Adrenergic signaling in muscularis macrophages limits infection-induced neuronal loss

Summary

Enteric–associated neurons (EAN) are closely associated with immune cells and continuously monitor and modulate homeostatic intestinal functions, including motility and nutrient sensing. Bidirectional interactions between neuronal and immune cells are altered during disease processes such as neurodegeneration or irritable bowel syndrome. We investigated the effects of infection-induced inflammation on intrinsic EAN (iEAN) and the role of intestinal muscularis macrophages (MM) in this context. Using murine models of enteric infections, we observed long-term gastrointestinal symptoms including reduced motility and loss of excitatory iEAN, which was mediated by an Nlrp6- and Casp11-dependent mechanism, depended on infection history, and could be reversed by microbiota manipulation. MM responded to luminal infection by upregulating a neuroprotective program via β₂-adrenergic receptor (β₂-AR) signaling and mediated neuronal protection through an arginase 1-polyamine axis. Our results identify a mechanism of neuronal death post–infection and point to a role for tissue–resident MM in limiting neuronal damage.

Introduction

The gastrointestinal (GI) tract comprises the largest environmental interface of the body and is posed with the unique challenge of maintaining tolerance to dietary and microbial antigens while protecting against pathogen invasion. Coordinated tolerance and resistance mechanisms serve to prevent pathogen dissemination, limit GI damage, and initiate repair responses induced by pathogenic burden or injury (Medzhitov et al., 2012). In the GI tract, infections can result in functional GI disorders post-pathogen clearance, and often enteric neurons, which occupy the intestinal tissue in large numbers, are targeted (Balemans et al., 2017). Clinical observations indicate that between 6–17% of individuals with irritable bowel syndrome (IBS) develop symptoms after episodes of enteric infections; while around 10% of people with bacterial gastroenteritis develop IBS (Holschneider et al., 2011; Ohman and Simren, 2010). Further, HSV-1 and flavivirus strains with neuronal tropism, including West Nile and Zika virus, have also been reported to induce long-term intestinal dysmotility (Khoury-Hanold et al., 2016; White et al., 2018). The clinical presentation of post-infectious IBS includes unresolved low-grade intestinal inflammation, gastrointestinal motility impairment, and nerve damage (Beatty et al., 2014; Holschneider et al., 2011). However, the underlying mechanisms involved in infection–induced neuronal damage are incompletely understood.

The intestinal immune and nervous systems sense and integrate luminal cues, and regulate physiological processes, including gastrointestinal motility. Recent evidence suggests that intestinal–resident macrophage populations play a role in normal functioning of enteric neurons in the absence of infections (De Schepper et al., 2018; Muller et al., 2014). Specifically, muscularis macrophages (MM), located within and surrounding the myenteric plexus, were shown to regulate the activity of enteric neurons and peristalsis via secretion of BMP2 in a microbiota–dependent manner (Muller et al., 2014). Additionally, we previously reported that MM displayed a tissue-protective gene expression profile, a signature that was enhanced following enteric infection, and involved β₂-adrenergic receptor (β₂-AR) signaling (Gabanyi et al., 2016). It remained to be determined, however, whether this neuron-macrophage crosstalk plays a role in infection–induced neuronal damage, or prevention thereof.

By utilizing imaging, cell sorting–independent transcriptomics, pharmacological and genetic gain– and loss–of-function approaches, surgical lesioning, and chemogenetic manipulations, we report that murine enteric infections result in a rapid and persistent loss of intrinsic enteric-associated neurons (iEAN), via the non-canonical inflammasome components Nlrp6 and Casp11, which is associated with long–term GI changes including intestinal dysmotility. Myeloid–specific targeting of β₂-AR, as well as arginase 1 (Arg1), coupled with pharmacological rescue experiments and exogenous activation of gut–sympathetic neurons implicated MM-adrenergic signaling in post-infectious enteric neuronal protection via production of polyamines. Our data uncover a novel mechanism of enteric neuronal cell death and identify a functional role for neuron–macrophage interactions in limiting infection-induced neuronal damage.

Results

Enteric pathogens trigger long-term impairment in GI physiology

Acute bacterial infections, including Salmonella enterica serovar Typhimurium, Shigella dysenteriae, and Campylobacter spp. have previously been linked to post–infectious IBS (Ohman and Simren, 2010). To characterize functional consequences of acute bacterial infection in mice, we used an attenuated strain of S. Typhimurium, spiB (Tsolis et al., 1999), which harbors a mutation in the type-III secretion system, impacting its intracellular replication. We chose to use spiB since wild-type S. Typhimurium rapidly invades and damages the intestinal wall, resulting in mortality in wild-type C57BL/6 mice, barring long-term functional studies post-pathogen clearance. We observed that orally administered spiB was undetected in the feces by 7–10 days post-infection (dpi) (Figure 1A, B). Infection with spiB caused mild intestinal inflammation as evidenced by increased levels of fecal lipocalin-2 (Lcn-2) which remained elevated after pathogen clearance (Figure 1C). This persistent inflammatory response was associated with lasting GI functional changes, including increased ileal ring contractility and a persistent delay in gastrointestinal transit time (GITT) (Figure 1D, E).

Figure 1: Enteric Infections Trigger Intrinsic Neuronal Loss and Dysmotility.

(A) Experimental design for (B-I). (B-H) C57BL/6J mice were orally gavaged with PBS or 10⁹ colony forming units (CFU) of Salmonella spiB. (B) Quantification of fecal CFU. (C) Quantification of fecal lipocalin-2 (Lcn-2) levels as assessed by ELISA post-infection. (D) Ileal ring myography assessed on day 18 post-infection. (E) Total gastrointestinal transit time. Experiments were ended at 450 min (dashed line); (F-I) Neuronal quantification as assessed by IF staining (ANNA-1) on the indicated days post-infection in the (F) ileum (G) colon, (H) duodenum MP; (I) ileum SMP. (J) Neuronal quantification (ANNA-1 staining) in the ileum myenteric plexus of C57BL/6J mice orally gavaged PBS or wild-type Salmonella Typhimurium (S. tm) 21 days post-spiB infection and sacrificed on day 10 post-re-infection. Shaded area indicates mean day 7 iEAN numbers +/− SEM of all control C57BL6/J mice (F). (K) Neuronal quantification in the ileum myenteric plexus of C57BL/6J mice orally gavaged with PBS, Y. pseudotuberculosis (Y.p.), spiB or spiB or Y.p. followed by infection with Y.p. or spiB, respectively and sacrificed 10 days post-secondary infection. (L) Neuronal quantification in the ileum myenteric plexus of C57BL/6J mice reconstituted with a normal microbiota or not after spiB infection. MP – Myenteric plexus; SMP – Submucosal plexus; Data are representative of at least 5 mice per condition. Data were analyzed by unpaired t-test or ANOVA with Tukey’s posthoc test and are shown as mean ± SD; n.s. - not significant, *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001. See also Figure S1.

