Main

HUS is characterized by hemolytic anemia, thrombocytopenia, and acute renal failure(1). The epidemic form of HUS in children is associated mostly with a prodromal disease of bloody diarrhea and has been termed D+, whereas the nonprodromal form has been termed D-. Infection with a VT-producing Escherichia coli has been strongly implicated in the etiology of the D+ form of HUS(2). Endothelial damage of glomeruli and kidney arterioles plays a pivotal role in the pathogenesis of HUS(3). The complex mechanisms leading to microvascular lesions in the kidney are not yet fully understood. In vitro studies with human umbilical vein, adult vein, and glomerular microvascular endothelial cells indicate that VT cytotoxicity in human endothelial cells requires the additional stimulation with inflammatory mediators, such as TNF-α. Induction of the globotriaosylceramide receptor, known to be the functional receptor for VT, proved to be the mechanism by which TNF-α induces VT cytotoxicity(4, 5). These data suggest that inflammatory mediators play an important role in the pathogenesis of HUS. This suggestion is strengthened by observations that endothelial cells respond to inflammatory mediators in many other ways, such as the induction of leukocyte adhesion molecule expression, the inhibition of antithrombotic properties, and the synthesis of growth factors, cytokines, the inducible species of nitric oxide synthase and cyclooxygenase(6–8). In vivo data supporting the involvement of inflammatory mediators are provided by several authors who demonstrated elevated concentrations of inflammatory mediators in predominantly urine of HUS patients(9–14). Besides, a local role for inflammatory mediators was suggested by Harel et al.(15), reporting that Shiga-like toxin injection into transgenic mice, bearing a CAT reporter gene coupled to a TNF-α promotor, induced CAT activity in the kidney but not in other tissues. The source of the local accumulation of these cytokines may be MOs and macrophages. In vitro, these cells produce and release a variety of inflammatory mediators upon stimulation with VT and lipopolysaccharide(16, 17). Moreover, once activated MOs and macrophages release a variety of other vasoactive agents(18), which may contribute to the endothelial damage observed in HUS patients. Whereas only recently the role of MO activation has been suggested in the pathogenesis of HUS(16, 17, 19), the pathogenic role of PMN activation has been well described[reviewed in Siegler(20)]. Therefore, the influx of leukocytes into the kidney may be a key event in the pathogenesis of HUS. The precise molecular mechanism for the recruitment of leukocytes into the glomeruli is unclear. Accumulating evidence suggests that chemokines, which are classified into two subfamilies (C-X-C and C-C) depending on the position of conserved cysteine residues, are prime candidates. Among these chemokines are MCP-1 (member of the C-C family) and IL-8 (member of the C-X-C family) potent, and specific chemotactic and activating factors for MOs and PMNs, respectively(21). To elucidate the role of chemokines in the pathogenesis of HUS, we determined urinary levels of MCP-1 and IL-8 in 15 HUS patients and correlated these with serum levels and clinical parameters. Furthermore, we studied the actual presence of MOs and PMNs in renal tissue of three HUS patients.

METHODS

Demographic and clinical data of HUS patients(Table 1). Sixteen, consecutive patients presenting with HUS and treated at the University Hospital Nijmegen, Department of Pediatric Nephrology, during the time period 1991-1995 were evaluated in this study. There were seven girls and nine boys, aged between 0.9 and 9.1 y. The criteria for HUS were defined as hemolytic anemia with burr cells in the peripheral blood smear, thrombocytopenia, and acute renal failure. Fifteen children presented with the D+ form, whereas only one patient did not have a diarrheal prodrome (D-). Evidence for an infection with a VT-producing E. coli O157 was found in 14 children, as determined by detection of antibodies to O157 lipopolysaccharide in serum. In agreement with previous reports, the total number of white blood cells was elevated (>15 × 109/L) in 63% of the HUS children (mean value 18.1 ± 9.7). In addition, elevated numbers of PMNs (>10.5 × 109/L) and MOs(>1.0 × 109/L) were observed in the 56 and 31% of the HUS children, respectively (mean values of 11.2 ± 6.9 and 1.3 ± 1.6). Fulfilling the criteria for HUS, the number of erythrocytes and platelets, and renal function were significantly decreased on admission. The mild form of renal damage as classified according to the description of Gianantonio et al.(22) was observed in seven children, whereas the moderate and severe form were observed in six and three children, respectively. Dialysis was required in 11 children, the duration of which ranged from 6 to 21 d. Neurologic dysfunction (seizures, coma, and blindness) was noticed in four children, of which one displayed also pancreas dysfunction (diabetes mellitus) as an extrarenal manifestation of HUS. All children were treated with urokinase and received supportive treatment(control of hypertension, dialysis when required). After 2 y, a follow-up study was performed in 13 children. In three children no information was available, one because of lack of cooperation and the other two children because they were diagnosed with HUS by the end of 1995 and therefore did not yet have their 2-y follow-up. A reduced GFR (<80 mL/min per 1.73 m2), as estimated from the formula: height (cm)/serum creatinine(μmol/L) × 39 was observed in three children. All children tested displayed a normal capacity to concentrate the urine, as determined by the desamino-8-d-arginine vasopressin test (>807 mOsmol/kg). Proteinuria(>150 mg/24 h) was observed in four children. Two children suffered from hypertension requiring antihypertensive treatment. One child displayed apparent neurologic sequelae (absences).

