Enteric nervous system

FIG. 2. The effects of rotavirus infection on polarized epithelial cells. Data from Jourdan et al (1998), Brunet et al (2000), Lundgren et al (2000) and Obert et al (2000).

in a neuraminidase-independent fashion infect polarized epithelial cells efficiently through both the apical and basolateral surfaces. Rotavirus strains whose infectivity is reduced by treatment of cells with neuraminidase (e.g. RRV, SA11, bovine NCDV, porcine OSU) only infect polarized epithelial cells efficiently through the apical surface (Ciarlet et al 2000a). These results suggest that distinct rotavirus receptors are present on different cell membranes. Rotavirus infection of polarized cells has other effects including disruption of cell interactions and cell integrity, perturbation of cellular protein trafficking, induction of chemokine responses and induction of intestinal secretion (Fig. 2). These outcomes may be differentially affected depending on whether cells are infected only apically or both apically and basolaterally. These newly recognized differences in virus—cell interactions between neuraminidase-sensitive and -resistant rotaviruses may be important for pathogenesis as well as for selection of virus vaccine strains.

How does the rotavirus enterotoxin NSP4 cause disease?

In 1996, the rotavirus non-structural protein, NSP4 was reported to function as a viral enterotoxin (Ball et al 1996). The discovery that NSP4 functions as an enterotoxin was unexpected because no other viral enterotoxins had been described. The remainder of this article briefly reviews the salient features about

TABLE 2 Properties of Rotavirus NSP4 or NSP4 peptide 114-135

Functions in viral morphogenesis; mediating the acquisition of a transient membrane envelope as subviral particles bud into the ER

Mobilizes [Ca2+ ]i release from internal stores (ER)

Associated with virulence based on studies of reassortants

Induces an age-dependent diarrhoea in mice and rats when administered by intraperitoneal or intraluminal routes but not when given intramuscularly

Does not induce histological changes in the intestine when diarrhoea is present

Induces age-dependent Cl— secretion in the intestinal mucosa of young mice

Alters plasma membrane permeability and is cytotoxic to cells

Mutations in NSP4 from virulent/avirulent pairs of virus are associated with altered virus virulence

NSP4 induced age-dependent diarrhoea and age-dependent Cl— permeability changes in mice lacking the CFTR channel

Children make antibodies and cellular immune responses to NSP4

A cleavage product is secreted into the medium of virus-infected cells

NSP4 induces homotypic and heterotypic protection against virus-induced diarrhoea

Exogenous NSP4 alters F-actin organizaton and affects transepithelial resistance in polarized epithelial cells

Used as a agonist to detect novel Ca2+-binding protein in tumour cells

Is a novel toxin? No primary sequence similarity to other known toxins. Is a novel secretory agonist

Hoshino et al (1995)

Ball et al (1996)

Ball et al (1996)

Ball et al (1996)

Tian et al (1994, 1995), Newton et al (1997) Kirkwood et al (1996),

Zhang et al (1998) Morris et al (1999)

Richardson et al (1993), Johansen et al (1999) Zhang et al (2000) Zeng et al (2001)

Tafazoli et al (2001)

this first viral enterotoxin and other pleiotropic properties of NSP4 related to pathogenesis that continue to be discovered (Table 2).