Changes in contractility and GITT can be associated with altered neuronal activity and nerve damage (Travagli and Anselmi, 2016). Intrinsic enteric-associated neurons (iEAN), commonly referred to as the enteric nervous system are organized in two distinct networks, the submucosal or Meissner’s and myenteric or Auerbach’s plexuses (Furness et al., 2013). In order to determine the impact of intestinal infection on iEAN, we quantified enteric neurons along the GI tract, in both the submucosal and myenteric plexuses. Following spiB infection, we observed a 20–30% reduction in myenteric neurons 7 dpi in the ileum and colon, both of which are major sites of Salmonella invasion; reduced neuronal counts were observed up to 126 dpi (Figure 1F, G). In contrast, neuronal numbers were preserved in the proximal small intestine where Salmonella invasion normally does not occur, and the submucosal plexus; additionally, heat-killed spiB did not result in myenteric iEAN loss (Figure 1H–I and Figure S1A). Loss of iEAN was also observed after infection with another enteric bacterial pathogen, Yersinia pseudotuberculosis (Figure S1B). We further interrogated whether we could recapitulate our findings in the context of the clinically relevant protozoan pathogens Toxoplasma gondii and Trypanosoma cruzi, known to induce GI-related pathologies (Dutra et al., 2009). These protozoans also induced intestinal inflammation, increased ileal ring contractility, delayed GITT and significant neuronal loss (Figure S1C–G and data not shown). By contrast, infection with the intestinal helminth Strongyloides venezuelensis, that displays distinct duodenal tropism (Esterhazy et al., 2019), did not lead to iEAN loss in either the proximal or distal intestine (Fig S1H). Together, these data indicate that iEAN loss may be a conserved feature of a many, though not all, enteric infections; we chose to continue with spiB as an infection model due to the inherent neurotropism of T. gondii and T. cruzi.

While it is traditionally thought that mammalian enteric neurogenesis ceases postnatally, recent studies suggest continuous turnover of iEAN, for instance by replenishment from Nestin+ stem cells (Kulkarni et al., 2017). However, in the present study, neither were we able to detect iEAN number recovery up to four months following spiB clearance, nor did we observe changes in Nestin^GFP+ cell networks in the myenteric plexus (Figure S1I). As an orthogonal approach, we evaluated the involvement of enteric glia, which have been shown to replenish iEAN in response to injury (Laranjeira et al., 2011), in neuronal maintenance after spiB infection. Using an inducible Sox10^RiboTag fate mapping strategy, we found that approximately 10% of iEAN were HA+ four months following tamoxifen administration to adult Sox10^RiboTag control mice, suggesting a glial origin of these cells. However, we did not observe iEAN recovery, nor changes in the frequency of HA+ in the ileum myenteric plexus of Sox10^RiboTag mice infected with spiB in the same period of time (Figure S1J). These results indicate that without further manipulations, iEAN loss is persistent in this model of infection.

Given the significant reduction in ileal and colonic myenteric neurons resulting from a single spiB infection, we asked whether subsequent infections would exacerbate iEAN damage. Following clearance of spiB, we infected mice with wild-type Salmonella, after which we observed no additional iEAN loss, indicating a possible restructuring or adaptation of tissue cells preventing further damage (Figure 1J). To exclude the contribution of adaptive immunological memory, we performed infection-challenge with different pathogens, spiB and Y. pseudotuberculosis, bacterial pathogens that resulted in iEAN loss post primary infections. Primary infection with Y. pseudotuberculosis resulted in iEAN protection against secondary challenge with spiB, and conversely, initial iEAN loss with primary spiB infection was not enhanced with subsequent Y. pseudotuberculosis infection (Figure 1K). The above results demonstrate that infections with different enteric pathogens induce prolonged low-grade inflammation and GI dysmotility, which are associated with a rapid and self-limiting, yet persistent loss of iEAN.

Human post-infectious IBS generally leads to long-term, yet reversible GI symptoms and neuropathy (Beatty et al., 2014; Ohman and Simren, 2010). However, here, we did not observe recovery of neurons after a single infection. Because of reported effects of microbiota, including recent work from our group, on the development and maintenance of enteric neurons (Muller et al., 2019; Obata and Pachnis, 2016), we assessed whether persistent changes in microbial composition (dysbiosis) could explain the lack of iEAN recovery post infection. We infected SPF mice with spiB, administered broad-spectrum antibiotics after bacterial clearance, and subsequently re-colonized the animals with SPF microbiota. This reconstitution of a “normal” or pre-infection microbiota allowed for iEAN recovery, effects that were correlated with a reduction in a member of the Erysipelotrichaceae family and Bacteriodales order when compared to untreated–infected mice (Figure 1L, S1K, L). These data point to a role for the gut microbiota in the recovery of iEAN post infection.

