A new study led by researchers at Baylor College of Medicine and published in the Proceedings of the National Academy of Sciences reveals that human noroviruses, the leading viral cause of foodborne illness and acute diarrhea around the world, infect cells of the small intestine by piggybacking on a normal cellular process called endocytosis that cells use to acquire materials from their environment.
The study found that two compounds present in bile – bile acids and the fat ceramide – are necessary for successful viral infection of a laboratory model of the human small intestine.
In addition, the researchers report for the first time that bile acids also stimulate endocytosis in the small intestine.
The findings support further exploration of the possibility of reducing norovirus infection by modulating the levels of bile acids and/or ceramide.
“Human noroviruses invade cells of the small intestine where they replicate and cause gastrointestinal problems,” said co-first author Victoria R. Tenge, graduate student of molecular virology and microbiology in Dr. Mary Estes’s laboratory.
“Previous work from our lab showed that certain strains of norovirus required bile, a yellowish fluid produced by the liver that helps digest fats in the small intestine.
In the current study, we investigated which bile components were involved in promoting norovirus infection.”
The researchers worked with human enteroids, a laboratory model of human intestinal cells that retains properties of the small intestine and is physiologically active.
“Mini-guts, as we call them, closely represent actual small intestine tissue, and, importantly, they support norovirus growth, allowing researchers to study how this virus causes disease,” said co-first author Dr. Umesh Karandikar, a research scientist in the Estes lab.
Creating a stage that favors viral infection
The researchers discovered that bile acids and ceramide in bile were necessary for viral infection.
“Interestingly, we also discovered that bile acids stimulated the process of endocytosis in mini-guts. Our findings led us to propose that as bile acids activate endocytosis, they create a stage that norovirus takes advantage of by riding along with it to enter the cells and subsequently replicate, causing disease,” said corresponding author, Dr. Mary K. Estes, Cullen Foundation Endowed Professor Chair of Human and Molecular Virology at Baylor College of Medicine and emeritus founding director of the Texas Medical Center Digestive Diseases Center. “Bile acid-induced endocytosis in the small intestine was not previously appreciated.”
“This strategy works well for a food-borne virus,” said co-first author Dr. Kosuke Murakami, who was working in the Estes lab during most of this project. He is currently at the National Institute of Infectious Diseases in Tokyo.
“As people ingest food, the body’s normal response is to secrete bile into the small intestine. Noroviruses contaminating food piggyback on this natural bodily response to invade cells in the small intestine, replicate and cause disease.”
Working with mini-guts not only showed new insights into how norovirus causes disease, but also illuminated details about the basic biological process of endocytosis in the small intestine that had not been reported before.
“Our findings suggest the possibility that modulating the amount of bile acids and/or ceramide could help reduce norovirus infection,” Tenge said.
“This strategy might be particularly helpful to people who have norovirus infections for months, even years,” Karandikar said.
The induction of the host innate response plays an essential role in the suppression of pathogen infection. The synthesis of interferons (IFN) and the subsequent signalling cascades that leads to the induction of IFN-stimulated genes (ISGs), determine the outcome of viral infection (1, 2). An understanding of the mechanisms underlying the interplay between pathogens and innate immune responses is vital to understanding viral pathogenesis and can greatly aid the identification of potential therapeutic and/or preventive strategies.
Human noroviruses (HuNoV) are widely recognised as the leading cause of viral gastroenteritis worldwide (3). Noroviruses are classified into at least seven genogroups based on the sequence of the major capsid protein VP1 and regions within ORF1 (3–5).
HuNoVs belong to one of three norovirus genogroups (GI, GII, or GIV), which are further divided into >25 genetic clusters or genotypes (6–8). Epidemiological studies reveal that over 75% of confirmed human norovirus infections are associated with HuNoV GII (9, 10).
Whilst norovirus gastroenteritis typically results in an acute and self-limiting disease, the socioeconomic impact in both developed and developing countries is estimated to be more than $60.3 billion per annum (11).
HuNoV infection is particularly severe and prolonged in immunocompromised patients, including young children, elderly, or patients receiving treatment for cancer. In these cases infections can last from months to years (12, 13).
Our understanding of the molecular mechanisms that control HuNoV infection has been limited by the lack of robust culture systems that facilitate the detailed analysis of the viral life cycle.
As a result, murine norovirus (MNV) and other members of the Caliciviridae family of positive sense RNA viruses, such as feline calicivirus (FCV) and porcine sapovirus (PSaV), are often used as surrogate models (14–17). MNV, FCV and PSaV can all be efficiently cultured in immortalised cells and are amenable to reverse genetics (16–20). These model systems have been critical to understanding many aspects of the life cycle of members of the Caliciviridae (15).