Table 1 Demographic and clinical data of HUS patients
Full size table

Histological studies. Kidney specimens of three different HUS patients were obtained from renal biopsy material taken within 7, 16, or 21 d after the start of the symptoms. Whereas two of these children presented with the D- form of HUS, one child had diarrhea associated HUS. The diagnosis of HUS was established by combining clinical data and standard pathologic methods (light microscopy, immunofluorescence, and in two cases supplementation with EM). The presence of monocytes/ macrophages was demonstrated histochemically on cryostat sections (2 μm) by an indirect immunofluorescence technique. The primary antibody (monoclonal anti-CD68 immunoglobulin 1:100; clone KP1; DAKO; Glostrup, Denmark) was incubated for 30 min at 20 °C in PBS/1% albumin. After washing with PBS, binding was visualized by incubating the sections for 30 min at 20 °C with FITC-labeled secondary antibody (1:20 diluted in PBS/1% albumin). Sections were washed with PBS, embedded in Vectashield (Burlinghame, CA) and examined with a Zeiss fluorescence microscope. Each staining experiment included a negative control by omitting the first antibody. The presence of PMNs was demonstrated by the Leder method, which assesses naphthol AS-D chloroacetate esterase activity(23).

Detection of MOs and PMNs in renal biopsy specimen by EM. For EM, renal biopsy material of two children with the D- form of HUS was fixed in 2.5% buffered glutaraldehyde, dehydrated in graded ethanols, and embedded in Epon 812. Subsequently, thin sections were cut with a ultramicrotome (Reichert) and stained with uranyl acetate and lead citrate. Sections were examined with a Jeol 1200 EX electron microscope.

Urine and blood sampling. Twenty-four-hour urinary samples of 15 HUS patients were collected, starting on either the day of admission or as soon as the anuria disappeared. Subsequently, serial samples were collected until recovery. Samples were stored in aliquots at -40 °C until assayed. Also, 24-h urinary samples were obtained from 17 healthy control children(mean age, 5.3 ± 1.3 y) and 19 children with a variety of inflammatory and noninflammatory glomerulopathies (mean age, 10.7 ± 3.3 y).

Serum samples were drawn on the day of admission from 14 HUS patients. To determine normal blood chemokine levels in children, serum samples were obtained from 14 healthy control children (mean age 8.3 ± 3.2). Furthermore, serum samples were drawn from 18 children suffering from a variety of inflammatory and noninflammatory glomerulopathies (mean age 9.2± 4.8) and nine children with end-stage renal disease treated by CAPD(mean age 8.3 ± 3.5). All blood samples were centrifuged for 10 min at 3000 × g and the sera were immediately stored in aliquots at-80 °C until assays were performed.

Immunochemical studies. Urinary MCP-1 and IL-8 immunoassay. Urinary MCP-1 and IL-8 were determined by sandwich ELISAs as described by van Setten et al.(24). Pilot experiments showed that a 1:5-100 dilution of urinary samples was needed to ensure the results to fall on the linear portion of the standard curve as well as to guarantee parallelism for different dilutions. Recoveries of recombinant MCP-1 and IL-8 spiked in urine were >95%. Sensitivities of MCP-1 and IL-8 ELISAs were 0.25 ng/mL and 25 pg/mL, respectively. Urinary MCP-1 and IL-8 values were expressed as ng/mmol creatinine. In urinary samples, in which chemokine concentrations appeared to be below the detection limit, chemokine values were calculated based on the lowest standard (0.25 ng/mL and 25 pg/mL, respectively). This resulted in an overestimation of urinary chemokine concentrations of mostly control children.