Enteric infection–induced neuronal loss is subtype-specific

iEAN comprise a numerous and heterogeneous population of neurons that monitor and respond to various environmental cues, including mechanical stretch and luminal metabolites (Furness et al., 2013). We first asked whether the observed loss of iEAN could be explained by a decrease of ELAVL3/4 protein expression. We interbred pan-neuronal Snap25^Cre, with Rpl22^lsl-HA (RiboTag) mice (Sanz et al., 2009), which express a hemagglutinin (HA)–tagged ribosomal subunit 22. Immunofluorescence (IF) analysis of HA+ cells in the myenteric plexus confirmed identical loss of iEAN as previously determined by ANNA-1 staining (Figure 2A) (Muller et al., 2019). In cases of overt intestinal inflammation, an indiscriminate loss of iEAN has previously been reported (Mawe, 2015). To investigate whether, in the case of enteric infection, specific neuronal subsets were preferentially lost, we used confocal IF imaging to evaluate the impact on excitatory and inhibitory iEAN populations (Zeisel et al., 2018). Imaging of the ileum myenteric plexus from Slc17a6^td-Tomato (VGLUT2) reporter mice revealed a significant reduction in total number and percentage of excitatory VGLUT2+ neurons post spiB infection (Figure 2B). In contrast, quantification of inhibitory nNOS+ and somatostatin (SST)+ iEAN revealed no changes (nNOS), and a relative increase (SST) post-infection (Figure 2C, D, Figure S2A, B). These data suggest a preferential loss of excitatory neuronal subsets post infection, resulting in changes in the neurochemical representation of iEAN, and may provide an explanation for the observed GI functional changes.

Figure 2: Enteric Infections Cause Subtype-Specific Neuronal Loss.

(A-D) Mice were orally gavaged with PBS or spiB and analyzed 7 dpi. (A) Left: Representative confocal IF images of the ileum myenteric plexus stained with anti-ANNA-1 (red) and anti-HA (green) from Snap25^RiboTag mice treated with (top) PBS or infected with (bottom) spiB. Right: Quantification of the (top) number of HA+ neurons per mm² or (bottom) percent HA/ANNA-1 overlap. (B) Left: Representative confocal IF image of the ileum myenteric plexus stained with anti-ANNA-1 (grey) from VGLUT2^td-Tomato mice treated with (top) PBS or infected with (bottom) spiB. Right: Quantification of the number of VGLUT2+ neurons per mm² and as a percentage of ANNA-1+ neurons. (C, D) Quantification of the number of (C) nNOS+ and (D) somatostatin (SST)+ neurons per mm² and as a percentage of ANNA-1+ neurons. Data are representative of at least 3 mice per condition. Data were analyzed by unpaired Student’s t-test and are shown as mean ± SD; *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001. See also Figure S2 and Supplemental Information.

iEAN loss post–enteric infection is NLRP6– and caspase 11–dependent

To gain insights into a possible mechanism involved in infection-induced iEAN loss, we used neuronal–specific translating ribosomal affinity purification (TRAP) using Snap25^RiboTag mice, which allow neuronal-specific immunoprecipitation of actively-translated mRNA (Muller et al., 2019). We compared iEAN with extrinsic neurons in the nodose ganglia (NG) of Snap25^RiboTag mice. Expression of HA–tagged ribosomes was confirmed in neurons of the myenteric plexus and NG of Snap25^RiboTag mice (Figure S3A). RNA sequencing analysis of immunoprecipitated intact mRNA bound to HA–tagged ribosomes (TRAP-seq) revealed iEAN–specific enrichment in genes encoding neuropeptides, such as Neuromedin U (Nmu), as well as components of the inflammasome pathway (both canonical and non-canonical) including Nlrp6, Pycard, Casp1, and Casp11 (Figure 3A). These data indicate that iEAN possess the machinery for engaging an inflammasome– and Casp1/11–mediated cell death. Imaging analyses for the expression of the inflammasome adaptor ASC (PYCARD) using anti-ASC antibody in WT mice, or Rosa26^ASC:mCitrine (ASC^mCitrine) reporter mice, confirmed the expression of this inflammasome component by iEAN of both naïve and spiB–infected mice (Figure S3B, C). To visualize the Nlrp6 expression pattern in iEAN, we performed fluorescence in situ hybridization (FISH) on ileum and colon sections using RNAscope® probes specific to Elavl4 (pan-neuronal) and Nlrp6. We observed dense localization of Nlrp6 transcripts in the epithelium of the ileum and colon, similar to what has previously been reported (Levy et al., 2015). In addition, we visualized Nlrp6 transcripts in areas of Elavl4–expressing cells in the muscularis, supporting the expression of Nlrp6 by myenteric neurons (Figure 3B). Analysis of data from single cell transcriptional profiling of iEAN (Zeisel et al., 2018) indicated that Nlrp6 is highly enriched in excitatory iEAN compared with additional iEAN subsets (Figure S3D), providing an explanation for preferential loss of excitatory iEAN subsets during enteric infections.

Figure 3: Infection–Induced iEAN Death is Dependent on NLRP6 and Caspase 11 Expression.

(A) Volcano plot of differentially-expressed genes of TRAP-seq from neurons of the nodose ganglion (NG) and ileum myenteric iEAN isolated from Snap25^RiboTag mice. Grey dots highlight all genes analyzed; red dots highlight genes significantly differentially expressed in ileum iEAN. Number of samples are indicated in parentheses. (B) Fluorescence in situ hybridization RNAScope® with probes for Nlrp6 and Elavl4 (pan-neuronal) in the ileum and colon from C57BL6/J mice. iEAN are outlined with a dashed line. Scale bars = 50 μm. Images representative of at least n=2 per group. (C) Scheme of the NLRP6 inflammasome pathway. (D) Neuronal quantification as assessed by IF staining (ANNA-1) in the ileum myenteric plexus on day 7 post-gavage of PBS or spiB of Casp1^–/– Casp11^–/– (ICE^–/–) mice or heterozygous controls (ICE^+/–). (E) Neuronal quantification (ANNA-1 staining) in the ileum myenteric plexus of Casp11^–/– mice on day 7 post-spiB or -PBS gavage. (F) Scheme of i.v. AAV9–mediated transduction of peripheral neurons. (G, H) Representative confocal IF images of ileal (G) and colonic (H) myenteric plexus of Casp11^flox/flox mice given i.v. injection of AAV9-hSyn-Cre-GFP or control viruses, stained with anti-ANNA-1 (red) antibodies. (I, J) (left) Neuronal quantification in the myenteric plexus of ileal (I) and colonic (J) segments from iEAN^ΔCasp11 or Casp11^flox/flox mice on day 7 post-spiB infection. (right) Quantification of the number of Cre-GFP transduced cells in the ileum myenteric plexus of iEAN^ΔCasp11 mice. (K) Neuronal quantification in the ileum myenteric plexus from Snap25Cre^–, Snap25^Nlrp6flox/+ or Snap25^ΔNlrp6 mice on day 7 post-spiB infection. Data are representative of 3–6 mice per condition. Data were analyzed by unpaired t-test or ANOVA with Tukey’s posthoc test and are shown as mean ± SD; n.s. - not significant, **p ≤ 0.01, ***p ≤ 0.001, ****p≤ 0.0001. See also Figure S3 and Supplemental Information.

