Soluble Toll-Like Receptor 4 Impairs the Interaction of Shiga Toxin 2a with Human Serum Amyloid P Component

Shiga toxin 2a (Stx2a) is the main virulence factor produced by pathogenic Escherichia coli strains (Stx-producing E. coli, STEC) responsible for hemorrhagic colitis and the life-threatening sequela hemolytic uremic syndrome in children. The toxin released in the intestine by STEC targets the globotriaosylceramide receptor (Gb3Cer) present on the endothelial cells of the brain and the kidney after a transient blood phase during which Stx2a interacts with blood components, such as neutrophils, which, conversely, recognize Stx through Toll-like receptor 4 (TLR4). Among non-cellular blood constituents, human amyloid P component (HuSAP) is considered a negative modulating factor that specifically binds Stx2a and impairs its toxic action. Here, we show that the soluble extracellular domain of TLR4 inhibits the binding of Stx2a to neutrophils, assessed by indirect flow cytometric analysis. Moreover, by using human sensitive Gb3Cer-expressing cells (Raji cells) we found that the complex Stx2a/soluble TLR4 escaped from capture by HuSAP allowing the toxin to target and damage human cells, as assayed by measuring translation inhibition, the typical Stx-induced functional impairment. Thus, soluble TLR4 stood out as a positive modulating factor for Stx2a. In the paper, these findings have been discussed in the context of the pathogenesis of hemolytic uremic syndrome.


Introduction
Shiga toxins (Stx) are powerful exotoxins involved in the pathogenesis of severe illnesses in children, such as hemorrhagic colitis and hemolytic uremic syndrome (HUS), related to infections by pathogenic Escherichia coli strains (Stx-producing E. coli, STEC) [1][2][3]. These bacteria colonize the human gut after ingestion of contaminated food and intimately adhere to the epithelial lining of the

Effect of the Soluble Recombinant Extracellular Domain of TLR4 on the Binding of Shiga Toxins to Neutrophils
Human neutrophils from healthy donors were treated with 60 nM Stx1a and Stx2a to reach full saturation of receptors [31]. The presence of 10-fold excess of 70 kDa sTLR4 inhibited the interaction of both toxins with neutrophils ( Figure 1), further confirming the identity of TLR4 as a Stx receptor. The presence of the co-receptor MD2 did not significantly modify the inhibitory power of the 70 kDa sTLR4. The effect of 70 kDa sTLR4 on the binding of Stx to neutrophils is specific since the decoy form of TLR2 [24] was not able to confer the protection.
Toxins 2018, 10, x FOR PEER REVIEW 3 of 9 sTLR4. The effect of 70 kDa sTLR4 on the binding of Stx to neutrophils is specific since the decoy form of TLR2 [24] was not able to confer the protection. with Stx1a or Stx2a were (mean ± SD, n = 3) 4.35 ± 0.14 or 3.90 ± 0.33, respectively. The data shown in the figure are expressed as binding percentages (mean ± SD, n = 3, unpaired t-test) *** p < 0.0001, ** p < 0.001.

