Astrovirus infection dysregulates goblet cell-associated antigen passages

Summary

Astroviruses are a major cause of pediatric diarrhea worldwide. Goblet cells are a primary target for astrovirus infection. In addition to secreting mucus, goblet cells facilitate oral tolerance through the formation of goblet cell-associated antigen passages (GAPs). Previous work showed that GAPs close in response to Salmonella infection to minimize its dissemination and prevent inappropriate inflammatory responses to dietary antigens. In contrast, we observed that GAPs remain open following astrovirus infection, likely due to the downregulation of epidermal growth factor receptor that mediates GAP closure. Further, Salmonella-induced GAP closure was reduced by astrovirus co-infection, and we observed higher bacterial dissemination in co-infected animals. Together, our study indicates that astrovirus-infected goblet cells are refractory to GAP inhibition, which impairs the protective host responses that combat a co-infecting bacterial pathogen.

Graphical abstract

Highlights

• •Goblet cell-associated antigen passages remain open after astrovirus infection
• •Astrovirus downregulates a key signaling pathway that regulates antigen passages
• •Salmonella-induced closure of antigen passages is reduced by astrovirus co-infection
• •Co-infection results in higher Salmonella dissemination

Microbiology; Cell biology

Introduction

Human astroviruses are a major cause of pediatric diarrhea worldwide.^1,2 While vastly underdiagnosed,³ evidence of sustained community-level transmission of astrovirus is supported by recent wastewater surveillance showing that astroviruses are the most highly abundant enteric virus detected.⁴ Further, seroprevalence studies have shown that all children experience at least one astrovirus infection by age 9.^5,6 Despite being so widespread, astroviruses are one of the least characterized enteric RNA viruses, and we lack a clear understanding of how these early childhood infections affect gut immunity.

Using murine^7,8 and enteroid^9,10 models for astrovirus, we and others have shown that the virus infects small intestinal goblet cells, specialized epithelial cells that secrete mucus. In addition to making mucus, goblet cells facilitate oral tolerance, the process by which the mucosal immune system tolerizes to dietary antigens.^11,12,13 Previous studies have shown that goblet cell-associated antigen passages (GAPs) close in response to Salmonella infection to prevent inappropriate inflammatory responses to innocuous dietary antigens and translocation across the gut epithelium.¹⁴ However, it is unknown whether this protective host response is triggered during an enteric virus infection. This is an important open question given that young children establish oral tolerance in the peri-weaning phase, a time during which they experience 2 to 5 diarrheal episodes each year.¹⁵

Prior work has shown that GAPs are developmentally regulated by microbial sensing via the signaling adaptor MyD88 and activation of the epidermal growth factor receptor (EGFR).^16,17,18 GAP formation is induced by acetylcholine acting on the M₄ muscarinic acetylcholine receptor (mAChR4) on goblet cells. Both MyD88 signaling and activation of EGFR lead to the inhibition of mAChR4 via p42/p44 mitogen-activated protein kinase (MAPK) activation.¹⁶ In the dynamic gut environment, there are likely many EGFR ligands and sources of cholinergic stimuli from neighboring cells and microbes that act to regulate the frequency of GAP induction.^19,20 A previous study demonstrated that IL-1β induced following Salmonella infection leads to MyD88-dependent transactivation of EGFR, resulting in GAP closure.¹⁴ Using the mouse model for astrovirus, we previously showed that IL-1β along with other pro-inflammatory cytokines are induced following infection.²¹ In rare instances, disseminated infection can occur after astrovirus infection, resulting in systemic, central nervous system disease that is often fatal.^22,23 Thus, we reasoned that GAPs would close following astrovirus infection.

In the following study, we tested whether GAPs close in response to astrovirus infection and assessed whether key pathways that regulate GAPs remain functional during infection. Surprisingly, we found that GAPs remain open during astrovirus infection and demonstrated that infected goblet cells are refractory to GAP inhibition. Further, we observed that astrovirus co-infection with Salmonella reduced GAP closure and was associated with increased bacterial translocation. Overall, our findings suggest that astrovirus infection can have direct consequences on a critical pathway for inducing tolerance and restricting co-infecting bacterial pathogens.

Results

Murine astrovirus does not trigger GAP closure

To evaluate if GAP frequency changes during enteric viral infection, groups of 3-week-old weanling C57BL/6J mice were orally inoculated with murine astrovirus (MuAstV) or PBS for mock infection. At 3, 6, 12, and 20 days post-infection (dpi), 10 kDa rhodamine-conjugated dextran was intraluminally injected into the duodenum, the main site of MuAstV replication,²⁴ of anesthetized animals to visualize goblet cells forming GAPs. Duodenal tissues were harvested, and histological analysis from tiled images was collected to enumerate GAPs (Figure 1A). At steady-state, 52% of goblet cells were forming GAPs in uninfected animals, comparable to the 51–58% observed in infected animals at 6, 12, and 20 dpi (Figure 1B), which respectively correspond to peak infection, post-peak infection, and virus clearance.²⁴ Interestingly, at the pre-peak infection timepoint (3 dpi), there was a significant increase in the percentage of goblet cells forming GAPs (67%) as compared to the uninfected animals (Figure 1B). We next examined whether GAP closure could be triggered upon initial inoculation. To test this, we performed intraluminal injections of MuAstV or PBS 1 h prior to injection with rhodamine-dextran, but we did not observe any change in GAP frequency (Figure 1C).