Recent efforts have led to the establishment of two HuNoV culture systems based on immortalised B cells (21, 22) and intestinal epithelial cells (IECs) generated from biopsy-derived human intestinal epithelial organoids (IEOs) (23).
Whilst authentic replication of HuNoV can be observed in both the B-cell and IEC-based culture systems, repeated long-term passage of HuNoV and the generation of high titre viral stocks is not possible, suggesting that replication is restricted in some manner.
In the current study we sought to better understand the cellular response to HuNoV infection and to identify pathways that restrict HuNoV replication in organoid-derived IECs. Using RNA-Seq we observed that HuNoV infection of IECs results in an interferon-mediated antiviral transcriptional response. We show for the first time that HuNoV replication in IECs is sensitive to both Type I and III interferon and that HuNoV replication is restricted by virus-induced innate response.
Pharmacological inhibition of the interferon response or genetic modification of organoids to prevent the activation of the interferon response significantly improved HuNoV replication in IECs. Furthermore, we show that ongoing HuNoV replication is enhanced by the inhibition of RNA Pol II mediated transcription. Overall this work provides new insights into the cellular responses to HuNoV infection of the gut epithelium and identifies modifications to the HuNoV culture system that significantly enhances its utility.
Discussion
The efficient cultivation of HuNoV has remained a challenge since the initial identification of the prototype norovirus, Norwalk virus, in 1972 (57). Norovirus infection of the natural host species is very efficient, typically requiring <20 virus particles to produce a robust infection whereby >108 viral RNA copies are shed per gram of stool within 24 hours (58, 59). Even in heterologous hosts (e.g. pigs) the HuNoV infectious dose has been estimated to be ∼ 2X 103 viral RNA copies (60).
Despite this, and despite enormous efforts, the ability to culture HuNoV efficiently has been a significant bottleneck in the study of HuNoV biology (57). Therefore, the ability to culture HuNoV has the potential to transform our understanding of many aspects of the norovirus life cycle, greatly enhance the capacity to develop therapeutics and allows the characterization of authentic viral neutralization titres following vaccination, rather than the current surrogate gold standard (21, 23).
The net result of >40 years research has resulted in the establishment of two culture systems for HuNoV that use patient stool samples as the inoculum. The first such system relies on the replication of HuNoV within immortalized B-cells and requires the presence of enteric bacteria or soluble HGBA-like molecules from their surface (21, 22). Whilst, we have been able to reproduce the culture of HuNoV in immortalised and primary B-cells to varying degree of success (data not shown), we note that attempts by other labs have not universally been successful (21).
The recently developed HuNoV culture system IECs derived from intestinal organoids (23) while experimentally challenging, has been used in a number of subsequent studies to examine the impact of disinfectants (61) and the monoclonal antibodies (62, 63).
This study set out to use organoid-based system to assess the cellular pathways that restrict HuNoV replication and to further refine the experimental conditions that allow optimal growth of HuNoV in culture. We found that HuNoV infection induces a robust innate response in IECs, in contrast to previous studies using transfection of purified HuNoV viral RNA into immortalized cells which concluded that the interferon response is unlikely to play a role (37).
While the conclusions drawn in this previous study may be valid, it is likely that the inefficient replication seen using transfected RNA, where less than 0.1% of transfected cells contain active replicating viral RNA, reduce the sensitivity of the experimental system.
This may be further confounded by unknown mutations that affect the robustness of the sensing pathways within immortalized cells. These previous observations, also contrast with our own findings that suggest that the ability of cells to respond to exogenous interferon negatively impacts on HuNoV replication (64, 65).
This conclusion was based on the finding that the IFNλ receptor is epigenetically suppressed in an immortalized intestinal cell line which efficiently replicates a HuNoV GI replicon and that genetic ablation of IFNλ receptor expression enhances HuNoV replication in immortalized cells (65). We also recently described the generation of a robust culture system in zebrafish larvae, in which we also observed MX and RSAD2 (viperin) induction (39).
Furthermore, it is well established that the interferon response is key to the control of MNV infection as mice lacking a competent innate response often succumb to lethal systemic MNV infections (66–68), demonstrating that the innate response is key to the restriction of norovirus infection to intestinal tissues in the mouse model (69).
The development of the HuNoV organoid culture system provides the first opportunity to assess the impact of HuNoV infection on IECs, the first port of entry into the natural host.
Here we have seen that HuNoV infection of IECs induces an IFN-like transcriptional response by examining the replication of single HuNoV GII.4 isolate in IECs derived from two independent terminal ileum organoid lines from two different donors (Fig. 3).
