If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Address correspondence to Maurizio Brigotti, Ph.D., Department of Experimental, Diagnostic and Specialty Medicine, University of Bologna, School of Medicine, General Pathology Bldg., Via San Giacomo 14, Bologna 40126, Italy.
This review focuses on typical hemolytic uremic syndrome (HUS), a life-threatening sequela of human infections caused, particularly in children, by Shiga toxin–producing Escherichia coli strains. Thrombotic microangiopathy of the brain and the kidney is the end point of toxin action, resulting in the hallmarks of HUS (ie, thrombocytopenia, anemia, and acute renal failure). A growing body of evidence points to the role of extracellular vesicles released in the blood of patients by toxin-challenged circulating cells (monocytes, neutrophils, and erythrocytes) and platelets, as a key factor in the pathogenesis of HUS. This review provides i) an updated description of the pathogenesis of Shiga toxin–producing E. coli infections; ii) an analysis of blood cell–derived extracellular vesicles, and of their parent cells, as triggering factors in HUS; and iii) a model explaining why Shiga toxin–containing vesicles dock preferentially to the endothelia of target organs.
The study of the pathogenesis of a given disease outlines the triggering factors and describes the chain of events involved in the development of the pathologic condition. In the case of infectious diseases, the identification of the microbial pathogenic factors and of the resulting host-pathogen interactions clarifies the mechanisms underlying the pathogenetic process. In many cases, however, a particular disease is the consequence of the complex interactions of several virulence factors, whereby the illness may not occur if one of these factors is removed or if their interplay is impaired. This review focuses on typical hemolytic uremic syndrome (HUS), a life-threatening condition caused by pathogenic bacteria known as Shiga toxin–producing Escherichia coli (STEC), which release potent exotoxin, named Shiga toxins (Stx).
These Stx-independent injuries, known as attaching and effacing lesions, are critical for the subsequent pathogenetic steps, such as the synthesis and release of Stx, which cross the intestinal mucosa to reach the lamina propria.
Bloody diarrhea is the consequence of toxin actions on the endothelial lining of the intestine, culminating in typical histopathologic changes, such as mucosal and submucosal edema, hemorrhage, focal necrosis, and thrombotic microangiopathy.
Lesions to the microvasculature of the gut trigger bloody diarrhea during precocious toxemia, whereas HUS develops 5 to 7 days later when the endothelial cells of the kidney and brain are targeted by the toxin.
Endothelial damage/dysfunction related to Stx action causing thrombotic microangiopathic lesions in target organs is considered the major pathogenetic event in HUS and occurs when Stx cross the epithelial lining of the large bowel and are transported in circulation. These injuries narrow renal glomerular capillaries, thereby damaging passing erythrocytes, and consume platelets; therefore, acute renal failure, thrombocytopenia, and anemia ensue.
Only a few of these have been found in STEC isolates associated with severe human illness [eg, Stx1a (bloody diarrhea and rarely HUS) or Stx2a, mucus-activatable Stx2d, and the controversial Stx2c (bloody diarrhea and HUS)].
The release of extracellular vesicles containing Stx (and other virulence factors) by these stimulated cells is considered the trigger producing the transition from bloody diarrhea to life-threatening HUS.
The multiple interactions of Stx with circulating cells, and with target epithelial and endothelial cells, induce the release of proinflammatory cytokines and chemokines capable of up-regulating the expression of specific toxin receptors (see the next paragraph), of exacerbating tissue damage in targeted organs, and of recruiting inflammatory cells in injured tissues.
Stx are bipartite exotoxins composed of a single A chain, which is endowed with the enzymatic activity, noncovalently bound to five B chains forming a pentameric ring that surrounds the COOH-terminus of the A subunit.
The latter is a proenzyme that is enzymatically cleaved into two fragments (A1 and A2) linked by a disulfide bond that is reduced within cells, permitting the A1 fragment to express its deadenylating activity on 28S rRNA in ribosomes and on DNA in chromatin, leading to irreversible arrest of translation and to formation of nuclear apurinic sites, respectively.
Stx-intoxicated endothelial cells not only have impaired translation, but also activate a wide array of responses to injury, which have a great impact in the pathogenesis of HUS (eg, release of proinflammatory cytokines,
Receptor-dependent toxicity is a key concept explaining why these powerful toxins cause damages to selected tissues in animals and humans. The B pentamers of the human subtypes Stx1a and Stx2a show a specific and preferential binding to globotriaosylceramide (Gb3Cer), a member of globoseries glycosphingolipids.