Figure 1.

MuAstV infection and direct intraluminal injection do not trigger GAP closure

(A) GAP formation identified by rhodamine-dextran and wheat germ agglutinin (WGA) co-staining within goblet cells in the duodenum of infected weanlings at 6 dpi (left) and 12 dpi (right), scale bars, 20 μm.

(B) GAP quantification per goblet cell during the infection in comparison to mock-infected animals (n = 6–14 animals/group).

(C) Intraluminal injection of MuAstV or PBS (mock) followed by GAP quantification 1 h later (left) and representative images (right), n = 3–5 animals/group, scale bars, 80 μm.

(D) GAP quantification after vehicle (PBS) and tropicamide (Trop) administration followed by GAP quantification 1 h later (left) and representative images (right), n = 4–5 animals/group, scale bars, 40 μm. Data are summarized as mean ± SEM and represent at least 2 independent experiments except for C and D, which represent single experiments. Statistical testing for (B) Dunnett’s 1-way ANOVA, (C) Welch’s t test, and (D) Mann-Whitney U; ∗p < 0.05 and ∗∗p < 0.005; ns, non-significant.

Because prior studies of GAP regulation primarily focused on adult mice (7–12-weeks-old²⁵), we next wanted to confirm that we were able to observe a reduction in GAP frequency in the duodenums of 3-week-old weanlings used in our model. To this end, we performed intraperitoneal injections of tropicamide, a muscarinic receptor antagonist that inhibits GAP formation.¹⁶ Histological quantifications showed that compared to vehicle control (PBS), GAP frequency was significantly reduced following tropicamide treatment (Figure 1D). Together, these data indicate that while we can visualize GAP closure with a known GAP inhibitor, astrovirus infection does not trigger GAPs to close.

GAPs remain open during neonatal murine astrovirus infection

Because previous studies have shown the GAPs are regulated during early life through a combination of maternal and microbiota influences,¹⁷ we next sought to examine whether MuAstV infection triggers GAP closure during this earlier developmental window. In the small intestine, GAPs were previously shown to be closed until postnatal day (PND) 12. We initially examined uninfected neonatal C57BL/6J mice at PND 8 and identified many GAPs in the duodenum but much fewer in the jejunal and ileal regions of the small intestine (Figure 2A). To gain finer resolution, we further divided each small intestinal region into 3 subsections and quantified GAPs in each (Figure 2B). Whereas 55–60% of goblet cells form GAPs in the duodenal sections, this percentage was reduced in the jejunal (36–40%) and ileal sections (7–22%), indicating a region-specific trend in GAP formation. Next, to examine whether GAP frequency changes during infection in the duodenum, PND 8 mice were infected with astrovirus or mock-infected with PBS by oral gavage, and rhodamine-dextran was injected into the duodenums of anesthetized animals to visualize goblet cells forming GAPs at 4 dpi, corresponding to PND 12. There was no significant difference in GAP frequency between infected mice and uninfected mice (Figure 2C), which was consistent with what we observed in weanlings (Figure 1B).

Figure 2.

GAPs decrease in frequency from the duodenum to ileum in neonatal mice and remain open during MuAstV infection

(A) GAP formation identified by rhodamine-dextran and wheat germ agglutinin (WGA) co-staining within goblet cells in the duodenum (scale bars, 100 μm), jejunum, and ileum (scale bars, 200 μm) collected from neonates at PND 8.

(B) GAP quantification based on intestinal location within neonates at PND 8 (n = 6–7 animals/region).

(C) GAPs per goblet cell in mock- and MuAstV-infected mice collected at 4 dpi, corresponding to PND 12 (n = 7 animals/group). Data are summarized as mean ± SEM and represent at least 2 independent replicate experiments. Statistical testing used Welch’s t test; ns, non-significant.

Productively infected goblet cells form GAPs

Given that both direct infection and the infection environment (e.g., innate sensing, cytokine signaling) could lead to GAP closure, we initially enumerated GAP frequency globally, irrespective of whether the goblet cell was infected (Figure 1B). We next wanted to determine whether (1) productively infected goblet cells still formed GAPs, and (2) if the frequency of infected cells forming GAPs was lower than steady-state levels in uninfected goblet cells. Because visualization of the virus via in situ hybridization (ISH) requires processing steps that would reduce the rhodamine-dextran signal that is needed to enumerate GAPs, we utilized a serial imaging strategy. Using tissues from infected animals that had been injected with rhodamine-dextran to label goblet cells forming GAPs, we first generated tiled fluorescent images to visualize GAPs and then performed ISH using a MuAstV-specific probe, before re-imaging the tissue. Analyzing the tiled images side-by-side, we identified infected cells forming GAPs (Figure 3A). As an additional approach, we used an anti-dsRNA antibody to indicate cells forming virus replication complexes, confirming that infected cells were capable of forming GAPs (Figure 3B). We next performed serial imaging analysis on acutely infected weanlings (6 dpi) and neonates (4 dpi), as well as persistently infected athymic CD1 Nude and muMT mice that lack mature T and B cells, respectively. Overall, we found that an average of 55% of infected goblet cells formed GAPs in C57BL/6J mice, in comparison to over 80% in the two immunocompromised mouse strains (Figure 3C). We additionally quantified GAP frequencies in uninfected goblet cells within these same animals and found no significant differences between infected and uninfected goblet cells in C57BL/6J or CD1 Nude animals (Figure 3C). However, we did observe a significantly higher frequency of GAP formation in the infected goblet cells of muMT animals (Figure 3C). Together, these data indicate that GAPs within infected goblet cells continue to form at steady-state levels or higher. We also examined whether the stage of infection within the cell was related to GAP formation, but there appeared to be no significant differences when we classified the level of infection as low, medium, or high, and GAP formation (Figures 3D and 3E). Overall, these data indicate that regardless of age, host immune status, or stage of virus infection, GAPs continue to form in astrovirus-infected cells.