We chose the terminal ileum-derived organoids as our source of IECs as our data to date would suggest that GII.4 HuNoV replicates more efficiently in IECs derived from this gut segment whereas the GII.3 isolate replicated more efficiently in duodenal lines (Fig. 1D). Whether this difference was organoid line or viral strain specific, or suggests differing tropism is unknown, however this observation was consistent across several different duodenal or ileal organoid lines (data not shown).
Under the conditions used in the current study, the overall number of genes altered more than 2-fold in response to infection was relatively modest, 70 and 162 for TI365 and TI1006 respectively. We found that the transcriptional response induced in each organoid line was highly comparable, with a substantial overlap in the induced genes (Fig. 4).
The use of UV inactivated inoculum allowed us to control for any non-specific effects of the other components of the filtered stool sample. Given the heterogeneity of any given stool sample, including this was essential to ensuring the observations were robust and represented alterations due to sensing of active viral replication intermediates.
The rather modest number of genes induced, likely reflects the heterogenous nature of the IEC cultures and that not all cells in any given monolayer are permissive to infection. We estimate that ∼30% of cells were infected under the conditions used for the gene expression analysis which is similar to previous reports (23).
The inclusion of Rux or TPL increased the overall number of infected cells in any given culture to ∼50% but even under the modified conditions, we have been unable to obtain higher levels of infection (data not shown). We hypothesize that obtaining higher levels of infection will likely require more uniform cultures, consisting primarily of enterocytes, the target cell for HuNoV (23).
The mechanism by which HuNoV is sensed by the infected cells is not currently known, however data from MNV suggests a clear role for Mda5-mediated sensing in the restriction of norovirus replication both in cell culture and in vivo (70). The sensing of MNV RNA occurs in a process that requires the HOIL1 component of the linear ubiquitin chain assembly complex (LUBAC) complex (71).
Other components of the RNA sensing pathways have been implicated in the innate response to MNV including MAVS, IRF3 and IRF7 (70, 71) but the role they play in sensing of HuNoV RNA is unknown. In addition to targeting STAT1 for degradation (72), the PIV5 V protein is known to also inhibit the activity of Mda5 (44). Whilst not directly assessed, it is therefore likely that the stimulation is of HuNoV replication in the presence of the PIV5 V protein is a combined result of both of these activities. Further studies using gene edited organoid lines will be required to better define the relative contribution of each component in the sensing of HuNoV.
The most highly induced gene in response to HuNoV infection in both organoid lines was IFI44L, a novel tumour suppressor (73) previously show to have modest antiviral activity against HCV (74) and RSV (75, 76). IFI44L was also potently upregulated in IECS infected with human rotavirus (HRV) (77).
Surprisingly, despite inducing a potent interferon response in IECs, HRV is not restricted by the endogenously produced IFN (77), an effect that has been hypothesized to be due to viral regulatory mechanisms that suppress the downstream activities of the induced genes. A number of the genes induced in response to HuNoV infection of IECs have previously been shown to have anti-viral activity against noroviruses.
GBP4 and GBP1 were both induced following GII.4 infection of both organoid lines (Fig 3). The GBPs are interferon induced guanylate-binding proteins that are targeted to membranes of vacuoles that contain intracellular fungi or bacterial pathogens (78, 79), where they frequently result in the disruption of the pathogen-containing vacuoles (79).
GBPs are targeted to the MNV replication complex in an interferon dependent manner that requires components of the autophagy pathway and exert their antiviral activity via an unknown mechanism (80).
GBP2 was also identified as a norovirus restriction factor in a CRISPR based activation screen where it was found to have potent antiviral activity against two strains of MNV (81). Further studies will be required to determine if GBPs have similar antiviral effects during HuNoV infection.
The IFIT proteins IFIT1-3 were also significantly induced in response to HuNoV Infection of IECs (Fig 3, Table S1). The IFITs are a family of interferon stimulated RNA binding proteins that, at least in humans, are thought to inhibit the translation of foreign RNAs by binding to 5’ termini and preventing translation initiation (82, 83).
In the context of norovirus infection, we have recently shown that the translation of norovirus VPg-linked RNA genome is not sensitive to IFIT1-mediated restriction (84), most likely due to the mechanism by the novel VPg-dependent manner with which norovirus RNA is translated (84). However, we did observe that IFIT1 in some way enhanced the IFN-mediated suppression of norovirus replication through an as yet undefined mechanism (84).