To evaluate if iEAN loss during infection is caspase-mediated, we first systemically administered a pan caspase inhibitor (zVAD-FMK) spiB-infected mice, which resulted in a reduction of iEAN loss (Figure S3E). We then directly addressed the role of caspases 1 and 11 in infection–induced iEAN loss by infecting caspase 1/11 (ICE)-deficient or haplosufficient (ICE^+/–) mice with spiB. While ICE^+/– mice exhibited pronounced iEAN loss 7 dpi, spiB-infected ICE^–/– mice were completely protected from neuronal loss, despite similar bacterial clearance patterns (Figure 3C, D, and Supplemental Information). We further dissected the role of Casp1 and 11 separately by utilizing the 129S1/Sv mouse strain, which carries an inactivating mutation in the Casp11 locus (human Casp4) (Kenneth et al., 2012). 129 mice infected with spiB exhibited no loss of colonic iEAN numbers as compared to non-infected controls (Figure S3F–J). This iEAN protection was independent of the ability of 129 mice to survive WT Salmonella infection due to expression of functional Nramp (Brown et al., 2013), as CBA/J mice, which also express a functional Nramp, but do not carry a Casp11 mutation, exhibited significant neuronal loss when infected with spiB, despite similar clearance of the bacteria (Figure S3F–K, and data not shown). Furthermore, spiB infection did not result in iEAN loss in single Casp11–deficient mice (Figure 3E). These data suggest Casp11–mediated cell death as a main mechanism involved in iEAN loss following Salmonella infection.

To define whether the NLRP6 inflammasome and Casp11-mediated mechanisms proposed above are neuron-intrinsic, we utilized two in vivo approaches of neuronal-specific gene deletion. Intravenous (i.v.) injection of AAV9 with neuronal-specific promoters can target iEAN, (Gombash et al., 2014), which we confirmed with i.v. administration of AAV9-hSyn-eGFP-Cre to Rosa26^lsl-tdTomato reporter mice (Fig. S3L). To remove Casp11 from iEAN, we i.v.-injected Casp11^flox/flox mice with AAV9-hSyn-eGFP-Cre or control AAV9-hSyn-eGFP viruses (Fig. 3F–H). We detected Cre–expressing AAV9 in 60–90% of myenteric iEAN (Fig. 3G–J). Upon infection with spiB, we found no iEAN loss in mice injected with Cre–expressing AAV9 as compared to mice receiving control virus (Fig. 3I, J), suggesting that neuronal–intrinsic Casp11 expression is required for Salmonella-induced iEAN loss.

To address whether post–infectious iEAN loss is dependent on neuronal NLRP6, we crossed Snap25^RiboTag mice with Nlrp6-floxed (Nlrp6^flox/flox) mice (Fig. S3M). We infected Snap25^ΔNlrp6, Snap25^{RiboTagNlrp6flox/+} heterozygous and (Cre-) Nlrp6^flox/flox control mice with spiB and quantified iEAN numbers. While both heterozygous and Cre^– controls showed reduced iEAN numbers upon spiB infection, Snap25^ΔNlrp6 did not display iEAN loss (Fig. 3K). Together, the above data implicate neuronal–specific NLRP6 and Casp11 as main effectors of iEAN death following Salmonella infection.

MM respond to luminal infection to limit neuronal damage

Tissue–resident MM, the most abundant immune cell population in the myenteric region, are closely juxtaposed to iEAN, and have been linked to normal functioning of these neurons during homeostasis (De Schepper et al., 2018; Gabanyi et al., 2016; Muller et al., 2014). Our previous observations suggested that MM possess a tissue-protective gene signature, including expression of Retnla, Mrc1, Cd163 and Il10 at steady state (Gabanyi et al., 2016). To investigate a possible role of MM in infection–induced iEAN loss, we first depleted MM using an antibody blocking colony stimulating factor 1 receptor (CSF1R)–signaling, AFS98 (α-CSF1R), using a dose that preferentially depletes MM over lamina propria macrophages (Muller et al., 2014) (Figure 4A, Figure S4A, B). Continuous α-CSF1R-mediated MM depletion did not impact iEAN numbers during steady state, but resulted in an enhanced iEAN loss upon spiB infection as compared to mice treated with control antibody, despite similar bacterial load (Figure 4B and Supplemental Information). These results indicate that, while short–term depletion of MM does not impact iEAN survival in the unperturbed state, MM may play an iEAN–protective role during enteric infection.

Figure 4: Tissue-Resident MM Respond to Luminal Infection to Limit Neuronal Loss.