Figure 3.

Murine astrovirus-infected goblet cells form GAPs

(A) Representative images of serial imaging analysis with rhodamine-dextran (left, white arrows) denoting GAPs in infected cells as indicated by ISH signal for the MuAstV genome (right, black arrows), scale bars, 40 μm.

(B) Representative image of an infected goblet cell forming a GAP as indicated by co-staining for rhodamine-dextran and anti-dsRNA, scale bars, 5 μm.

(C) GAP quantification from infected and uninfected goblet cells from wildtype C57BL/6J (triangle = neonate, circle = weanling) and immunocompromised mice (n = 3–5 animals/group).

(D) Representative images of goblet cells at progressive stages of infection and GAP formation based on serial imaging analysis. Bottom images contain villi with goblet cells forming (left, white arrowhead) and not forming (left, white arrow), although both cells are robustly infected (right, black arrowhead and arrow), scale bars, 20 μm.

(E) GAP quantifications according to level of infection (low, medium, high) as depicted in D (n = 5 animals). Data are summarized as mean ± SEM. Statistical testing used in (C) paired t test and (E) Tukey’s 1-way ANOVA; ∗p < 0.01; ns, non-significant.

Astrovirus infection overrides GAP inhibition

Given that GAPs remain open in goblet cells infected with MuAstV (Figure 3), we next investigated whether the mechanisms that regulate GAP closure are disrupted following infection. Previous studies showed that GAP closure occurs via IL-1β and EGFR signaling cascades that inhibit mAChR4 responsiveness to acetylcholine in goblet cells.^16,17,18 At 6 dpi, we treated groups of infected and uninfected mice with tropicamide, a mAChR4 antagonist, as well as IL-1β, prior to enumerating GAPs. In comparison to uninfected mice given PBS, we noted that infected and uninfected mice treated with tropicamide and IL-1β had significantly fewer GAPs (Figures 4A–4C). However, whereas uninfected mice treated with EGF exhibited a significant decrease in GAP frequency as anticipated, infected mice treated with EGF did not show a significant reduction in GAPs (Figure 4D). These data indicate that while direct mAChR4 inhibition and IL-1β signaling still function to close GAPs during astrovirus infection, EGF-mediated closure is disrupted.

Figure 4.

Astrovirus infection overrides mechanisms of GAP closure

(A) Representative images show GAPs forming in uninfected PBS-treated weanlings, identified by rhodamine-dextran and wheat germ agglutinin (WGA) co-staining, which are less frequently identified within the duodenums of infected (7 dpi) tropicamide (Trop) and IL-1β-treated animals, but not EGF-treated animals, scale bars, 80 μm. GAP quantification from uninfected PBS-treated animals relative to (B) tropicamide-treated uninfected and infected animals (n = 6 animals/group), (C) IL-1β-treated uninfected and infected animals (n = 7 animals/group), and (D) EGF-treated uninfected and infected animals (n = 5–7 animals/group).

(E) Egfr expression in goblet cell subclusters from infected (I) and uninfected (U) mice based on single-cell transcriptomics collected at 6 dpi. Previously published data,⁷ with susceptible goblet cells (clusters 1–3) and non-susceptible goblet cells (cluster 4) indicated by MuAstV expression. Average expression and percent of cell population are indicated by color gradient and size of dot, respectively.

(F) Representative images with inset of ISH using an Egfr-specific probe in goblet cells identified in duodenal tissues collected from mock and MuAstV-infected animals at 6 dpi (n = 3–4 animals/group), scale bars, 170 μm.

(G) Egfr expression quantified by ISH puncta counted from goblet cells identified in duodenal tissues in F.

(H) EGFR levels quantified by ELISA from intestinal epithelial cell lysates collected from mock- and MuAstV-infected animals (n = 5–6 animals/group).

(I) Representative images of a GAP forming in an infected goblet cell as identified by co-staining of rhodamine-dextran and WGA (left, white arrow) followed by serial imaging of red ISH signal for the MuAstV genome (right, black arrow), scale bars, 15 μm.

(J) Using images gathered from serial imaging analysis as shown in I, GAP frequencies were quantified within infected and uninfected goblet cells from the same infected animals treated with GAP inhibitors (tropicamide, IL-1β, EGF) (n = 3–6 animals/group). Data are summarized as mean ± SEM (B–D) or median with interquartile range (dashed lines, G) and represent at least 2 independent experiments except for E–H, which represent single experiments. Statistical testing for (B) Kruskal-Wallis test with Dunn’s multiple comparisons, (C and D) Dunnett’s 1-way ANOVA, (G) Mann-Whitney U test, (H) Welch’s t test, and (J) paired t test, ∗∗∗∗p < 0.0001, ∗∗p < 0.01, and ∗p < 0.05; ns, non-significant.