The rotavirus non-structural protein NSP4 was initially identified as a nonstructural glycoprotein with a topology that spans the membrane of the endoplasmic reticulum (ER) and which has a distinct role in virus assembly. The functioning of this protein was examined because rotaviruses undergo a unique morphogenesis in which newly made subviral particles bud into the ER, and during this process they obtain a transient membrane envelope that is subsequently lost within the lumen of the ER as particles mature. This process also involves Ca2+ to maintain the integrity and specific conformation of the two outer capsid proteins necessary for the correct association of proteins as the virus matures in the ER. NSP4 has multiple domains. NSP4 is a 20 kDa primary translation product, is co-translationally glycosylated to 29 kDa, and oligosaccharide processing yields the mature 28 kDa protein (Ericson et al 1983). The 175 amino acid backbone of NSP4 consists of an uncleaved signal sequence, three N-terminal hydrophobic domains, and a predicted amphipathic a-helix that overlaps a folded coiled-coil region (Chan et al 1988, Taylor et al 1996). The H2 transmembrane domain traverses the ER bilayer and the cytoplasmic C terminus of NSP4 functions as an intracellular receptor in viral morphogenesis (Au et al 1989, Meyer et al 1989, Taylor et al 1993, Bergmann et al 1989). NSP4 mobilizes intracellular Ca2+ ([Ca2+];) release from the ER and this may also affect viral morphogenesis (Tian et al 1994, 1995). NSP4 also possesses a membrane destabilization activity when incubated with liposomes that simulate ER membranes, and it has been hypothesized that this activity may play a role in the removal of the transient envelope from budding particles (Tian et al 1996). Finally, NSP4 may facilitate cell death and virus release from cells (Newton et al 1997), and polarized cells treated with NSP4 undergo reorganization of filamentous actin and lose transepithelial cell resistance due to alterations in tight junctions (Tafazoli et al 2001).

The discovery that NSP4 is an enterotoxin was made serendipitously during studies aimed at dissecting the molecular mechanisms by which NSP4 functions in morphogenesis. A synthetic peptide of NSP4 that contains amino acid residues 114—135 was conjugated to a carrier and injected into mice to produce a new antiserum. Although the mice received peptide (and no virus), they got diarrhoea. Pursuit of this observation showed that both the full-length protein and the 114—135 peptide share several properties that are consistent with NSP4 being an enterotoxin, and recent studies have found novel pleiotropic properties of NSP4 important for pathogenesis.

The protein induces diarrhoea in pups with a DD50 of 0.56 nmols while the DD50 of the 114—135 peptide is more than 10-fold higher, indicating that this peptide contains only a part of the active domain of the protein. The exact size of the active toxin domain remains unclear. Properties that define bacterial enterotoxins include their ability to stimulate net secretion in intestinal segments in the absence of inducing histological changes. NSP4 possesses these properties, confirming that it functions as an enterotoxin. The early studies and in vitro analyses of the response of human intestinal (HT29) cells to exogenously added NSP4 led to a working model for the mechanism of action of this enterotoxin (Ball et al 1996, Estes & Morris 1999). We hypothesized that an extracellular form of NSP4 functions as an enterotoxin after it is released from virus-infected cells. We proposed that extracellular NSP4 binds to an unidentified receptor on secretory intestinal cells, and activates a signalling pathway that mobilizes [Ca2+]; and leads to Cl_ secretion.

Our model predicted: (i) NSP4 is released from virus-infected cells either by cell lysis or secretion; (ii) avirulent rotaviruses may possess mutant NSP4 genes; (iii) intestinal cells possess a receptor for NSP4; (iv) age-dependent disease results from a change in the signalling pathway; and (v) antibody to NSP4 will reduce rotavirus-induced disease.