(A, B) C57BL/6J mice were orally gavaged with PBS or spiB and treated with anti-CSF1R or IgG isotype control antibodies. (A) Representative confocal IF staining of the ileum myenteric region stained with anti-MHC-II (green) and anti-ANNA-1 (red). Images representative of at least n=5 per condition; (B) Neuronal quantification of ileum myenteric plexus of injection-habituated non-infected and spiB-infected mice receiving anti-CSF1R or IgG isotype control; (C) Representative confocal IF images of the ileum myenteric plexus of tamoxifen-treated Plp1^iDTA or control mice using anti-S100b (green). Samples obtained at 6 days post-gavage of spiB. (D) Neuronal quantification of ileum myenteric plexus from Plp1^iDTA or control mice 6 days post-spiB. (E) Representative confocal IF images from the ileum myenteric plexus of C57BL/6J mice using anti-MHC II (green) and anti-ANNA-1 (red). Samples obtained at 3 hours post-gavage of PBS or spiB. (F) Neuronal quantification in the ileum myenteric plexus of C57BL/6J mice continuously receiving NaCl or salbutamol by osmotic pumps, and orally gavaged PBS or spiB during drug treatment. Samples obtained on day 7 post-infection. (G) Neuronal quantification in the ileum myenteric plexus of C57BL/6J mice receiving salbutamol delivered by subcutaneous osmotic pumps for 14 days. Mice were gavaged PBS or spiB on day 7 post-surgical pump implantation and received anti-CSF1R or IgG isotype control antibodies. (B, F, G) Shaded area indicates mean day 7 iEAN numbers +/− SEM of all control C57BL6/J mice analyzed in Figure 1F. Data are representative of at least n=3 per condition. Data were analyzed by unpaired t-test and are shown as mean ± SD; n.s. - not significant, *p ≤ 0.05, **p ≤ 0.01, ****p≤ 0.0001. See also Figure S4 and Supplemental Information.

Glial cells, both in the CNS and periphery, have been reported to mediate neuronal protection (Brown et al., 2016). We thus evaluated their involvement in infection–induced iEAN loss or limiting thereof. To target enteric glia, we used mice carrying tamoxifen–inducible Cre under the glia–specific Plp1 promotor (Plp1^CreER) and the lox-stop-lox-DTA transgene in the Rosa26 locus (Rao et al., 2017). Depletion of glia in Plp1^iDTA mice before infection with spiB led to a similar loss of iEAN as tamoxifen–treated controls, indicating that enteric glia play neither a neuroprotective, nor -detrimental role in this context (Figure 4C, D).

Next, we used confocal imaging to investigate MM dynamics early upon infection, which revealed a continuing presence and intercalation into ganglia of the myenteric plexus (Figure 4E, Figure S4C). To assess whether inflammatory monocytes recruited from the circulation are involved in infection-induced neuronal damage, we used CCR2^–/– mice, which exhibit an impairment in mobilizing inflammatory monocytes into tissues (Boring et al., 1997). CCR2^–/– mice failed to clear spiB infection, likely reflecting reduced resistance in the mucosal layer (Dunay et al., 2008), and showed a mild acceleration, rather than a delay in GITT (Figure S4D, E). Nonetheless, CCR2^–/– mice showed similar iEAN loss as WT control mice (˜25%), suggesting that the persistent luminal pathogen load does not cause an increase in neuronal damage (Figure S4F). These data highlight that intestinal macrophages recently differentiated from circulating monocyte precursors may contribute to spiB clearance mechanisms but are not required for post–infectious iEAN death. Together, the above results suggest a critical response of resident MM to luminal pathogenic stimulation; while likely not essential for resistance mechanisms, they appear to prevent excessive neuronal damage.

β2-AR signaling in MM constrains infection–induced inflammation and neuronal death

We previously reported that the anti-inflammatory gene-expression profile of MM was further enhanced following enteric infection via β_2-adrenergic receptors (β₂-AR) (Gabanyi et al., 2016). We investigated the capacity of β₂-AR signaling to mediate tissue protection by coupling pharmacological and genetic modulation of this pathway. First, we infected C57BL/6 mice with spiB while continuously administering a selective β₂-AR agonist (salbutamol) via osmotic pumps. While not affecting pathogen load, salbutamol treatment protected mice from iEAN loss (Figure 4F and Supplemental Information). To directly assess the role of MM in salbutamol–mediated neuroprotection, we depleted MM using α-CSF1R in mice receiving salbutamol. While we observed a rescue of iEAN death in mice treated with IgG isotype control antibody, MM depletion led to a loss of salbutamol-mediated iEAN protection (Figure 4G). Together, these data suggest that MM are critical for β₂-AR–mediated iEAN protection.

We complemented these findings using a genetic approach, by interbreeding mice carrying Cre recombinase under the myeloid Lyz2 promoter (LysM^Cre) with Adrb2^flox/flox mice (Hinoi et al., 2008). The specificity of Cre-targeting of macrophages in the intestinal muscularis was confirmed by immunofluorescence and TRAP-seq analysis of LysM^{ΔAdrb2:RiboTag} and LysM^{Adrbfl/:RiboTag} mice which demonstrated macrophage-restricted HA expression, enrichment for macrophage-specific genes and specific loss of Adrb2 (Figure 5A–C). Additionally, while we found occasional ectopic Cre expression in extrinsic ganglia, we were unable to detect iEAN recombination (Figure S5A). We did not detect infiltrating neutrophils, also targeted by Lyz2, in the intestinal muscularis at steady state or early post-infection, suggesting that neutrophil-β₂-AR signaling does not play a major role in this model (data not shown). Moreover, we found no differences in iEAN numbers or GITT in between LysM^ΔAdrb2 and Cre- littermates during homeostasis (Figure S5B, C). LysM^ΔAdrb2 and Cre- littermates exhibited similar spiB clearance upon infection (Figure 5D). However, LysM^ΔAdrb2 mice had further increased fecal Lcn-2 levels post infection, suggesting a role for myeloid-specific β₂-AR signaling in modulating infection–induced inflammation (Figure 5E). This was accompanied by prolonged GITT, which persisted long-term post-spiB clearance, an increased loss of iEAN and enhanced alterations in ileal ring contractility (Figure 5F, G, Figure S5D). We then infected LysM^ΔAdrb2 mice and Cre- littermates receiving continuous salbutamol treatment, which resulted in prevention of iEAN loss in Cre-littermates carrying intact myeloid β₂-AR. However, this neuroprotection was absent in salbutamol–treated LysM^ΔAdrb2 mice (Figure 5H). These data establish a role for β₂-AR signaling in intestinal macrophages that contributes to a neuroprotective program following enteric infection.

Figure 5: Loss of β₂–AR Signaling in MM Exacerbates Neuronal Damage Following Enteric Infection.