Effect of the Soluble Recombinant Extracellular Domain of TLR4 on the Activity of Shiga Toxin 2a on Target Cells
To study the effect of 70 kDa sTLR4 on the biological activity of Stx2a, we used Stx-sensitive human cells that express Gb3Cer, namely Raji cells. According to a previously described method [32], cells were treated for a short time (3 h) with Stx, and then cell translation was measured after the addition of a radioactive amino acid. The extent of protein synthesis inhibition with respect to control untreated cells was determined. The presence of 70 kDa sTLR4 (5 μg/mL) or of the complex 70 kDa sTLR4/MD2 (5 μg/mL) did not modify the IC50 of Stx2a on protein synthesis in Raji cells ( Table 1), suggesting that a putative complex between Stx2a and cell-free soluble TLR4 is active in intoxicating target cells. Conversely, it is well known that Stx2a, but not Stx1a, acting on the same cellular model was greatly impaired by the presence of human serum containing HuSAP [32]. The effect of HuSAP on Stx2a is dramatic since even low concentration of purified HuSAP (10 nM) significantly impaired the cytotoxicity of Stx2a on Raji cells (Table 1). To verify if the binding of soluble TLR4 to Stx2a allows the toxin to escape from the capture by HuSAP, we pre-incubated Stx2a (0.5 nM) with a 1000 molar excess of both factors in different combinations and, after complexes formation, aliquots of the mixture were withdrawn and added to Raji cells to obtain a concentration of Stx2a (10 pM) approximately equal to its IC80, the inhibitor concentration that decreases translation by 80%. The strong protection induced by HuSAP was significantly alleviated by pre-incubating Stx2a with 70 kDa sTLR4 ( Figure 2). Moreover, the simultaneous addition of HuSAP, 70 kDa sTLR4, and Stx2a to the assay allowed similar effects ( Figure 2). Thus, the soluble form of TLR4 may impair HuSAP protection by creating physical hindrance for binding between HuSAP and Stx2a. It should be noted that if the opposite experiment was performed, i.e. pre-incubation of the toxin with HuSAP and then addition of 70 kDa sTLR4 to the assay, no action of the toxin was permitted. Hence, soluble TLR4 is not able to displace the toxin from the HuSAP/Stx2a complex. One can conclude that the simple release of soluble TLR4 in patients′ sera after monocyte/neutrophil were (mean ± SD, n = 3) 4.35 ± 0.14 or 3.90 ± 0.33, respectively. The data shown in the figure are expressed as binding percentages (mean ± SD, n = 3, unpaired t-test) *** p < 0.0001, ** p < 0.001.

Effect of the Soluble Recombinant Extracellular Domain of TLR4 on the Activity of Shiga Toxin 2a on Target Cells
To study the effect of 70 kDa sTLR4 on the biological activity of Stx2a, we used Stx-sensitive human cells that express Gb3Cer, namely Raji cells. According to a previously described method [32], cells were treated for a short time (3 h) with Stx, and then cell translation was measured after the addition of a radioactive amino acid. The extent of protein synthesis inhibition with respect to control untreated cells was determined. The presence of 70 kDa sTLR4 (5 µg/mL) or of the complex 70 kDa sTLR4/MD2 (5 µg/mL) did not modify the IC 50 of Stx2a on protein synthesis in Raji cells ( Table 1), suggesting that a putative complex between Stx2a and cell-free soluble TLR4 is active in intoxicating target cells. Conversely, it is well known that Stx2a, but not Stx1a, acting on the same cellular model was greatly impaired by the presence of human serum containing HuSAP [32]. The effect of HuSAP on Stx2a is dramatic since even low concentration of purified HuSAP (10 nM) significantly impaired the cytotoxicity of Stx2a on Raji cells (Table 1). To verify if the binding of soluble TLR4 to Stx2a allows the toxin to escape from the capture by HuSAP, we pre-incubated Stx2a (0.5 nM) with a 1000 molar excess of both factors in different combinations and, after complexes formation, aliquots of the mixture were withdrawn and added to Raji cells to obtain a concentration of Stx2a (10 pM) approximately equal to its IC 80 , the inhibitor concentration that decreases translation by 80%. The strong protection induced by HuSAP was significantly alleviated by pre-incubating Stx2a with 70 kDa sTLR4 ( Figure 2). Moreover, the simultaneous addition of HuSAP, 70 kDa sTLR4, and Stx2a to the assay allowed similar effects ( Figure 2). Thus, the soluble form of TLR4 may impair HuSAP protection by creating physical hindrance for binding between HuSAP and Stx2a. It should be noted that if the opposite experiment was performed, i.e. pre-incubation of the toxin with HuSAP and then addition of 70 kDa sTLR4 to the assay, no action of the toxin was permitted. Hence, soluble TLR4 is not able to displace the toxin from the HuSAP/Stx2a complex. One can conclude that the simple release of soluble TLR4 in patients sera after monocyte/neutrophil challenge induced by LPS or Stx would have no effect once the binding of HuSAP with Stx2a had occurred. On the other hand, soluble TLR4 could interfere with HuSAP binding to freshly entered Stx2a in blood stream from the intestine. Although these experiments allowed us to conceptualize the reciprocal relationship between Stx2a and protective and/or inhibitory factors, they have been performed in conditions far from those observed in human sera. Thus, we calculated the IC 50 of Stx2a on Raji cell translation in the presence of normal human serum. It was shown that simultaneous addition of Stx2a and high concentration of 70 kDa sTLR4 (10 nM) to Raji cells containing 10% human serum had no effect (Table 1) on IC 50 compared with Raji cells without adding soluble TLR4. Conversely, the addition of Stx2a pre-incubated with 70 kDa sTLR4 significantly decreased the IC 50 (Table 1). Therefore, a substantial increase of Stx2a toxicity would be achieved if TLR4/Stx2a complex is formed before Stx2a interacts with HuSAP. challenge induced by LPS or Stx would have no effect once the binding of HuSAP with Stx2a had occurred. On the other hand, soluble TLR4 could interfere with HuSAP binding to freshly entered Stx2a in blood stream from the intestine. Although these experiments allowed us to conceptualize the reciprocal relationship between Stx2a and protective and/or inhibitory factors, they have been performed in conditions far from those observed in human sera. Thus, we calculated the IC50 of Stx2a on Raji cell translation in the presence of normal human serum. It was shown that simultaneous addition of Stx2a and high concentration of 70 kDa sTLR4 (10 nM) to Raji cells containing 10% human serum had no effect (Table 1) on IC50 compared with Raji cells without adding soluble TLR4.
Conversely, the addition of Stx2a pre-incubated with 70 kDa sTLR4 significantly decreased the IC50 (Table 1). Therefore, a substantial increase of Stx2a toxicity would be achieved if TLR4/Stx2a complex is formed before Stx2a interacts with HuSAP.