We next questioned whether the loss of EGF responsiveness could be due to downregulation of EGFR following MuAstV infection. We previously performed single-cell transcriptomics on duodenal epithelial cells collected from uninfected and infected animals at 6 dpi and computationally defined four subclusters of goblet cells, three of which were susceptible to MuAstV infection.⁷ We queried this dataset to examine Egfr expression in these goblet cell subpopulations and observed that there was lower expression in all susceptible goblet cell (clusters 1–3), whereas there was no change in expression in the non-susceptible goblet cells (cluster 4) (Figure 4E). As a complementary approach, we additionally quantified Egfr expression by ISH and found significantly fewer puncta inside of goblet cells from infected animals as compared to mock-infected animals at 6 dpi (Figures 4F and 4G). Likewise, we observed significantly lower amounts of EGFR protein levels in intestinal epithelial cell lysates collected from infected animals as compared to mock-infected animals at 7 dpi (Figure 4H). Because we also noted fewer Egfr puncta in neighboring non-goblet cells within infected animals, this could support a more global response to infection that contributed to the lower EGFR protein observed in intestinal epithelial cell lysates. Overall, these data indicate that EGF-mediated GAP closure is disrupted during astrovirus infection, likely due to a global downregulation of EGFR.

Next, to examine whether GAP regulation was specifically disrupted within MuAstV-infected goblet cells, we quantified GAPs from a subset of the infected animals that were given the different GAP inhibitors. Using serial image analysis, we observed that in all treatment conditions, infected goblet cells still formed GAPs (Figure 4I). We quantified GAP frequencies in the infected and uninfected goblet cell populations within the same infected animals and found no significant differences in GAP frequencies among tropicamide- or EGF-treated animals (Figure 4J). However, we did observe a significantly higher frequency of GAP formation in the infected goblet cells as compared to uninfected goblet cells from IL-1β-treated animals (Figure 4J). Together, these data indicate that (1) EGF regulation of GAPs is disrupted in the duodenum during MuAstV infection, likely due to the downregulation of EGFR, and (2) infected goblet cells are refractory to all forms of GAP inhibition.

Astrovirus co-infection with Salmonella Typhimurium keeps GAPs open and facilitates dissemination

Given our observation that MuAstV infection disrupts GAP regulation, we questioned whether this could override the protective response to close GAPs following Salmonella infection. To address this, we enumerated GAPs following intraluminal injection with Salmonella enterica serovar Typhimurium or PBS (Figure 5A) in mock and MuAstV-infected mice at 7 dpi (Figure 5B). In comparison to uninfected mice injected with PBS, which had an average of 66% of goblet cells forming GAPs, we observed that GAP frequency significantly dropped to 33% in mice infected with Salmonella (Figure 5C), consistent with prior data.¹⁴ In contrast, we observed that MuAstV-infected animals inoculated with Salmonella still had 51% of goblet cells forming GAPs, which was significantly higher than uninfected animals inoculated with Salmonella and comparable to the 58% observed in MuAstV-infected mice injected with PBS (Figure 5C). These data suggest that co-infection with MuAstV can block a significant portion of Salmonella-induced GAP closure.

Figure 5.

Astrovirus co-infection with Salmonella Typhimurium blocks GAP closure and facilitates dissemination

(A) Representative images of weanling mice given an intraluminal injection with Salmonella (Sal) expressing GFP or PBS control in MuAstV or PBS mock-infected mice, scale bars, 80 μm.

(B) Representative images of GAPs forming in MuAstV or PBS mock-infected mice given Salmonella or PBS by intraluminal injection, followed by GAP quantification 1 h later. GAPs identified by rhodamine-dextran and wheat germ agglutinin (WGA) co-staining, scale bars, 80 μm.

(C) GAP quantification from MuAstV or PBS mock-infected mice given Salmonella or PBS by intraluminal injection (n = 3–7 animals/group).

(D) Weight changes at 4 dpi in animals orally gavaged with PBS, Salmonella alone, or co-infection with MuAstV and Salmonella (n = 4–8/group). Colony-forming units (CFU) quantified from (E) spleens and (F) small intestines harvested from these groups of mice at 4 dpi (n = 4–17/group). GAP and weight data are summarized as mean ± SEM, whereas CFU data are summarized as geometric mean with 95% confidence intervals. Data in all panels represent at least 2 independent experiments. Replicate experiments are denoted by different shapes in D–F. Statistical testing for (C) Dunnett’s 1-way ANOVA, (D) Welch’s t test, and (E and F) Mann-Whitney U test; ∗∗∗p < 0.0005, ∗∗p < 0.005, and ∗p < 0.05; ns, non-significant.

Previous data indicate that Salmonella predominantly translocates across the small intestinal epithelium and can use GAPs as a portal for entry.¹⁴ Based on this information, we hypothesized that MuAstV followed by Salmonella co-infection would result in higher extra-gastrointestinal dissemination due to GAP dysregulation. To test this hypothesis, we infected mice with MuAstV or PBS for mock infection, and at 6 dpi, both groups of mice were orally inoculated with Salmonella. After 4 days, both the co-infected animals and animals given Salmonella alone lost weight, and this was not statistically different (Figure 5D). Small intestines, small intestinal contents, cecum, and spleens were harvested to determine whether co-infection led to higher colonization and dissemination, respectively. We observed that co-infected animals harbored significantly higher colony-forming units (CFUs) in the spleen in comparison to animals infected with Salmonella alone (Figure 5E), whereas there was no significant difference in CFUs in the small intestines, small intestinal contents, or cecums (Figures 5F and S1). Overall, these data indicate that astrovirus infection can disrupt the ability of the host to limit extra-intestinal dissemination of pathogenic bacteria.