Serum MCP-1 and IL-8 immunoassay. Pilot experiments indicated that chemokine levels in HUS patients were only slightly elevated. Because commercially available human ELISA kits for MCP-1 and IL-8 (Quantikine; R&D systems Inc., Abingdon, UK) showed high sensitivities (5 and 18 pg/mL, respectively) and reliable results in particular when the samples were measured undiluted, serum chemokine levels were determined using these commercially available kits.

Statistical analysis. Urinary and serum chemokine levels are expressed as mean ± SD. The Mann-Whitney U test was used to analyze statistical significance of differences found between the urinary and serum chemokine results of the different groups of children. The correlation between urinary MCP-1 and IL-8 levels was calculated using Spearman's rank test. Statistical level of significance was defined as p < 0.05.

RESULTS

Detection of MOs and PMNs in renal biopsy specimens of HUS children. To examine the presence of MOs and PMNs, kidney specimens were studied by histochemical examination, using CD68 antigen as marker for MOs and a Leder staining to detect PMNs. Whereas renal tissue of control children demonstrated only small numbers of MOs and PMNs in the glomeruli (<3 cells per glomerular cross section; data not shown), renal tissues of all three HUS patients revealed the presence of increased numbers of MOs (Fig. 1A) and to a lesser extent of PMNs (Fig. 1B) in the glomeruli. The staining for MOs was specific because no staining was observed when the primary antibody was omitted. In addition, EM examination of renal biopsy specimens obtained from two children with the D- form of HUS demonstrated the presence of cells in the capillary lumen of the glomeruli, most probably representing MOs (Fig. 1C) and PMNs (Fig. 1D). Both histochemical and EM data showed that MOs represented the majority of white blood cells infiltrating the glomeruli of HUS children.

Figure 1
figure 1

Histologic (A and B) and EM(C and D) detection of MOs and PMNs in renal biopsy specimens obtained from children with HUS. (A) Section stained with the murine anti-CD68 antibody showing the presence of monocytes-macrophages in a representative glomerulus of a D+ HUS patient (×400,arrow indicates a CD68+ cell). (B) Section stained according to the method of Leder showing the presence of a PMN in a representative glomerulus of a D+ HUS patient (×350,arrow indicates a positive cell). Similar histologic results were obtained with renal biopsy specimens of two other HUS patients. (C) In addition to EM alterations characteristic for HUS, this section of a child with the D- form of HUS demonstrates the presence of cells occluding the capillary lumen of a representative glomerulus, suggestive for MOs(×4500; U, urinary space; arrowhead, deposition of basement membrane material; asterisk (*) indicates a cell occluding the capillary lumen, suggestive of a MO). (D) This section from a child with the D- form of HUS demonstrates the presence of an infiltrating PMN in a representative glomerulus (×37 00; arrowhead, deposition of basement membrane material; asterisk (*) indicates a PMN in the capillary lumen). Both histologic and EM data demonstrate that MOs represent the majority of white blood cells infiltrating the glomeruli.

Full size image

Urinary levels of MCP-1 and IL-8. To determine whether MCP-1 and IL-8 are involved in the observed increased presence of leukocytes in renal tissue of HUS patients, both chemokines were measured in urine of HUS patients and compared with urinary levels in healthy control children and children with a variety of inflammatory and noninflammatory glomerulopathies. Only limited amounts of MCP-1 (<45 ± 20 ng/mmol creatinine) and IL-8(<4 ± 2 ng/mmol creatinine) were detected in urine of 17 tested normal healthy control children. However, urinary levels of MCP-1 and IL-8 were significantly elevated in urine of all HUS patients examined.

Initial samples demonstrated increased levels of MCP-1 in all HUS patients, whereas IL-8 was significantly elevated in 10 out of 15 HUS patients. The mean MCP-1 and IL-8 concentrations in initial urine samples of all HUS patients were 2570 ± 1705 ng of MCP-1/mmol of creatinine (p < 0.001) and 120 ± 170 ng IL-8/mmol creatinine (p < 0.001). HUS children suffering from renal failure accompanied by anuria showed significantly higher MCP-1 levels in initial urine samples than those children with renal failure without anuria (Fig. 2). Similar results were obtained when initial urinary IL-8 levels of the two groups were compared.