(A) IF staining of MM from the ileum of LysM^RiboTag mice using anti-MHC II (green) and anti-hemagglutinin (HA, red) antibodies. White arrows indicate MHC II+ HA- macrophages. Scale bars = 50 μm. Images representative of n=2 mice per group. (B) Volcano plot of differential expression analysis (DEseq2) of TRAPseq from LysM^{Adrb2fl/+:RiboTag} mice comparing input (ileum tissue) to immunoprecipitated (i.p.) transcripts (red dots) All i.p. samples indicated by red dots are log2FoldChange > 0.5 and padj > 0.05. Relevant macrophage genes highlighted in black text and Nlrp6 (neuronal specific) highlighted in blue text. (C) Transcripts per million (TPM) as calculated by Kallisto alignment of Adrb2 transcript comparing input (ileum tissue) or immunoprecipitated transcripts from LysM^{Adrb2fl/+:RiboTag} and LysM^{ΔAdrb2:RiboTag} mice. (D-H) LysM^ΔAdrb2 and wild-type (WT) littermate control mice were orally infected with spiB. (D) Quantification of fecal colony forming units (CFU) on the indicated days post-infection (E) Quantification of fecal lipocalin-2 (Lcn-2) on the indicated days post-infection. (F) Total gastrointestinal transit time; experiments were ended at 450 min (dashed line); (G, H) Neuronal quantification in the ileum myenteric plexus on day 7 post-spiB infection of (G) LysM^ΔAdrb2 mice and WT littermates; shaded area indicates mean day 7 iEAN numbers +/− SEM of non-infected LysM^ΔAdrb2 mice and WT littermates (Figure S5B); (H) LysM^ΔAdrb2 and WT littermates receiving salbutamol via subcutaneous osmotic pumps and infected with spiB post-implantation; shaded area indicates mean day 7 iEAN numbers +/− SEM of non-infected LysM^ΔAdrb2 and WT littermate mice (Figure S5B). Unless indicated otherwise, data are representative of 3–6 mice per condition, were analyzed by unpaired t-test or ANOVA with Tukey’s posthoc test and are shown as mean ± SD; *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p≤ 0.0001. See also Figure S5 and Supplemental Information.

Local sympathetic activation mediates iEAN protection during enteric infection

We noted that mice i.p.-injected multiple times with IgG isotype antibody (as in MM depletion experiments) or mice anesthetized with isoflurane (as in control pump experiments) did not lose usual ˜25% iEAN numbers post spiB infection, suggesting that stress–induced local or systemic catecholamine release could trigger the same protective pathway in MM. Indeed, a short pulse of isoflurane anesthesia or i.p. injections of PBS or IgG isotype antibody were sufficient to prevent iEAN death following spiB infection, indicating that stress signals able to potentiate β₂-AR signaling may help prevent post–infectious iEAN loss (Figure S6A, B). In light of these findings, we sought to determine whether systemic release of catecholamines from the hypothalamic-pituitary-adrenal axis could play a role in neuronal protection during infection. We performed bilateral adrenalectomy on wild-type mice and compared iEAN loss upon spiB infection with sham-operated controls. Rather than enhanced iEAN loss, as observed in LysM^ΔAdrb2 mice, adrenalectomized animals showed reduced neuronal loss, suggesting that systemic catecholamines were not required for iEAN protection, and furthermore, that additional stress hormones produced by the adrenal glands (corticosteroids) could play a detrimental role in enteric neuronal survival during intestinal infections (Fig. 6A–C).

Figure 6: Activation of Gut Sympathetic Neurons Protects iEAN Death Following Enteric Infection.

(A) Scheme of hypothalamic-pituitary-adrenal (HPA) axis and local sympathetic innervation of the intestine. (B, D, E) Mice were orally infected with spiB. Neuronal quantification in the ileum myenteric plexus on day 7 post-spiB infection. (B) Adrenalectomized or sham-operated C57BL/6J mice. (C) Images demonstrating successful removal of adrenal glands from C57BL/6J mice. (D, E) SNS^hM4Di (D) or Phox2b^hM4D (E) mice were treated with 10 mg/kg Compound 21 (C21) 24 hours before spiB infection. NE – noradrenaline, E – epinephrine (adrenaline), CG-SMG – celiac-superior mesenteric ganglion; Data are representative of at least n=3 per condition. Data were analyzed by unpaired t-test and are shown as mean ± SD; **p ≤ 0.01, ***p ≤ 0.001. See also Figure S6 and Supplemental Information.

We previously demonstrated that spiB infection led to a significant increase in the number of recently activated, cFos+ sympathetic neurons in the celiac-superior mesenteric ganglion (CG-SMG)(Gabanyi et al., 2016). We thus assessed whether stimulation of these neurons is sufficient to drive MM-β₂-AR-mediated tissue protection. We utilized two chemogenetic mouse models in which administration of the synthetic ligand, Compound 21, to mice carrying an inactivating DREADD driven by promoters active in sensory nodose ganglia neurons (SNS^hM4Di or Phox2b^hM4Di) result in a significant increase in cFos expression by gut-projecting sympathetic neurons (Muller et al., 2019). We found that DREADD-induced activation of gut-projecting neurons prior to spiB infection both in SNS^hM4Di and Phox2b^hM4Di mice significantly rescued iEAN loss (Fig. 6D, E). These results provide further support of a role for locally–released catecholamines and β₂AR signaling in MM in limiting inflammation–induced iEAN loss.

An adrenergic-arginase 1-polyamine axis in MM limits infection–induced neuronal loss