Discussion

Despite being one of the most common causes of pediatric diarrhea,^1,2 we understand little about how astroviruses affect gut homeostasis and immunity. Because astroviruses replicate in goblet cells,^7,8,9,10 we questioned whether infection could alter the formation of GAPs. GAPs are critical for establishing oral tolerance during early life,^13,17 which coincides with the developmental age that has the highest astrovirus disease burden.²⁶ Our study demonstrates that, unlike Salmonella infection, GAPs remain open throughout astrovirus infection. We further determined the key regulatory pathways disrupted by astrovirus and the consequences this has on impairing host responses against a bacterial co-infection. Overall, these results reveal a previously unappreciated facet of astrovirus pathogenesis and open up new questions surrounding how they disrupt gut homeostasis during early-life development.

In this study, we examined four time points during astrovirus infection, enabling us to examine how the different phases of infection could impact GAP formation. We found that GAPs remain open throughout acute MuAstV infection and even increased in frequency at the beginning stages of the infection (3 dpi) compared with uninfected animals. We also observed that GAPs remain open during persistent infection in immunocompromised mice. In addition, direct administration of the virus to the gut lumen did not induce GAP closure. So why do GAPs not close? To address this, we used three forms of pharmacologic inhibition, EGF, IL-1β, and tropicamide, to understand how astrovirus infection may alter GAP regulation. We found that EGF could not induce GAP closure, likely because EGFR expression is downregulated following infection. EGFR plays a pivotal role in GAP regulation, as previous studies showed that goblet cell-specific EGFR knockout animals fail to close GAPs in response to microbial stimuli and IL-1β.¹⁴ Further, administration of the EGFR inhibitor, tyrphostin AG1478, blocks EGFR activation, which in turn prevents GAP closure.¹⁶ It is notable that while IL-1β closed GAPs during MuAstV infection, we did not observe a reduced frequency of GAP formation in the infected goblet cell population. These data could suggest that IL-1β closes GAPs through an alternative pathway in the cell that is not EGFR-dependent. Alternatively, the response to IL-1β could be dose-dependent, and higher doses administered exogenously could still close GAPs globally after the downregulation of EGFR following MuAstV infection. However, given that we further observed that infected goblet cells were also refractory to tropicamide inhibition of mAChR4, this could suggest that there is a cell-intrinsic disruption at a point where these GAP regulatory pathways converge. Alternatively, or in addition, there may be differences in how viruses are sensed within the goblet cell as compared to bacteria. It is notable that toll-like receptor (TLR) expression is lower in small intestinal goblet cells in comparison to colonic goblet cells, with the dsRNA-sensing TLR3 being the only exception.¹⁶ However, TLR3 does not signal through MyD88, and therefore may not lead to EGFR trans-activation and GAP closure.¹⁶ Notably, both TLR7 and 8, which sense ssRNA and signal through MyD88, are not expressed in small intestinal goblet cells,¹⁶ which could contribute to the lack of GAP closure after infection.

Using a neonatal model of infection, our studies additionally demonstrated that GAPs continue to form in infected goblet cells regardless of age. Previous studies showed that whereas small intestinal goblet cells in adults remain open during steady-state conditions, GAPs in neonatal animals prior to PND 10 are closed.¹⁷ Consistent with these data, at PND 8, we observed that GAPs remained open in the duodenum but were mostly closed at the distal end of the ileum. This region-specific difference in GAP formation has not been previously observed, but could be explained by differential expression of EGFR, the bioavailability of EGF in milk,¹⁸ or stimuli that signal through EGFR that are region-specific in the small intestine. In addition, the change in GAP frequency from the duodenum to ileum in neonates correlates with bacterial abundance^27,28 and would also suggest that the majority of early-life antigen sampling occurs in the duodenum, in addition to the colon, as shown previously.¹⁷ Because astrovirus infection changes EGFR expression, this could significantly alter the immune programming during early life that promotes the development of oral tolerance. In addition, the GAP dysregulation caused by astrovirus infection could result in more early-life opportunities that facilitate bacterial translocation, which was recently shown to be an important disease mechanism in a model of late-onset sepsis.¹⁸

