is responsible for enhanced renal Gb3 expression.Further characterization of the molecular mechanisms responsiblefor this pattern of enzyme activity may ultimately lead to developmentof therapies that reduce renal Gb3 levels and prevent and/orameliorate renal injury in D+HUS and Fabry’s disease.
Acknowledgments
This research was support by National Institutes of Health grantsDK52043 and DK58953 (both to DEK).
Ruggenenti P, Noris M, Remuzzi G: Thrombotic microangiopathy, hemolytic uremic syndrome, and thrombotic thrombocytopenic purpura. Kidney Int 60: 831–846, 2001[CrossRef][Medline]
Obrig TG, Louise CB, Lingwood CA, Boyd B, Barley-Maloney L, Daniel TO: Endothelial heterogeneity in shiga toxin receptors and responses. J Biol Chem 268: 15484–15488, 1993[Abstract/Free Full Text]
van de Kar N, Monnens LAH, Karmali MA, van Hinsbergh VWM: Tumor necrosis factor and interleukin-1 induce expression of the verocytotoxin receptor globotriaosylceramide on human endothelial cells: implications for the pathogenesis of the hemolytic uremic syndrome. Blood 80: 2755–2764, 1992[Abstract/Free Full Text]
van de Kar N, Kooistra T, Vermeer M, Lesslauer W, Monnens LAH, van Hinsbergh VWM: Tumor necrosis factor induces endothelial galactosyl transferase activity and verocytotoxin receptors. Role of specific tumor necrosis factor receptors and protein kinase C. Blood 85: 734–743, 1995[Abstract/Free Full Text]
van Setten P, van Hinsbergh V, van der Velden T, van de Kar N, Vermeer M, Mahan J, Assmann K, van den Heuvel L, Monnens L: Effects of TNF on verocytotoxin cytotoxicity in purified glomerular microvascular endothelial cells. Kidney Int 51: 1245–1256, 1997[Medline]
Kojima Y, Fukumoto S, Furukawa K, Okajima T, Wiels J, Yokoyama K, Suzuki Y, Urano T, Ohta M, Furukawa K: Molecular cloning of globotriaosylceramide/CD77 synthase, a glycosyltransferase that initiates the synthesis of globo series glycosphingolipids. J Biol Chem 275: 15152–15156, 2000[Abstract/Free Full Text]
Bishop DF, Calhoun DH, Bernstein HS, Hantzopoulos P, Quinn M, Desnick RJ: Human -galactosidase A: Nucleotide sequence of cDNA clone encoding the mature enzyme. Proc Natl Acad Sci 83: 4859–4863, 1986[Abstract/Free Full Text]
Takizawa M, Nomura T, Wakisaka E, Aoki J, Arai H, Inoue K, Hattori A, Matsuo N: cDNA cloning and expression of human lactosylceramide synthase. Biochem Biophys Res Commun 1438: 301–304, 1999
Ichikawa S, Sakiyama H, Suzuki G, Hidari KI, Hirabayashi Y: Expression cloning of a cDNA for human ceramide glucosyltransferase that catalyzes the first glycosylation step of glycosphingolipid synthesis. Proc Natl Acad Sci USA 93: 4638–4643, 1996[Abstract/Free Full Text]
King A, Murray S, Acheson DWK, Keusch G: Sensitivity of human glomerular endothelial cells and tubular epithelial cells to shiga toxin-1 and -2: Implications for pathogenesis of acute renal failure. J Am Soc Nephrol 9: 580A, 1998
Hughes A, Stricklett P, Kohan D: Cytotoxic effect of Shiga toxin-1 on human proximal tubule cells. Kidney Int 54: 426–437, 1998[CrossRef][Medline]
Ohshima T, Murray GJ, Swaim WD, Longenecker G, Quirk JM, Cardarelli CO, Sugimoto Y, Pastan I, Gottesman MM, Brady RO, Kulkarni A: -Galactosidase A deficient mice: A model of Fabry disease. Proc Assoc Am Physicians 94: 2540–2544, 1997
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]
Ramegowda M, Samuel J, Tesh V: Interaction of shiga toxins with human brain microvascular endothelial cells: Cytokines as sensitizing agents. J Infect Dis 180: 1205–1213, 1999[CrossRef][Medline]
Abe A, Arend LJ, Lee L, Lingwood C, Brady RO, Shayman JA: Glycosphingolipid depletion in Fabry disease lymphoblasts with potent inhibitors of glucosylceramide synthase. Kidney Int 57: 446–454, 2000[Medline]
Shayman J, Lee L, Abe A, Shu L: Inhibitors of glucosylceramide synthase. Methods Enzymol 311: 373–387, 2000[Medline]
Dressler G, Douglass E: Pax-2 is a DNA-binding protein expressed in embryonic kidney and Wilms tumor. Proc Natl Acad Sci USA 89: 1179–1183, 1992[Abstract/Free Full Text]
Mobassaleh M, Koul O, Mishra K, McCluer RH, Keusch GT: Developmentally regulated Gb3 galactosyltransferase and -galactosidase determine shiga toxin receptors in intestine. Am Physiol 267: G618–G624, 1994
Received for publication April 15, 2002.
Accepted for publication June 13, 2002.
This article has been cited by other articles:
A. X. Garg, L. Moist, D. Matsell, H. R. Thiessen-Philbrook, R. B. Haynes, R. S. Suri, M. Salvadori, J. Ray, W. F. Clark, and for The Walkerton Health Study Investigators Risk of hypertension and reduced kidney function after acute gastroenteritis from bacteria-contaminated drinking water
Can. Med. Assoc. J.,
August 2, 2005;
173(3):
261 - 268.
[Abstract][Full Text][PDF]
Errata
J. Am. Soc. Nephrol.,
May 1, 2003;
14(5):
A8 - 8.
[Full Text][PDF]
Top
Abstract
Introduction
-galactosidase activities than HBEC, whereas lactosylceramide synthase (GalT2) activity was higher in HBEC. Ceramide glucosyltransferase (CGT) activity was similar between the two cell types. The higher HPT GalT6 activity was associated with increased GalT6 mRNA steady-state levels, but no difference in GalT6 mRNA half-life. The lower HPT 
-1,4, glucose-ceramide (Gb3). Once bound, the toxin is internalized, and ultimately is cytotoxic, at least in part, by directly inhibiting peptide elongation. The observations that intact kidney and isolated renal cells express relatively high amounts of Gb3 on the cell surface has been used to explain renal targeting in D+HUS (2–6). Although this scenario is generally accepted, it remains unknown as to why the kidney contains such large amounts of Gb3. 
3 h at 60°C in 50% formamide, 5x SSC, 5x Denhardt’s solution, 1% sodium dodecyl sulfate, and 100 µg/mL salmon sperm DNA. Fresh solution was added for hybridization along with radioactively labeled probe. For probes, cDNA was made from human proximal tubule cell total RNA using oligo dT mRNA primer and SuperScript II reverse transcriptase (Invitrogen, Carlsbad, CA). The cDNA was then used as a template for PCR amplification of the coding region of the gene using specific primers: (1) GalT6 (GeneBank Accession No. AB037883): Forward, 5'- GAT CTG GGG ATA CCA TGT CCA AG -3' and Reverse, 5'- CAG TAG CGG GCA TGC AGC TGG -3' , which yields a product size of 1040 bp; (2) 