To dissect possible mechanisms by which MM β₂-AR signaling is involved in preventing excessive tissue damage post–infection, we analyzed LysM^ΔAdrb2 MM gene expression profiles. MM sorted from Cre- littermates responded to spiB infection by upregulating tissue protective genes upon infection, but we did not observe this upregulation in MM sorted from infected LysM^ΔAdrb2 mice (Figure 7A). Arginase 1 (Arg1) is known to mediate the production of neuroprotective polyamines such as spermine (Cai et al., 2002), which in turn was described to suppress NLRP6–inflammasome activation (Levy et al., 2015). To investigate the participation of polyamines in iEAN cell death, we supplemented the drinking water of spiB-infected mice with spermine or with Difluoromethylornithine (DFMO), which inhibits polyamine biosynthesis by selective and irreversible inhibition of ornithine decarboxylase 1 (ODC1) (Koomoa et al., 2013). Bacterial load and clearance patterns were similar in either treatment condition; however, mice that received DFMO exhibited enhanced iEAN loss, while those receiving spermine showed a significant rescue post-spiB infection (Figure 7B, C, and Supplemental Information). Finally, we genetically addressed whether Arg1 activity in MM was required for their protective role in infection–induced neuronal damage in the intestine by interbreeding LysM^Cre and Arg1^flox/flox mice (El Kasmi et al., 2008). We did not observe differences in iEAN numbers in LysM^ΔArg1 mice as compared to Cre^–controls during homeostasis (Figure S7A). Upon spiB infection, LysM^ΔArg1 also showed similar pathogen load and clearance patterns, but a trend to increased GITT (Figure S7B and Supplemental Information). However, loss of myeloid Arg1 heightened iEAN loss post-spiB infection (Figure 7D). Together, these results point to a functional role for a MM β2-AR–Arg1-polyamine axis in in limiting infection–induced enteric neuronal cell death.

Figure 7: An Adrenergic-Arginase 1-Polyamine Axis in MM Limits Infection-Induced Neuronal Death.

A) mRNA expression analysis of the indicated genes by quantitative real-time PCR in LysM^ΔAdrb2 and wild-type (WT) littermate control mice non-infected (NI) or 2 hours post-infection with spiB. (B-D) Neuronal quantification in the ileum myenteric plexus on day 7 post-infection with spiB of (B) C57BL/6J mice with access to regular drinking water or water supplemented with difluoromethylornithine (DFMO) over the course of infection; shaded area indicates mean day 7 iEAN numbers +/− SEM of all control C57BL6/J mice (Fig. 1F); (C) C57BL/6J mice receiving regular drinking water or water supplemented with spermine over the course of infection; shaded area bounded by dashed lines indicates mean day 7 iEAN numbers +/− SEM of all control C57BL6/J mice; (D) LysM^ΔArg1 mice and WT littermates. Shaded area indicates mean day 7 iEAN numbers +/− SEM of non-infected LysM^ΔArg1 and WT littermate control mice (Figure S7A). Data are representative of 3–6 mice per condition. Data were analyzed by unpaired t-test or ANOVA with Bonferroni correction and are shown as mean ± SD; *p ≤ 0.05, **p ≤ 0.01, ****p≤ 0.0001. See also Figure S7 and Supplemental Information.

Discussion

Inflammation–induced neuronal damage

Overt inflammation in response to infection can result in chronic immunopathology. In the GI tract, bacterial, viral and protozoan infections are associated with intestinal neuropathy and the development of post-infectious IBS (Beatty et al., 2014; Holschneider et al., 2011; Ohman and Simren, 2010; White et al., 2018). However, the understanding of underlying mechanisms involved in this process is largely lacking. Here, we provide evidence that muscularis macrophages located in close proximity to enteric neurons prevent infection– or inflammation–induced tissue damage; perhaps an analogous function to that described for microglia in the CNS (Davalos et al., 2005; Wang et al., 2015).

Most mechanisms proposed for inflammation–induced neuronal damage thus far are related to CNS neuroinflammation. The apparent link observed in our study, between having a “normal” microbiota and iEAN recovery offers an interesting direction for studies of central and peripheral neuropathies. For instance, what are the mechanisms utilized by the gut microbiota to maintain, or rescue, neuronal populations? Do specific bacterial species play a beneficial or detrimental role in this process? Although certain inflammatory processes are essential for neuronal function and tissue physiology, including neuronal pruning and tissue repair (Salter and Stevens, 2017), excessive inflammation often results in the production of neurotoxic factors such as interleukin 1 beta (IL-1β) and tumor necrosis factor alpha (TNF-α) (Crotti and Glass, 2015). While some of these could play a role in enteric infections, our studies rather point to NLRP6 ligands, or alternative triggers of Casp11 cleavage, as mediators of infection–induced iEAN cell death. These results corroborate previous work describing a protective phenotype in Pycard^–/– mice, which did not lose enteric neurons following DNBS-induced colitis, a process that was shown not to involve the NLPR3 inflammasome (Gulbransen et al., 2012). Unlike the indiscriminate iEAN loss previously observed in chemically–induced murine colitis models (Mawe, 2015), our studies indicate infection-induced location- and subset-specific loss of excitatory iEAN. These data warrant further study and additional subset-specific analysis of EAN in the steady state or under pathological conditions. Nevertheless, we speculate that the mechanism of neuronal cell death involving components of the inflammasome machinery put forward here in the context of enteric infection may be engaged in other contexts. For instance, it is possible that iEAN–mediated production of IL-1β or IL-18, via Casp1/11 activation, participate in pathogen resistance mechanisms (R. Flavell, personal communication). Additionally, heterologous infection experiments suggest the existence of either a form of tissue memory (trained immunity), or a limited subset of enteric neurons expressing such inflammasome machinery – possibilities that warrant further investigation. Furthermore, it remains to be determined whether neurodegenerative processes in the CNS also involve this or analogous pathways (Song and Colonna, 2018).

Monocyte-macrophage lineage and enteric glia in neuronal damage

Previous studies have uncovered a potential role of monocytes, macrophages and microglia in the induction, or modulation of inflammation–induced neuropathy, CNS damage, and subsequent changes in animal behavior (Aguzzi et al., 2013; D’Mello et al., 2013). Tissue-resident macrophage populations in the brain express PRRs and are believed to contribute significantly to dendritic spine remodeling (Parkhurst et al., 2013; Song and Colonna, 2018). The balance between the production of pro-inflammatory and tissue-protective factors by macrophage populations is likely to determine the pathological versus reparative role of these cells. In a model mimicking CNS virus infection, depletion of monocytes abrogated dendritic spine loss (Parkhurst et al., 2013), indicating that in the CNS multiple cells types can mediate similar effects. In the gut, lack of inflammatory monocytes failed to impact iEAN loss, suggestive of a dependence on longer-lived resident macrophage populations, whose presence in the gut were only recently proposed (De Schepper et al., 2018; Shaw et al., 2018). Under conditions of stress or inflammation, activation of enteric glia has also been linked to neuronal damage (Brown et al., 2016; Gulbransen et al., 2012; Mayo et al., 2014), yet little is known about the relevance of glial-immune cell interactions in the GI tract. While our experiments depleting enteric glia did not suggest a major role in the infection–induced iEAN loss, it remains plausible that surrounding enteric glia, or MM, induce neuronal damage in other contexts as previously attributed to microglia in the brain (Jin et al., 2016).