Is there a benefit for the virus to keep GAPs open? It is possible that, similar to Salmonella, GAPs could serve as a portal for virus dissemination. There have been 15 human cases of astrovirus neurological disease²³ and 11 cases of viremia identified to date,^{29,30,31,32,33,34,35} but the route of this disseminated disease remains unclear. Alternatively, GAPs could aid in establishing and promoting virus entry and it is intriguing to consider whether goblet cells forming GAPs are more likely to become infected. It is notable that previous studies have shown that villus goblet cells are more likely to form GAPs in the small intestine, which is the same population that MuAstV targets for infection.^21,25 It is also possible that the early stages of the infection could benefit from co-opting the tolerogenic process mediated by GAPs to initially evade host immunity, which would be consistent with their ability to establish prolonged infections.²⁴ Regardless of the potential benefit for viral infection, our co-infection experiments showed that astrovirus could block GAP closure following Salmonella infection and promote dissemination that was independent of colonization in the intestine, highlighting an important consequence of GAP dysregulation caused by astrovirus. While it is possible that other unmeasured facets of astrovirus infection beyond GAP dysregulation could explain the enhanced dissemination by Salmonella, it was previously shown that loss of goblet cells, and by association, GAPs, results in significantly reduced bacterial dissemination.¹⁴ These data should be placed in the context of recent data from a multi-site birth cohort that showed 77% of astrovirus-positive diarrheal samples were co-detected with at least one other enteropathogen.²⁶ Additional work is needed to understand the complexity of these co-infections and whether these interactions can improve our understanding of diarrheal disease.

Finally, our data indicate that there could be continued antigen sampling in the context of an inflammatory environment following astrovirus infection, particularly during critical developmental phases. Oral tolerance is established early in life, during which antigen sampling of new dietary antigens via GAPs is essential, as it has previously been shown that forcing GAPs open during Salmonella infection leads to the induction of inflammatory, dietary antigen-specific T cells.^13,14 Given that the first exposure to astrovirus during childhood coincides with the etiological time window for the onset of digestive diseases, future studies are needed to determine whether astroviruses could trigger a break in tolerance, as has been shown for other enteric viruses via bystander activation.^36,37 Overall, the continued GAP formation in the context of enteric virus infection highlights the need for additional study to understand the consequences on the development of oral tolerance.

Limitations of the study

The findings of this study should be framed within its limitations. First, there is currently only one method to measure GAPs, and that is by fluorescence microscopy. The development of orthogonal approaches would be highly beneficial to complement this single method. Second, although our previous studies showed no sex-based differences in MuAstV infection,²⁴ we used female-only muMT and male-only groups of CD1 nude mice for a subset of experiments and it is possible that there are sex-based differences to be discovered. Second, while our study was designed to test whether GAPs close after astrovirus infection, we did not specifically test whether infected goblet cells form GAPs at higher frequencies than uninfected goblet cells, although we observed evidence of this in a subset of experiments. Additional studies are needed to determine whether this is a consistent finding and to define the mechanism by which this could occur. Third, although our data support a model by which GAPs remain open following astrovirus infection due to reduced EGFR expression, additional studies are needed to define what aspect of the infection drives the downregulation of EGFR. Finally, it is possible that other unmeasured facets of astrovirus infection beyond GAP dysregulation could also contribute to the enhanced Salmonella dissemination we observed. Future studies aimed at visualizing Salmonella translocating across astrovirus-infected goblet cells as well as testing whether closing GAPs with an inducible mAChR4 conditional knockout can reduce Salmonella dissemination after co-infection would shed light on this open question.

Resource availability

Lead contact

Further information and requests for resources should be directed to and will be fulfilled by the lead contact, Valerie Cortez (vccortez@ucsc.edu).

Materials availability

All unique reagents generated in this study are available on request with a completed materials transfer agreement.

Data and code availability

All data reported in this paper will be shared by the lead contact upon request. This paper does not report original code. Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Acknowledgments

This research was supported by National Institutes of Health grant nos. K22 AI156116 and R21 DK139534, a Hellman Fellowship, and a Society for Mucosal Immunology Technique-Sharing grant to V.C. The authors thank Dr. Mary O’Riordan for sharing the Salmonella enterica serovar Typhimurium (SL1344/pFPV25.1), Dr. Manuela Raffatellu for advice on the Salmonella inoculations, and Dr. David Boyd for helpful comments on the manuscript. We additionally thank Dr. Ben Abrams of the UCSC Life Sciences Microscopy Center (RRID: SCR_021135) and the UCSC Vivarium Staff.

Author contributions

J.G., N.H.-O., M.O., A.P., C.V., and V.C. performed the experiments and analyzed the data. V.A., K.G.M., R.D.N., and V.C. designed and supervised this study. J.G. and V.C. wrote the paper with input from all authors.

Declaration of interests

R.D.N. and K.G.M. are inventors on patent US11,241,480 Methods for Modulation of Dietary and Microbial Exposure With Compounds Comprising an EGFR Ligand.