Further work has provided evidence for each of these predictions. A cleavage product of NSP4 can be detected in the medium of virus-infected cells and this protein retains enterotoxin activity (Zhang et al 2000). Sequence analyses of virulent/avirulent pairs of viruses detected sequence changes in the NSP4 gene in amino acid residues (aa) 136 and 138, and expressed NSP4 with these mutations shows reduced enterotoxin activity (Zhang et al 1998). These results also indicate that the enterotoxin domain extends beyond aa 114—135 and help explain why the initial NSP4 114—135 peptide was not as active as the full-length protein. A glutathione S-tranferase (GST) fusion protein containing aa 86—175 of the murine NSP4 also causes diarrhoea in mice (Horie et al 1999). Binding experiments of NSP4 to cells has shown that cells possess a receptor for NSP4, but the receptor has not yet been identified (C. Q.-Y. Zeng, M. Zhang & M. K. Estes, unpublished results 1999). Human intestinal cells exposed to exogenously added NSP4 initiate a signalling pathway that involves activation of phospholipase C (PLC), elevation in inositol 1,4,5-trisphosphate (InsP3) and mobilization of [Ca2+]i (Dong et al 1997). Crypt cells isolated from mice also respond in a similar manner and mobilization of [Ca2+]; occurs in both young and older mice (Morris et al 1999). Studies in mice lacking the cystic fibrosis transmembrane regulator (CFTR) channel provided evidence that age-dependent disease results from an event downstream of mobilization of [Ca2+];. Since the CFTR channel is a cAMP-regulated Cl_ channel , it was predicted that such mice should still get diarrhoea based on rotavirus and NSP4 stimulating a Ca2+-activated Cl_ channel. CFTR knockout mice get diarrhoea following rotavirus infection or NSP4 treatment, indicating that a different Cl_ channel than CFTR mediates this effect (Angel et al 1998, Morris etal 1999). CFTR knockout mice do not respond to other classical secretory agonists so the activity of NSP4 in these mice shows that NSP4 is a novel secretory agonist (Morris et al 1999). Cl_ secretion is age-dependent in CFTR knockout mice, indicating that age-dependent disease may result from an age-dependent induction, activation or regulation of this Cl_ channel (Morris et al 1999). This pathway would be one that is activated when NSP4 is released from virus-infected cells, possibly by cell lysis, and the affected cells are hypothesized to be the secretory crypt cells. Other cell types, including villus enterocytes, may also be affected in a similar manner by NSP4. Finally, protection studies using the neonatal mouse model have shown that antibody to NSP4 can induce protection against disease caused by infection with either simian or a highly virulent murine rotavirus (Zeng et al 2001). Protection consists of a reduction in the number of pups that develop diarrhoea, the severity of disease and the duration of the diarrhoea. Exogenous NSP4 has also been shown to alter polarized cell transepithelial cell resistance (Tafazoli et al 2001) and to inhibit the Na+-D-glucose symporter in brush border membranes (Halaihel et al 2000), two additional properties that may contribute to pathogenesis. An updated model of NSP4 action summarizes these new effects (Fig. 3).

Additional effects of NSP4 most likely occur by alternative pathway(s) that lead to changes in plasma membrane Cl_ permeability or other effects on epithelial cell function that are activated during viral infection when NSP4 is initially expressed as a transmembrane ER-specific glycoprotein within infected enterocytes. Endogenous expression of intracellular NSP4 in cells can also mobilize [Ca2+]; from internal stores but this is a separate process that is not affected by PLC inhibitors (Tian et al 1995). Changes in plasma membrane permeability are seen in virus-infected cells, but it is unclear whether these effects result from expression of NSP4 or another viral protein (Michelangeli et al 1991, 1995). Future studies of the effects of intracellularly expressed NSP4 may unravel other functions and properties of this multifaceted protein.

Can our new knowledge of the mechanisms ofpathogenesis be used to improve methods of treatment and prevention of rotavirus-induced diarrhoeal disease?

The discovery that rotaviruses have multiple effects on polarized epithelial cells and that these viruses produce an enterotoxin that mimics many of the same effects as the virus, raises the possibility that this protein could be useful in the development of new methods to prevent or treat rotavirus-induced disease. Other viral proteins may also be involved and if additional viral proteins play a role in the complex signalling cascades, they might also be useful targets to prevent disease. The discovery of the enterotoxin has stimulated new ideas and work on rotavirus pathogenesis but many questions remain to be answered in greater detail. It is clear that NSP4 is a novel multifunctional virulence factor and secretory agonist that can stimulate several cell signalling pathways, and that this protein has been used to identify novel calcium binding proteins in tumour cells (Xu et al 1999). Recently, involvement of the enteric nervous system in rotavirus diarrhoea has been reported (Lundgren et al 2000), although the mediators that trigger this process remain unknown. NSP4 is one possible factor. Further work is needed to dissect all the targets in the NSP4-induced signalling cascades and to understand the consequences and targets of the different signalling pathways induced by intracellular and extracellular NSP4. Such studies may lead to common mechanisms shared with bacterial pathogens that could lead to development of broadly active drugs to prevent diarrhoea caused by more than one pathogen. Understanding the mechanism of action of NSP4 may lead to new treatments for