Figure 2
figure 2

Urinary chemokine levels in HUS children suffering from renal failure with and without anuria. In this figure are shown mean MCP-1(A and B) and IL-8 (C and D) values in 24-h urinary samples of HUS children with (double hatched bars; n = 8) and without (hatched bars; n = 7) anuria. (A and C) Chemokine levels in initially produced 24-h urinary samples.(B and D) Maximum urinary chemokine levels reached during the course of the disease. Data are expressed as mean ± SD (ng/mmol creatinine). *p < 0.05.

Full size image

Serial measurements demonstrated either steadily increasing levels of MCP-1 and IL-8 during the course of the disease, reaching a maximum, followed by a gradual decline or a gradual decline until recovery. Figure 3 shows the time course of MCP-1 and IL-8 levels in 24-h urinary samples of two representative patients out of 15. The patient in Figure 3,A and B, was admitted to the hospital relatively early in the course of the disease, showing initially only a slightly increased serum creatinine value followed by further deterioration of renal function without periods of anuria and finally recovery of renal function. The patient shown in Figure 3,C and D, demonstrated a significantly increased serum creatinine on admission, followed by further deterioration of renal function with a short period of anuria and finally recovery of renal function. Similar time courses for urinary chemokines were observed in 13 other HUS patients.

Figure 3
figure 3

Time course of urinary chemokines in two representative HUS children out of 15. The patient in A and B was admitted to the hospital relatively early in the course of the disease, showing only slightly increased serum creatinine and urea on admission(B). Subsequently, renal function further deteriorated but no anuria occurred, after which the patient recovered from renal dysfunction. The patient in C and D demonstrate a significantly elevated serum creatinine on admission, followed by further deterioration of renal function with a short period of anuria and finally recovery of renal function.(A and C) Chemokine levels in 24-h urinary samples in ng/mmol creatinine (circles, MCP-1; triangles, IL-8);(B and D) serum creatinine and urea [(circles, creatinine (μmol/L); triangles, urea (mmol/L)].

Full size image

The highest chemokine levels reached during the course of the disease are shown in Figure 4 and Table 2.Figure 4 depicts maximum urinary chemokine levels of 15 individual HUS children, demonstrating impressive increases of urinary MCP-1 and IL-8 in HUS children compared with control children (p < 0.001). HUS children with severe renal failure accompanied by anuria showed higher maximum concentrations of both chemokines than those HUS patients with mild renal failure without anuria (Fig. 2). This effect may be enhanced by accumulation during the anuric phase. HUS children with extrarenal manifestations of HUS demonstrated higher, though not significantly higher, maximum urinary chemokine levels than children without extrarenal manifestations. No significant differences in maximum urinary chemokine concentrations were observed between children with normal renal function and those with reduced renal function 2 y after onset of the disease (data not shown). Maximum urinary MCP-1 and IL-8 levels of all HUS patients were significantly correlated showing a Spearman's rank correlation coefficient of 0.64 (p = 0.01). Among the 19 children with glomerulopathy were patients with diseases commonly associated with glomerular inflammation as well as patients with characteristically noninflammatory glomerular lesions. As demonstrated in Figure 4A children with glomerulopathy had significantly elevated levels of MCP-1 compared with healthy control children (p < 0.001). Only a minority of these children had raised levels of urinary IL-8 (Fig. 4B). Overall, urinary IL-8 levels were not significantly elevated compared with healthy control children (p = 0.31). When maximum urinary chemokine values of HUS children and children with glomerulopathy were compared, it became obvious that urinary chemokine levels in acute renal failure patients suffering from HUS were significantly higher than chemokine levels in nonacute renal disease patients (p < 0.001; Fig. 4 and Table 2).

Figure 4
figure 4

Urinary chemokine levels in healthy control children(1), children with a variety of inflammatory and noninflammatory nephropathies (2), and HUS children (maximum values (3).(A and B) MCP-1 and IL-8 values (ng/mmol creatinine), respectively. NS, not significant.