Declaration of generative AI and AI-assisted technologies in the writing process

No generative AI or AI-assisted technologies were used in the writing process.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
anti-dsRNA/ssRNA [P6] Rabbit mAb	Absolute Antibody	Ab03115-23.0
Wheat germ agglutinin-fluorescien (FITC)	Vector Laboratories	FL-1021-5
Bacterial and virus strains
Murine astrovirus (MuAstV)	Cortez Lab, UC Santa Cruz	SJ001
Salmonella enterica expressing eGFP	Auerbuch-Stone Lab, UC Santa Cruz	SL1344 pFPV25.1
Chemicals, peptides, and recombinant proteins
Tropicamide	Thermo Scientific	CAT#AAJ6113203
Carbamylcholine chloride	Thermo Scientific	CAT#AC108240050
Recombinant Mouse IL-1β	VWR	CAT#71005-114
Recombinant Mouse EGF	R&D	CAT#2028EG200
10 kDa Dextran-Tetramethylrhodamine, lysine fixable	Invitrogen	CAT#D1817
Formalin, Buffered, 10%	Fisher	CAT#SF1004
VECTASHIELD Vibrance Antifade Mounting Medium with DAPI 10 ml	Vector Laboratories	CAT#H-1200-10
DMSO	Sigma Aldrich	CAT#D2650
Ketamine Hydrochloride	Sigma	CAT#K4138-10ML
DAPI 1mg lyophylized	Cell Signaling	CAT#4083S
Critical commercial assays
QIAmp Viral RNA Mini Kit	Qiagen	CAT#52906
In-situ hybridization RNAscope 2.5 HD Assay - RED	Advanced Cell Diagnostics	CAT#322373
Mouse EGFR DuoSet ELISA kit	R&D	CAT#DY1280-05
Experimental models: Organisms/strains
Mouse muMT	The Jackson Laboratory	Strain#002288; RRID: IMSR_JAX:002288
Mouse CD1 Nude	Charles River Laboratories	Strain#086; RRID:IMSR_CRL:86
Mouse Wild-type C57BL/6J	The Jackson Laboratory	Strain#000664; RRID: IMSR_JAX:000664
Oligonucleotides
MuAstV qRT-PCR forward primer	Integrated DNA Technologies (IDT)	TACATCGAGCGGGTGGTCGC
MuAstV qRT-PCR reverse primer	Integrated DNA Technologies (IDT)	GTGTCACTAACGCGCACCTTTTCA
MuAstV qRT-PCR probe	Life Technologies	6-FAM-TTTGGCATGTGGGTTAA-(MBGNFQ)
Software and algorithms
Prism	GraphPad	N/A
ImageJ	Fiji	N/A

Experimental model and study participant details

Male and female wild-type C57BL/6J (#000664) breeder mice were originally purchased from The Jackson Laboratory. For all experimental MuAstV infections in C57BL/6J, mice were bred in-house and litters of male and female 3-4-week-old weanlings were split and used for experimental and control groups. While our previous studies have indicated that there is no sex-bias in MuAstV infection,²⁴ we still combined smaller litters were combined and spread across experimental and control groups to achieve sex parity. Female 4-week-old muMT (#002288) mice were purchased from The Jackson Laboratory, and male 3-week-old CD1 Nude mice (#087) were purchased from Charles River Laboratories and experiments described were performed within 3-4 weeks following their arrival. All animals were housed under specific pathogen free conditions (2-5 animals/cage) in individually ventilated Thoren microisolator cages. All caging with bedding (corn cob mixed with 10% Diamond-Dri absorbent squares) was autoclaved before irradiated food was added to cage tops and reverse osmosis water provided via lixit valves during breeding and maintenance or with individual Hydro-pac water pouches following infection. All animal experiments were approved by the University of California, Santa Cruz (UCSC) Institutional Animal Care and Use Committee (protocol #Cortv2406dn). UCSC is fully accredited by the Association for the Assessment and Accreditation of Laboratory Animal Care International (AAALAC-I) and has an approved Animal Welfare Assurance Statement on file with the Office of Laboratory Animal Welfare (D16-00493). These guidelines were established by the Institute of Laboratory Animal Resources and were approved by the Governing Board of the U.S. National Research Council.

Method details

MuAstV infections

For all experiments, mice were first confirmed to be negative for MuAstV (strain SJ001²⁴) by qRT-PCR screening of fresh feces. muMT mice arrived from The Jackson Laboratory MuAstV-positive and exhibited persistent infections based on consistent qRT-PCR positivity in weekly fecal samples collected for 3 weeks. For all experimental infections in C57BL/6J and CD1 Nude mice, 100 μL of fecal filtrate (SJ001, 100 mg/mL) was used for inoculation by oral gavage. CD1 Nude mice established persistent infections based on consistent qRT-PCR positivity in weekly fecal samples collected for 3 weeks. For experiments in neonates, 1 μL of fecal filtrate (SJ001, 100 mg/mL) was pipetted into the mouth using a p10 pipette on PND 8. Mock infections used PBS alone.

MuAstV inoculum preparation and quantification

Virus inoculum was prepared using homogenized feces from persistently infected CD1 Nude animals resuspended at 100 mg/mL in PBS and blended with zirconium beads using a Next Advance 50 Gold Bullet Blender (6 min, level 4). Fecal homogenates were clarified by centrifugation at 2,800 x g for 10 min and then the supernatant passed through 0.45 μm and then 0.22 μm filtration. To measure virus levels in the inoculum prep and to confirm infections following collection of fresh feces, RNA was extracted using QIAmp Viral RNA Mini Kit (Qiagen) and copies of the MuAstV genome were quantified using a g-block standard (Integrated DNA Technologies) in a one-step qRT-PCR using TaqMan Fast Advanced Master Mix Virus (Applied Biosystems) with primers (F: TACATCGAGCGGGTGGTCGC, R: GTGTCACTAACGCGCACCTTTTCA) and probe ((6-FAM)-TTTGGCATGTGGGTTAA-(MBGNFQ)³⁸ under the following conditions: 50°C for 5 min, 95°C for 20 s followed by 40 cycles of 95°C for 3 s and 60°C for 30 s on a BioRad CFX Connect Real Time System.²⁴