The focused expression of Gb3Cer by intestinal, renal, and cerebral endothelial cells and by other cells in the kidney (mesangial cells and tubular and glomerular epithelial cells) allows the toxin to target these organs preferentially by means of multivalent interactions of the glycolipid receptors with B subunits.
Therefore, a similar protecting interaction of Stx2 with the extracellular domain of membrane-anchored TLR4 can be expected. The implications of these Stx2-modulating activities on the pathogenesis of HUS are discussed in the following paragraphs.
Extracellular Vesicles Bearing Stx and Their Parent Cells
In STEC-infected patients, Stx come in contact with and bind to platelets
This impacts the formation of circulating platelet-monocyte and platelet-neutrophil aggregates containing Stx, as observed in children with HUS during the acute phase of disease, but not after recovery.
Under normal conditions, platelets and leukocytes do not interact in the circulation. Hence, the formation of aggregates between these blood components in STEC-infected patients could play an important role in thrombogenesis and inflammation.
The second more important consequence is the generation of blood cell–derived extracellular vesicles bearing Stx detected in plasma of patients during overt HUS by flow cytometry and visualized within renal endothelial cells by electron microscopy.
The toxins within or on the surface of host blood cell–derived extracellular vesicles can be transferred to target endothelial cells, as shown in in vitro studies (ie, extracellular vesicles containing Stx and derived from human platelets and leukocytes undergo endocytosis in human glomerular endothelial cells and release the toxin, leading to inhibition of protein synthesis and cell death).
This is an important point because this technique allows direct visual inspection of the vesicles, giving a more accurate measure of the size compared with other methods, such as flow cytometry or nanoparticle tracking.
A weak point of transmission electron microscopy is that it is difficult to ascertain whether the detected vesicles are typical or rare structures. Therefore, it is recommended to supplement transmission electron microscopy with an additional method, allowing detection and analysis of the whole vesicle population.
Besides apoptotic bodies, the other known extracellular vesicles may play a pivotal role in intercellular communication, providing receptors and soluble molecules (mRNA, lipids, and proteins) to recipient cells.
An analysis of the features of the extracellular vesicles found in HUS patients and of the circulating parent cells that release them is crucial to understand their contribution in the pathogenesis of HUS. In a 2-year–old child, platelets (7.9 fL),
greatly differ in size. Considering the size of HUS-related rounded 1-μm–diameter microvesicles (0.52 fL), the volumetric ratio of microvesicle/parent cell is roughly in the range 1:150 to 1:700 for all of the cells while platelets show a 1:15 ratio. Therefore, the number of extracellular vesicles potentially released by a single thrombocyte appears to be limited by the small size of these cell fragments deriving from megakaryocytes. Moreover, the different blood half-lives of erythrocytes (months), platelets (few days), classic monocytes (1 day), and neutrophils (few hours) also affect the daily availability of new vesicle-generating cells entering in circulation from bone marrow.
to the number of vesicle-generating cells is maximal for monocytes (0.4 to 1.3 vesicles per cell), intermediate for neutrophils (4 to 22 × 10−2 vesicles per cell), low for platelets (2.7 to 3.1 × 10−3 vesicles per cell), and extremely low for erythrocytes (7 to 9 × 10−5 vesicles per cell). Only a small percentage of microvesicles are Stx2-positive [ie, approximately 25% to 30%, except erythrocytes (9%)].