Full size image
Table 2 Urinary IL-8 and MCP-1 levels in healthy control children, HUS patients (maximum values), and children with a variety of inflammatory and noninflammatory nephropathies, respectively
Full size table

Serum levels of MCP-1 and IL-8. To evaluate whether the raised levels of urinary chemokines in HUS children represent a local inflammatory process limited to the kidney or are part of a systemic inflammatory response, serum chemokine levels were determined in HUS patients (Fig. 5). Healthy control children (n = 14), children with a variety of glomerulopathies (n = 18), and children treated with CAPD(n = 9) served as control groups. Only low levels of MCP-1 were detectable in serum of healthy control children, showing a median value of 200 pg/mL (mean 200 ± 81; range 79-361 pg/mL, Fig. 5A). Whereas one HUS patient showed a significantly elevated serum MCP-1 concentration (1448 pg/mL), the majority of HUS children displayed only slightly elevated concentrations of MCP-1. Overall, serum MCP-1 was significantly higher in HUS children compared with serum MCP-1 in healthy control children (median 270, mean 342 ± 331, range 115-1448 pg/mL,p < 0.05). MCP-1 levels in patients suffering from a variety of inflammatory and noninflammatory renal diseases were not significantly increased (median 208, mean 227 ± 108, range <31 to 460 pg/mL). CAPD patients displayed slightly but significantly elevated serum MCP-1 levels(median 300, mean 328 ± 71, range 238-425, p < 0.005) compared with healthy control children.

Figure 5
figure 5

Serum chemokine levels in healthy control children(1), children with a variety of inflammatory and noninflammatory nephropathies (2), HUS children on admission (3), and children treated by CAPD (4). (A and B) MCP-1 and IL-8 values (pg/mL), respectively. The detection limit of the assay is indicated by the dotted line. The median value is indicated by a solid line. NS, not significant.

Full size image

IL-8 was not detectable (<31 pg/mL) in serum of 11 healthy control children. Three other healthy control children showed slightly elevated levels of serum IL-8 (Fig. 5B). In two other control groups consisting of children with nephropathy and children on CAPD, serum IL-8 was not detectable (<31 pg/mL). However, IL-8 was detected and significantly elevated in initial serum samples of 11 out of 14 HUS patients (median 145, mean 1852 ± 5657 pg/mL, range <31 to 21 435, p < 0.005). Correlation analysis of initial urinary and serum IL-8 levels, which were available in 6 HUS patients, revealed that initial urinary and serum IL-8 concentrations were not significantly correlated (p > 0.05).

DISCUSSION

In the present study we have shown that MCP-1 and IL-8, potent and specific chemotactic, and activating factors for MOs and PMNs, respectively, are significantly elevated in urine of HUS patients. The highest chemokine levels were encountered in HUS children suffering from renal failure with anuria. Whereas serum IL-8 concentrations were significantly increased in the majority of HUS patients at admission, serum MCP-1 concentrations were only slightly elevated. Urinary chemokine levels did not correlate with serum chemokine levels. The actual presence of MOs and PMNs in glomeruli of HUS biopsy specimens was demonstrated by histochemical and EM analysis.

Experimental and clinical observations in various forms of kidney diseases have high lightened the importance of interactions of specific leukocyte populations with local renal cells in the disease process [reviewed in Banas et al.(25) and Schlondorff et al.(26)]. In HUS, predominantly the importance of PMN activation and cell-mediated endothelial injury has been described, whereas only recently the importance of MOs in the pathogenesis of HUS has been suggested(16, 17, 19). A significant correlation has been shown between PMN cell count at presentation and poor outcome of the disease(27). HUS-PMNs adhere more avidly to endothelium than control cells and induce endothelial cell injury(28). Recently, Morigi et al.(29) reported that VT-treated human umbilical vein endothelial cells, cultured under physiologic flow conditions, showed an increase of adhering leukocytes, most probably by an up-regulation of leukocyte adhesion molecules. Furthermore,α1-antitrypsin-complexed elastase is raised at presentation(30) and ultrastructural studies on HUS PMNs revealed a reduction in granules in some patients(31). Upon exposure to VT in vitro, MOs produce and release a variety of inflammatory mediators(16, 17) which have been shown to increase the sensitivity of glomerular endothelial cells in vitro for VT cytotoxicity by an up-regulation of the number of specific VT receptors on the cell surface(4, 5). In addition, MOs synthesize and release a variety of other vasoactive constituents, which may directly or indirectly contribute to the renal damage(18). These data, sustained by data on experimental animals(32, 33), indicate that both MOs and PMNs are involved in the pathogenesis of glomerular endothelial cell damage in HUS.

As previously described, we observed that not only PMN-(27) but also MO cell counts(34) are raised in HUS children. The number of patients was too small to observe significant differences in the number of MOs or PMNs between the group children with and without anuria. However, these data should be interpreted with some caution. Because in hemorrhagic colitis, even in the absence of renal disease, PMN counts are raised, increased numbers of PMNs may just reflect the severity of the intestinal infection. Moreover, peripheral white blood counts do not necessarily reflect their potential pathogenic role at a local level.