Histological analyses and GAP enumeration

GAP quantifications³⁹ were performed by intraluminal injection of 10 kD rhodamine-conjugated dextran using a 27G tuberculin needle in anesthetized mice (100 mg/kg ketamine with 7.5 mg/kg xylazine). After 20 minutes, mice were sacrificed, and tissues were fixed in 10% neutral buffered formalin and stored at 4°C for 24 hours. Tissues were then washed in PBS and placed in 15% sucrose for 3 hours. Tissues were embedded in O.C.T. compound and 6 μm thick sections were prepared on a Leica CM3050 S cryostat. Tissue sections were stained with DAPI and FITC-conjugated wheat germ agglutinin (WGA) and GAPs identified as dextran-filled columns measuring ∼20 μm (height) x 5 μm (diameter) traversing the epithelium, containing a nucleus and mucus (WGA+). Tiled images were obtained on a Leica DM5500 Widefield Microscope and ∼400 goblet cells were enumerated per weanling animal and ∼100 goblet cells were enumerated per neonatal mouse. For serial image analysis, coverslips were removed by soaking slides in PBS for 60 minutes at RT. To identify infected goblet cells, in situ hybridization was performed with a MuAstV genome-specific probe according to manufacturer specifications (ACD, RNAscope). The slides were then reimaged and analyzed using side by side comparisons between fluorescent images of GAPs and brightfield images of viral RNA to enumerate ∼100 uninfected goblet cells and ∼65 infected goblet cells forming GAPs per animal. Tissue sections from infected and uninfected animals were also analyzed by fluorescence in situ hybridization using an Egfr-specific probe according to manufacturer specifications (ACD, RNAscope). Images were acquired on an Echo Revolve fluorescence microscope. For each animal, ∼89 goblet cells were analyzed for EGFR puncta and identified by rhodamine-dextran and DAPI signal. Image stacks were processed in Fiji/ImageJ. Color channels were split, and background was subtracted using a rolling-ball radius of 15 pixels. A fixed intensity threshold (55-255) was applied uniformly across all images to generate binary masks. Masks were watershed-segmented per image to separate overlapping Egfr puncta. Puncta were quantified by particle analysis within manually defined goblet-cell regions of interest and the same parameters were applied to all samples.

EGFR quantification by ELISA from intestinal epithelial cells

Mouse duodenums were harvested on ice, washed with cold Hanks’ Balanced Salt Solution (HBSS), and incubated in strip buffer (HBSS, 0.75% bovine serum albumin, 15 mM HEPES, 5 mM EDTA, 5 mM DTT) for 20 min at 37 °C. Tissues were passed through a 100 μm cell strainer and washed with cold HBSS. Cells were resuspended in low salt lysis buffer (UltraPure water, 50 mM Tris pH 8.0, 150 mM NaCl, 1.0% Triton X-100) supplemented with Halt protease inhibitor cocktail, incubated at room temperature for 10 min, and centrifuged at 14,000 × g at 4 °C for 15 min. EGFR levels were quantified from cleared lysates using a Mouse EGFR DuoSet ELISA kit (R&D Systems) according to the manufacturer’s instructions.

GAP inhibitors

To close GAPs for short durations we used 3 pharmacological agents: tropicamide, IL-1β, and EGF. Tropicamide is a muscarinic receptor antagonist, which was administered (100 mg/kg in DMSO) in 3 intraperitoneal injections (every 20 minutes) over the course of 1 hour prior to an intraluminal injection with rhodamine-dextran to enumerate GAPs. IL-1β (100 ng in PBS) was administered in a single intraluminal injection for 1 hour prior to the rhodamine-dextran intraluminal injection. EGF (10 μg in PBS) was administered in a single intraluminal injection 20 minutes prior to the rhodamine-dextran intraluminal injection.

Salmonella Typhimurium co-infection

Salmonella enterica serovar Typhimurium (SL1344) expressing eGFP (pFPV25.1)⁴⁰ was grown in LB broth at 37°C shaking at 250 RPM overnight, followed by a 4 hr subculture in LB broth at 37°C shaking at 250 RPM. During the subculture incubation, food was withdrawn from the mice before they were gavaged with a target dose of 1x10⁹ CFU in 100 μl of PBS or PBS alone for mock infection. For GAP quantifications, anesthetized mice were intraluminally injected with ∼1x10⁹ bacteria in 100 μl of PBS or PBS alone for mock infection followed by a 1-hour incubation prior to rhodamine-dextran administration followed by tissue harvest. The actual dose of inoculated bacteria (8.43x10⁸-1.17x10⁹ CFU) was determined by serial dilution on LB agar plates incubated overnight at 37°C. Orally inoculated mice were initially weighed prior to food withdrawal and then daily to track weight loss following the infection. To measure colonization within the intestine and dissemination to the spleen, at 4 dpi small intestines, small intestinal contents, cecums, and spleens were harvested, weighed, and homogenized in 500 μl of PBS and CFU enumerated by serial dilution on LB agar plates incubated overnight at 37°C.

Quantification and statistical analysis

Statistical details of experiments can be found in the figures and figure legends, including the statistical tests used, exact value of n, what n represents (e.g., number of animals, number of cells, etc.), as well as the dispersion and precision measures (e.g., mean, geometric mean, SEM, confidence intervals). To compare across experimental groups, statistical analyses were based on normality tests and sample size and performed using GraphPad Prism version 10.5.0.