The steady-state blood count of extracellular vesicles is determined by a balance between their biogenesis and their delivery to target cells. Nevertheless, when whole blood samples from healthy donors (normal adult blood count ranges; Royal Wolverhampton Trust Pathology Services, https://www.royalwolverhampton.nhs.uk/services/service-directory-a-z/pathology-services/departments/haematology/haematology-normal-adult-reference-ranges, last accessed February19, 2021) or isolated red cells were treated in vitro with Stx2 in the absence of recipient target cells, similar values were obtained (monocytes, 0.3 to 1.1 vesicles per cell; neutrophils, 2 to 9 × 10−2 vesicles per cell; platelets, 1.8 to 5.5 × 10−3 vesicles per cell; erythrocytes, 1 × 10−4 vesicles per cell). Therefore, although a number of platelet-derived microvesicles prevail in patients' blood during overt HUS, the efficiency of vesicular shedding shown by these cell fragments is lower than that of monocytes and neutrophils. The same ranking in vesicle generation efficiency among human circulating cells was found when unstimulated human blood was analyzed by imaging flow cytometry (ImageStream, Amnis Corporation, Seattle, WA) after simultaneous staining of circulating cells and cell-derived vesicles,
excluding any interference caused by microvesicle isolation procedures in the data obtained with patients' blood. In conclusion, Stx2 enhances the basal level of microvesicle release by circulating cells but does not selectively change the vesicle-generating ability of a specific circulating cell. Besides Gb3Cer, TLR4 appears to contribute to vesicle generation by interacting with Stx, as demonstrated in Gb3Cer-lacking human neutrophils.
Extracellular Vesicles Bearing Stx Contain Further Virulence Factors Involved in the Pathogenesis of HUS
Leukocyte-platelet aggregates and microvesicles involved in HUS contain additional pathogenic factors apart from Stx. The deposition of C3 and C9 on the surface of circulating blood cell–derived microvesicles, originated from platelets, monocytes, and neutrophils, was observed in the plasma of 12 HUS patients in the acute phase.
These findings have been confirmed in experimental models with whole blood stimulated with Stx and/or STEC-lipopolysaccharide, showing that these virulence factors induced the deposition of surface-bound C3 and C9 on platelet-leukocyte aggregates and derived microvesicles.
In whole blood from four children with HUS, circulating platelet-monocyte and platelet-neutrophil aggregates expressing surface-bound tissue factor (TF) were observed during the acute phase of disease but not after recovery.
The contribution of complement factors and TF to the thrombotic microangiopathy of HUS is clear-cut. Activated complement factors fixed on the surface of extracellular vesicles can be transferred to endothelial target cells, causing injuries. Stx2 activates complement alternative pathway in the fluid phase and binds to factor H, thereby reducing its protective effect on cellular membranes.
TF carried by extracellular vesicles and delivered to target endothelial cells can bind and activate coagulation factor VII, leading to a cascade of reactions culminating in thrombin generation and thus moving the scale toward a prothrombotic state.
Pathotype of HUS-Triggering Extracellular Vesicles
An emerging concept in microbiology of STEC infections is that of pathotype. To describe strains commonly associated to HUS, it is recommended to refer to the cocktail of virulence factors involved in pathogenesis rather than to the E. coli serogroup, such as O157 or O26. It is becoming clear that the specific host clinical symptoms depend on the peculiar combinations of virulence factors (pathotype) produced by a specific STEC strain as well as on their allelic types.
Likewise, extracellular vesicles involved in the pathogenesis of HUS should not be considered, based on the type of generating cells, as neutrophil-, monocyte-, or platelet-derived microvesicles nor based on the absolute blood count. Paradoxically, a specific cell-derived microvesicle can have the highest blood count in HUS patients because of delivery systems for target cells. On the other hand, a sparsely represented circulating microvesicle can be promptly cleared from the blood by interactions with target cells. Although the parent cell may confer some of its features to the microvesicles, the specific combination of virulence factors (pathotype) present in the microvesicle/s mainly responsible for the transition from bloody diarrhea to HUS in STEC-infected patients should be investigated. HUS-triggering microvesicles are expected to bear TF, complement factors, and Stx2 on their surface and within the vesicle as toxic cargo. It is conceivable that the most poisoning HUS-triggering extracellular vesicle/s would have a golden ratio among the different necessary and interactive virulence factors, allowing targeting and full expression of the pathologic effects.
Is Cerebral Endothelium Differently Targeted in HUS?
Central nervous system involvement in STEC-infected patients and the role of Stx in this process are intriguing. Neurologic symptoms are more closely related to the direct action of Stx in the brain rather than to the consequences of renal failure (uremic encephalopathy), as demonstrated by their appearance before overt HUS in 15% of patients.
The underlying mechanisms could be thrombotic microangiopathy in the brain (weak evidence), direct toxic action of Stx2 on neural cells after blood-brain barrier disruption, or indirect inflammatory-mediated damage to the brain.
whereas HUS ensues in STEC-infected patients during the toxin-decreasing phase on circulating cells, as demonstrated by the reduced toxic cargo on neutrophils in patients with the most severe forms of HUS.