Released NSP4 interacts with adjacent cell receptors

Absorptive Enterocyte (villus)

FIG. 3. Revised model of enterotoxin action. Rotavirus infected enterocytes produce virus-specific proteins including NSP4. NSP4 is released from cells by lysis or secretion and this extracellular NSP4 acts in a paracellular fashion by interacting with a receptor for NSP4 on adjacent cells. These secretory cells are likely to be crypt cells but might also be villus enterocytes. This triggers a rapid response involving a signal transduction pathway that mobilizes intracellular Ca"^ and results in increased plasma membrane Cl~ permeability. The plasma membrane permeability is age-dependent based on studies in CFTR knockout mice (Morris et al 1999). Extracellular NSP4 also causes a delayed disruption of the organization of filamentous actin and transepithelial cell resistance (Tafazoli et al 2001). Model modified from Estes & Morris (1999).

* Signal Transduction ci I Pathway a

Plasma membrane [CI ] permeability

Disruption of Microvillar Microfilament Network

Secretory Cell (crypt)

Na-K-2CI~ Cot ran s porter

Na-K-2Cr C otra ns portar

FIG. 3. Revised model of enterotoxin action. Rotavirus infected enterocytes produce virus-specific proteins including NSP4. NSP4 is released from cells by lysis or secretion and this extracellular NSP4 acts in a paracellular fashion by interacting with a receptor for NSP4 on adjacent cells. These secretory cells are likely to be crypt cells but might also be villus enterocytes. This triggers a rapid response involving a signal transduction pathway that mobilizes intracellular Ca"^ and results in increased plasma membrane Cl~ permeability. The plasma membrane permeability is age-dependent based on studies in CFTR knockout mice (Morris et al 1999). Extracellular NSP4 also causes a delayed disruption of the organization of filamentous actin and transepithelial cell resistance (Tafazoli et al 2001). Model modified from Estes & Morris (1999).

rotavirus-induced diarrhoea. For example, it is possible that antibody treatment or new drugs might be developed to treat children or animals with rotavirus infection and diarrhoea, specifically immunocompromised children with chronic rotavirus diarrhoea. Such potential new therapies may require a more detailed understanding of the NSP4 receptor on cells, the NSP4 structure, and whether there are distinct signalling pathways for NSP4 action when the protein is expressed endogenously versus when cells are exposed exogenously to NSP4.

A remaining question is whether this enterotoxin action is important in rotavirus pathogenesis in children. A direct answer is currently not available, but precedents with bacterial toxins indicate that results in animals generally are relevant for humans, so it seems likely that NSP4 plays a role in rotavirus pathogenesis in humans. Future studies can determine if antibody to NSP4 correlates with protective immunity and if vaccination strategies that induce immunity to NSP4 improve vaccine efficacy. Children do make antibody and cellular immune responses to NSP4 (Richardson et al 1993, Johansen et al 1999). Although there is sequence variability in NSP4, and four genetic groups are now known (Ciarlet et al 2000b), the broad immunity induced by NSP4 in mice protects against viruses in at least two distinct NSP4 genogroups (Zeng et al 2001). Whether immunity to NSP4 will provide the long sought after correlate of protection for rotavirus remains an important issue that needs to be investigated.

The discovery of the rotavirus enterotoxin raises interest in knowing whether other viruses code for enterotoxins. This is obviously important for understanding the mechanisms of pathogenesis for other gastroenteritis viruses such as astroviruses, caliciviruses, coronaviruses and enteric adenoviruses. This question is also relevant for viruses such as HIV that cause a devastating enteropathy. Finally, these results emphasize the common mechanisms of pathogenesis shared among microbial pathogens.

The work on NSP4 as an enterotoxin summarized in this article includes that of former students, postdoctoral fellows or current collaborators including Kit-Sing Au, Judy Ball, Wai-Kit Chan, Yanjie Dong, Kari Johansen, Andrew P. Morris, Linda J. Saif, J. Scott, L. Svensson, Peng Tian, andMingdong Zhang. The critical work, friendship and scientific enthusiasm of these colleagues is greatly appreciated.

+1 0

Post a comment