To evaluate inflammatory processes at a local level, kidney specimens of HUS children were examined for the presence of leukocytes. Whereas the presence of PMNs in renal tissues of HUS patients has been previously reported(35–37), the presence of MOs was unclear. Compared with control renal tissues, HUS biopsies demonstrated increased numbers of MOs and to a lesser extent of PMNs, which seemed to specifically infiltrate the glomerular area. These results indeed suggest that these cells participate in the observed glomerular damage. It is not known whether the influx is a primary event early on in the pathogenic mechanism or secondary to disease processes in the kidney.

The molecular mechanisms responsible for the migration and recruitment of leukocytes comprise a cascade of events, including interactions between molecules expressed on leukocytes, the endothelium, and extracellular matrix. Chemokines appear to play a role in each of these steps(26). Whereas MCP-1 has been characterized as a specific chemoattractant and activating factor for MOs, IL-8 has been characterized as a regulator of PMN cell adhesion, migration, activation and degranulation(21). Thus, chemokines do not only participate in the recruitment of specific subtypes of leukocytes, but also modulate their potential injurious effect. In the present study, we demonstrated that chemokine levels are significantly higher in urine of HUS patients compared with chemokine levels in urine of control children. In the majority of HUS patients, initial urine samples displayed increased levels of both chemokines, suggesting that increased production has occurred. Serial measurements showed either steadily increasing levels of the chemokines during the course of the disease and a gradual decline or a gradual decline until recovery. HUS children with anuria showed significantly higher initial and maximum urinary chemokine levels than HUS children without anuria. Chemokine levels were neither significantly higher in HUS children with extrarenal manifestations of HUS nor in HUS children with poor renal outcome after two years. The correlation of urinary MCP-1 and IL-8 may implicate a general up-regulation of chemokine production and does not favor a distinct time course of expression. Corresponding previous observations(38, 39), raised levels of urinary chemokines were observed in children with a variety of glomerular diseases with and without proteinuria, although to a lesser extent than observed in HUS children suffering from acute renal failure. In general, raised urinary chemokine levels in children with a variety of glomerular diseases were not or only slightly associated with elevations in serum.

In agreement with data reported by Fitzpatrick et al.(10), we observed significantly elevated levels of IL-8 in initial serum samples of the majority of HUS patients. Increased serum levels of IL-8 have also been reported in patients with Crohn's disease or colitis ulcerosa(40), suggesting that circulating IL-8 may just reflect inflammatory processes in the gastrointestinal system. In contrast to initial serum IL-8, initial serum MCP-1 was only slightly elevated compared with serum MCP-1 in healthy control children. Thus, chemokines are predominantly and to a much higher extent increased in urine than serum of HUS children. Furthermore, initial urinary and serum chemokine levels did not significantly correlate. Therefore, it is conceivable that urinary chemokines are produced at a local level and excreted into the urine. This suggestion is strengthened by previous observations reporting that TNF-α and IL-6 are predominantly elevated in urine and not or only slightly elevated in serum of mostly severe cases of HUS(9–14). Moreover, fever is not a common finding in HUS patients.

A variety of renal cells including endothelial(41), mesangial(42), proximal tubular epithelial cells(43), fibroblasts(44), as well as infiltrating leukocytes(45) are reported to release chemokines. Chemokines are not expressed in most resting cells but are rapidly up-regulated on activation, typically induced by proinflammatory mediators such as TNF-α, IL-1β, interferon-γ, and lipopolysaccharide. It remains to be investigated whether VT itself directly induces chemokine production in renal cells. In vitro, mesangial cells have been shown not to respond to VT in terms of chemokine production(24). The biological effects of chemokines at a local level can be prolonged because chemokines are basic proteins which bind strongly to heparan sulfate proteoglycans on cell surfaces or to extracellular matrix.

The present data suggest an important local role for MOs and PMNs in the pathologic process of glomerular endothelial cell damage. The chemokines, MCP-1 and IL-8 may possibly be implicated in the pathogenesis of HUS through the recruitment and activation of MOs and PMNs, respectively, in the kidney. These observations may contribute to the understanding of the pathogenesis of HUS and raise the question whether therapeutic interventions (reviewed inRefs. 25, 26, and 46) are of interest to study in the future. It remains to be elucidated which factor(s) initiate(s) the inflammatory response in the glomeruli of HUS patients.