It can be hypothesized that higher amounts of blood Stx are necessary to induce sudden brain endothelial injuries before the toxin-decreasing phase on circulating cells or that the mechanisms of endothelial damage by Stx are not related to microvesicle release by circulating cells. This interesting topic needs to be further investigated, keeping in mind that fenestrated endothelial cells (renal glomeruli) and those composing the robust blood-brain barrier (brain) differ considerably.
The Observation of STEC-Infected Patients before HUS Provides a Higher Vantage Point
The important observations described in the previous paragraphs on the role of extracellular vesicles in the pathogenesis of HUS have been obtained after the diagnosis of HUS.
In this setting, the recognized virulence factors act during the evolution of the pathogenetic process, before the end point (HUS). Contrary to studies on patients in the acute phase of HUS in whom free Stx are rarely found in serum,
STEC-infected patients with bloody diarrhea show free Stx in blood detected by enzyme-linked immunosorbent assay (at ng/mL concentrations) and by functional assay on human cells (inhibition of protein synthesis).
by analyzing about 100 patients with STEC-associated gastroenteritis, found Stx2 in blood by enzyme-linked immunosorbent assay in patients who developed HUS and in some who recovered. According to other studies on HUS patients,
A quantitative investigation of neutrophil-associated Stx2 obtained by flow cytometry (0.2 to 1.7 × 10−9 mol/L, see equation below) or of cell-free circulating Stx2 detected by enzyme-linked immunosorbent assay (3.3 to 8.9 × 10−11 mol/L)
Kinetic analyses shows that the toxin amounts in circulating leukocytes decrease over time with an increase in the concentration of functionally active Stx in patients' sera, as well as that of the soluble decoy TLR4.
over cell-free Stx2 (see above), the toxicity found in patients' sera can be ascribed to Stx2/TLR4 complexes or to microvesicle-associated Stx2. Furthermore, detachment of Stx from leukocytes was also observed in patients who recovered (even showing proteinuria); thereby, this process, although representing the first step in the delivery to target organs, is not sufficient per se to trigger HUS.
The distinctive feature observed in patients who developed HUS was the sudden appearance of the vesicular form of Stx2 detected by enzyme-linked immunosorbent assay (capturing antibody against B-chains) in blood the day before the onset of HUS during the decreasing phase of toxin on leukocytes.
As with antigens, vesicular toxin was obtained by centrifuging patients' sera at the g-force required to isolate 1-μm microvesicles and was labeled particulate toxin in contrast to free soluble serum toxin. In these patients, the particulate toxin peak was accompanied by a corresponding peak of particulate TLR4 (microvesicle-associated TLR4), having a stoichiometry ratio with particulate Stx2 close to 1.
The amount of Stx associated with the circulating cells of the patients can decrease over time because of direct transfer of the toxic cargo to Gb3Cer-expressing recipient cells, as observed experimentally during neutrophil transmigration across endothelial monolayers
or release of free Stx (cleared by HuSAP), Stx-TLR4 complex (escaping HuSAP), or microvesicle-associated Stx (within vesicles or Gb3Cer and TLR4 bound). Given the low vesicle generation efficiency of circulating cells (maximal for monocytes, approximately 1 vesicle per cell) and the small amount of microvesicles found in patients' blood,
it is unlikely that a sudden decrease in the circulating toxic reservoir could be ascribed to vesicle generation alone. Despite this, there is a good correlation in patients between the toxic activity (translation inhibition in cells) of serum Stx2 and the amount of particulate toxin, but not of whole serum toxin (free toxin + particulate toxin).
A Receptor-Based Model for the Specific Targeting of Gb3Cer-Expressing Endothelial Cells by Stx2-Containing Microvesicles
The role of microvesicles in the pathogenesis of HUS was first explained by the specific binding of Stx2 to circulating monocytes, platelets, and red cells through Gb3Cer followed by a release phase and a targeting phase.
In this model, the toxic cargo, membrane-bound and internalized Stx2, and the associated pathogenic factors, TF and activated complement components, reach the target cells through non–receptor-mediated endocytosis or fusion. Both uptake mechanisms are largely unspecific; therefore, this model failed to explain why the histopathologic lesions observed in HUS patients are focused on few organs and tissues whose cells express Gb3Cer.