Discussion
In this paper the effects of the extracellular soluble domain of TLR4 on the biological activity of Stx and Stx interaction with blood components (neutrophils, HuSAP) involved in the pathogenesis of HUS have been investigated. The soluble form of TLR4 was found to impair neutrophil/Stx interactions. However, high concentrations of 70 kDa sTLR4 were required to achieve this negative modulating function: 600 nM (~40 µg/mL) when the receptors on neutrophils are 100% saturated (Figure 1) or 60 nM (~4 µg/mL) when the receptors are 50% saturated (data not shown). These high concentrations of sTLR4 appear well beyond the capability of circulating monocytes or neutrophils because the maximal concentrations of the top mediators induced by Stx in these cells usually range from~1 ng/mL (CXCL8, neutrophils) to~30 ng/mL (CXCL8, monocytes) [5]. Moreover, 60 nM 70 kDa sTLR4 corresponds to 3.6 × 10 13 molecules per milliliter, while the average numbers of neutrophils and monocytes in a two-year-old child's blood are about 3.5 × 10 6 and 5 × 10 5 per milliliter [33], respectively, suggesting each cell has to release 1-7 × 10 7 sTLR4 molecules to reach the needed level, which is highly unlikely. Based on preliminary ELISA on sera from patients with STEC-induced HUS the concentration of the soluble form of TLR4 was nanogram per milliliter level (data not shown). Thus, the hypothesis that the soluble TLR4 would inhibit the binding of freshly entered Stx in blood from the gut to neutrophils seems unlikely during the natural course of STEC-infections.
Our results demonstrated that cell-free TLR4 protected Stx2a from capturing by HuSAP in blood that inhibits Stx2a interaction with Gb3Cer on target cells. Once a Stx2a/HuSAP complex is formed, soluble TLR4 is no longer able to displace HuSAP. Conversely, formation of Stx2a/70 kDa sTLR4 complex allows the toxin to bind Gb3Cer. According to calculation of the reciprocal interactions between these factors based on the physiological concentrations of HuSAP and on experiments with human serum as source of HuSAP, it is also unlikely that soluble TLR4 would be able to prevent the formation of the Stx2a/HuSAP complex in patients' blood.
A possible explanation for Stx2a forming complex with sTLR4 before interacting with HuSAP is that certain amounts of soluble TLR4 could be produced during the intestinal phase of STEC-infections by LPS-or Stx2a-stimulated macrophages resident in the lamina propria of the gut. In this case, the toxin would translocate through the intestinal barrier as Stx2a/sTLR4 complex.
Another attractive hypothesis is that detachment of Stx2a already bound with TLR4 in circulating cells would allow Stx2a to escape from HuSAP negative modulating effect. This could happen (1) after proteolytic processing of the membrane TLR4 originating macromolecular complexes Stx2a/sTLR4 or (2) considering Stx2a as a membrane component bound to TLR4 present on microvesicles released by activated circulating cells after toxin challenge. It is worth noting that one of the most recent hypotheses on the pathogenesis of STEC-induced HUS proposed that Stx-containing microvesicles released by circulating cells may be the main trigger. In this model, circulating cells represent a reservoir of toxins available for the follow-up intoxication of the renal target cells after their release. At a molecular level, the binding of the extracellular domain of TLR4 exposed on microvesicles to Stx2a could protect toxin from binding by HuSAP, hence favoring its toxicity for target endothelial cells. As shown in Figure 2 and Table 1, the toxicity of Stx2a rescued by soluble TLR4 is only partial, however, over a sufficiently long period of time, this toxicity could accumulate to a level high enough to cause diseases. The role of blood cell-derived Stx-containing microvesicles and the relevance of the soluble form of TLR4 in HUS needs to be further validated in studies on patients during the natural course of STEC-infections.