The development of the B-cell and organoid culture system have opened up the opportunity to dissect the molecular mechanisms of norovirus genome replication and to better understand host responses to infection. Others have observed that HuNoV replication in organoid derived IECs is highly variable (85) which agrees with our own experience during the course of the current study as we observed significant levels of week to week variation in infectious yield from the same organoid lines for any single strain of HuNoV (data not shown). We have also observed, as have others, that not all HuNoV strains appear to replicate efficiently in IECs derived from any single organoid line, which likely reflects the natural biology of HuNoV as individual susceptibility varies within any given population (85).
What factors contribute to the relative permissiveness of any given organoid line to an isolate of HuNoV remains to be determined, but it is clear that Fut2 function appears essential for most HuNoV isolates as FUT2 negative lines were not permissive to the strains of viruses tested here (85, 86), data not shown).
It is also possible that strains vary in the degree to which they induce and are sensitive to, the interferon response, as is common for other positive sense RNA viruses. Our data would suggest that irrespective of this, the replication of all isolates examined appear to be improved by treatment of cultures with Rux or TPL (Figs. 6 and 7; Figs. S2 and S3; and data not shown).
To our knowledge, our study represents the first demonstration that the genetic modification of human intestinal organoids can improve viral replication. The expression of BVDV NPro and PIV5 V proteins in cells has been widely used as a way to enhance virus replication in immortalised cells via the inactivation of aspects of the innate response (18, 87, 88).
While the genetically modified organoids enhanced HuNoV replication by up to 30-fold in comparison to unmodified organoids we found that this varied between organoid lines examined (not shown). Surprisingly, we found that the process of differentiation, resulted in a significant increase in the basal levels of a number of ISGs (data not shown).
Therefore the reason for variation in the enhancement is unknown but it may relate to the ability of any given organoid line to respond effectively and produce a rapid and effective innate response. The ability to readily generate gene edited human intestinal organoids while possible, is still very much in its infancy (89), therefore the ability to overexpress viral innate immune antagonists provides a more rapid way of generating intestinal organoids with specific defects in innate immune pathways.
However, the simple inclusion of TPL or Rux appears to phenocopy the effect of overexpression of either NPro or V protein and can be readily applied to any organoid line. This low cost modification to culture conditions enhances the utility of the experimental system by improving the robustness of the replication.
The use of pharmacological inhibitors for the stimulation of viral infection has been described in many instances in immortalised cell lines (49, 56, 90), and more recently for viral infection of intestinal organoids (91). The mechanism of action of Rux is well defined as it specifically targets the JAK kinases (46).
In contrast, the mechanism of action of TPL is less well defined but recent data suggests a direct mode of action on RNA polymerase II-mediated transcription (55). TPL has previously been shown to stimulate the replication of VSV by the inhibition of the interferon induced transcriptional responses (56). While TPL is not clinically used due to problems with water solubility, a water soluble pro-drug minnelide, has been trialled as an anti-cancer treatment for a number of cancers including pancreatic cancer (92).
Norovirus infection has now been widely accepted as a significant cause of morbidity and mortality in immunocompromised patients (13). In such cases, patients on immunosuppressive therapy following organ or stem cell replacement therapies, or those undergoing treatment for cancer, often suffer from infection lasting months to years (13, 93).
Such infections have significant impact on the overall health of the affected patient, resulting in significant weight loss and a requirement for enhance nutritional support (94). Ruxolitinib, under the trade name Jakavi, is approved for the treatment of a range of diseases including splenomegaly in patients with myelofibrosis and has been shown to be effective in the treatment of chronic or acute (48, 95).
Our data could suggest that the sustained administration of Rux in patients where chronic norovirus has been detected, may exacerbate the disease. We note however that during a study examining the effect of Rux on NK cell function in patients with STAT1 gain of function mutations, a single patient with chronic norovirus infection appeared to clear the infection following Rux treatment (96). The impact of Rux treatment on viral loads, and whether clearance was spontaneous, or due to improved NK cell function was not reported.
In summary, we have demonstrated HuNoV replication in IECs is restricted by the interferon response and that modulation of this response through either the genetic manipulation of intestinal organoids or the inclusion of pharmacological inhibitors, enhances HuNoV replication. Overall this work provides new insights into the cellular pathways and processes that control the replication of HuNoV, and provides improved conditions for the culture of HuNoV, enhancing the robustness of the HuNoV organoid culture system.
More information: Kosuke Murakami el al., “Bile acids and ceramide overcome the entry restriction for GII.3 human norovirus replication in human intestinal enteroids,” PNAS (2019). www.pnas.org/cgi/doi/10.1073/pnas.1910138117