HUS is not considered a multi-organ disease. On the other hand, non–receptor-mediated uptake mechanisms have been invoked to explain why other organs apart from intestine, kidney, and brain are affected, albeit rarely, by the thrombotic microangiopathy (eg, pancreas, musculoskeletal system, and heart).
In this case, the specific targeting represents the main advantage. Stx associated with microvesicles are primarily localized within the microvesicles because permeabilization is required for maximal detection
This means that at least part of the toxin is surface bound. It is worth noting that Stx interplay with circulating cells is mediated by two different receptors (Gb3Cer and TLR4), which specifically engage two different parts of the bipartite AB5 toxin (Figure 1).
If Stx are bound to microvesicle surface through the glycolipid receptor, their B chains are engaged and presumably the A chain is exposed. Conversely, when the toxins are bound through TLR4 via A chain, the B subunit pentamer is available for specific binding to target cells. Typically microvesicles captured by B-chain specific antibodies are the ones found in the blood of patients the day before the fulfillment of the criteria for HUS diagnosis.
These findings are sustained by results obtained in vitro, suggesting that uptake of Stx2-positive microvesicles is not sufficient to induce cell death and that the recipient cell needs to possess Gb3Cer for the the toxin to exert its toxic effect.
The binding of Stx2 to membrane-associated TLR4 on microvesicles would prevent the interaction of the toxin to HuSAP, a specific inhibitor of the Gb3Cer-dependent toxicity of Stx2, hence favoring the docking to Gb3Cer-expressing renal endothelial cells.
The features of the microvesicles derived from platelets or monocytes are similar because these cells are endowed with both Stx2-interacting receptors (Figure 2). This means that both TLR4-anchored B pentamer–exposing Stx2 and Gb3Cer-anchored A chain–exposing Stx2 are present on the vesicle surface. Renal endothelial cells also express both these receptors, and when they are targeted by soluble Stx in vitro, TLR4 acts as coreceptor, hence cooperating with Gb3Cer during binding and intoxication.
In this light, the presence of surface-bound Stx2 exposing A or B chains alternatively would confer an advantage to the microvesicle in intoxicating the target cell. Finally, microvesicles derived from neutrophils expose on their surface only the TLR4-anchored B pentamer–exposing Stx2.
In conclusion, although the pathogenic cargo depends on the presence, amount, and specific proportion of the different virulence factors associated with the microvesicle, the delivering process is likely to be determined by the presence and spatial orientation of Stx2 exposed on the vesicle membrane. Similar to delivering a package, the delivery of Stx2 to target organs through microvesicles requires a stamp, a price (binding to blood cells); an address, (TLR4-anchored B pentamer–exposing Stx2); and a post box, that is Gb3Cer on target cells. Indeed, Stx are toxic components of the package as well as the leading factors necessary for the docking of the other virulence factors to renal endothelial cells.
This review focuses on the involvement of blood cell– or platelet-derived microvesicles in early pathogenesis of HUS (ie, the initial interactions of virulence factors, including Stx, with renal endothelial cells). In vitro experiments point to the involvement in HUS not only of microvesicle but also of Stx-containing exosomes released by target cells (Vero or HeLa cells) or innate immune cells (macrophage-like differentiated THP-1 cells) after toxin internalization and intracellular processing.
If confirmed in HUS patients, this might be an important mechanism for toxin spreading once the first cells have been targeted through the microvesicle-centered mechanisms reviewed herein. Moreover, Stx-containing exosomes might confer to adjacent nonintoxicated cells not only the toxic cargo but also ancillary factors that exacerbate tissue damage, as mRNAs coding for proinflammatory cytokine and chemokines.
Future research perspectives would include the definition of the pathotype of Stx2-bearing HUS-triggering extracellular vesicles (microvesicles and exosomes) and a direct demonstration of their involvement in the targeting of cerebral endothelia. These results will stimulate the development of new strategies for therapeutic interventions aimed at preventing the transition from bloody diarrhea to HUS in STEC-infected children. Blocking the interactions of Stx with circulating cells or the release of the toxic cargo from the circulating or target cells reservoirs by impairing the budding of pathogenic microvesicles and the release of Stx-containing exosomes would achieve this goal. Glycovesicles spiked with Gb3 neoglycolipids have been developed, which inhibit the binding of Stx to Gb3Cer.