Experimental Binding of Shiga Toxins to Human Neutrophils
Highly pure neutrophils (99.7%) were isolated under endotoxin-low conditions from buffy coats of healthy donors after centrifugation over Ficoll-Paque, followed by dextran sedimentation, hypotonic lysis of contaminating erythrocytes, and removing of any eventual contaminating cells by using the EasySep human neutrophil enrichment kit (Stemcell Technologies, Vancouver, BC, Canada) [11,36]. Healthy donors gave their informed consent for the use of their blood samples in the experiments described in this paper. For binding experiments with purified Stx1a and Stx2a, Eppendorf tubes were precoated with PBS containing 1% endotoxin-low (≤1 endotoxin unit/mg) BSA to avoid nonspecific loss of the toxins. Neutrophils (5 × 10 5 ) were incubated for 90 min at 37 • C with 60 nM Stx in 250 µL of the same buffer (PBS-BSA). After incubation, the cells were spun down at 200× g for 5 min and washed three times with 100 µL incubation buffer at 37 • C. Stx bound to neutrophils were detected by flow cytometry as describe in the paragraph below.

Detection of Shiga Toxins Bound to Neutrophils
Stx bound to neutrophils were detected by flow cytometry, as reported in a specific methodological article in which all technical details were carefully described [31]. The assay was validated by comparing control subjects and HUS patients in a blinded manner [31] and by challenging Stx-positive neutrophils with a negative control antibody [37]. Briefly, untreated or Stx-treated neutrophils were incubated with mouse monoclonal antibodies (IgG) against Stx1a (Stx1-13C4, Toxin Technology, Sarasota, FL, USA) or Stx2a (Stx2-BB12, Toxin Technology, Sarasota, FL, USA) in the presence of human serum to saturate Fc receptors on leukocytes. After incubation with FITC-goat anti-mouse IgG, flow cytometric analysis was used to reveal the neutrophil-bound fluorescence, allowing a highly sensitive detection of both Stx. Cells were visualized by a cytogram that combined forward scatter versus 90 • side scatter, and fluorescence was analyzed by a cytogram that combined 90 • side scatter and fluorescence and by a single-fluorescence histogram. Neutrophils were checked by staining with mAb to Ags associated to granulocytes (FITC-CD16 and FITC-CD65, Beckman Coulter, Miami, FL, USA). The mean channel value of fluorescence (MCV) was chosen as an objective parameter to measure the extent of binding of Stx to neutrophils.

Detection of the Functional Activity of Shiga Toxins in Whole Cells
The rapid intoxication kinetic of human Gb3Cer-expressing Raji cells and their high sensitivity to Stx were exploited for the detection of the toxic activity of Stx2a by measuring the inhibition of translation [32]. Protein synthesis was measured after 3 h incubation with different concentrations of Stx2a as the rate of incorporation of [ 3 H] leucine into proteins during 60 min incubation in complete medium as described previously [32] in the absence or in the presence of pooled human serum from three healthy donors as source of HuSAP.

Statistical Analysis
Statistical analysis was performed with GraphPad Prism 5 software (GraphPad Software, La Jolla, CA, USA). Continuous variables were described through means and SD. Differences in continuous variables were tested with a t-test after controlling the normality of their distribution. Correlation between variables was assessed using the Pearson correlation coefficient. A p value < 0.05 was considered statistically significant.