Both glial cells and tissue macrophages are also appreciated to prevent inflammation–induced damage in several contexts, as well as to initiate repair processes post-injury. In models of spinal cord injury, incoming monocytes are thought to mediate a functional recovery via secretion of anti-inflammatory mediators (Shechter et al., 2009), although an enhancement of inflammatory markers in macrophages was also shown to help CNS axonal regeneration (Gensel et al., 2009). Our data favor a neuroprotective role for MM via upregulation of Arg1 and polyamine production; an analogous observation to previous studies suggesting that “alternatively activated” macrophages promote axonal growth or regeneration after CNS injury (Cai et al., 2002; Kigerl et al., 2009). While multiple studies of the CNS find a direct role for polyamines in neuro-regeneration (Mahar and Cavalli, 2018), our data suggest a potential role for these molecules in preventing neuronal cell death modulation of the inflammasome. This possibility is supported by previous studies demonstrating that microbiota–derived spermine regulates NLRP6 signaling and intestinal epithelial cell secretion of IL-18 (Levy et al., 2015), which could also play tissue protective roles in the enteric infection models described here. Nonetheless, macrophages differentiated towards an “alternatively activated” phenotype could also negatively contribute to gut homeostasis during the resolution of inflammatory processes, for instance by enhancing fibrosis (Xue et al., 2015). Finally, because we previously demonstrated that macrophages located in the mucosal layer do not show an immediate upregulation of tissue protective genes upon enteric infection, an anatomically segregated model in which lamina propria macrophages help pathogen resistance and MM boost tolerance is conceivable.

Sensing and regulation of the catecholamine-β₂-AR pathway

The nervous and immune systems are the body’s main sensory interfaces, and both depend on cell-to-cell contacts and on soluble molecules that act on proximal or distant target cells, which include neurons and immune cells themselves (Besedovsky et al., 1983). The involvement of catecholaminergic neurons in anti-inflammatory responses has been demonstrated in different tissues, and are attributed to the broad sympathetic innervation in peripheral lymph nodes associated with expression of β₂-AR in immune cells (Pavlov et al., 2018). Our findings also suggest an intricate interplay between iEAN, macrophages and eEAN, particularly the extrinsic sympathetic ganglia, adding functional significance to observations of interactions between barrier tissue resident macrophages and peripheral neurons (Gabanyi et al., 2016; Muller et al., 2014). The source of catecholamines may be context dependent: enteric infections may result in the activation of extrinsic sympathetic neurons (Gabanyi et al., 2016); under conditions of systemic stress these signals could originate from the adrenal glands, although our adrenalectomy experiments argue against this possibility. The effects of stress are typically thought to be immunosuppressive although in the context of cellular damage and death, glucocorticoids can promote pro-inflammatory immune responses particularly in the CNS (Di Giovangiulio et al., 2015; Salgo et al., 2006). We found that acute stress in mice provided significant protection to iEAN during enteric infection, results that more broadly link physiological stress to control of inflammation. In light of these findings, stress should be considered as a confounding factor to control for in context of inflammation. Nevertheless, both of these pathways require neural input, yet the circuitry connecting luminal sensing to eEAN activation (and a possible involvement of the CNS), and how luminal stimulations can regulate it, remain to be determined (Muller et al., 2019; Pruss et al., 2017). Current literature suggests that sensing of luminal pathogens, or associated changes in the microenvironment, could be mediated by enteroendocrine cells (Kaelberer et al., 2018) or epithelial/mucosal associated sensory afferents (Chiu et al., 2013), both with the capacity to impact iEAN survival or initiate changes in eEAN activity. However, the components of luminal sensing still require further dissection to uncover relevant cell populations and receptors (Veiga-Fernandes and Mucida, 2016). Downstream of the initial components, our data identify a novel mechanism of neuronal cell death and point to a role for tissue–resident MM in limiting neuronal damage. Such intra-tissue adaptation of immune cells appears to help the maintenance of a proper balance between resistance and tolerance and these findings could have implications in additional tissues and contexts.

STAR Methods

Contact for Reagent and Resource Sharing

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Daniel Mucida (mucida@rockefeller.edu).

Experimental Procedures

Mice

C57BL/6J (000664), Casp1^−/− Casp11^−/−(B6N.129S2-Casp1tm1Flv/J, 016621), CCR2^−/− (B6.129S4-Ccr2^tm1lfc/J, 004999), LysM^Cre (B6.129P2-Lyz2^tm1(cre)Ifo/J), Arg1^flox/flox (C57BL/6-Arg1^tm1Pmu/J), Rosa26^tdTomato (B6.Cg-Gt(ROSA)26Sor^{tm14(CAG-tdTomato)Hze}/J), VGLUT2^Cre (Slc17a6^tm2(cre)Lowl/J), Rpl22^HA (B6N.129-Rpl22^tm1.1Psam/J), 129S1(129S1/SvImJ), CBA/J, Cx3cr1^GFP(Cx3cr1^{tm1Litt/LittJ}), Phox2b^cre (B6(Cg)-Tg(Phox2b-cre)3Jke/J), Snap25^Cre(Snap25^{tm2.1(cre)Hze}), NSG (NOD.Cg-Prkdc^scidIl2rgtm1^Wjl/SzJ), Casp11^−/− (Casp4tm1Yuan/J), and R26-CAG-ASC-citrine (B6.Cg-Gt(ROSA)26Sortm1.1(CAG-Pycard/mCitrine*,-CD2*)Dtg/J), Plp1^creERT (B6.Cg-Tg(Plp1-cre/ERT)3Pop/J), Rosa26^lsl-hM4Di (B6N.129-Gt(ROSA)26Sor^{tm1(CAG-CHRM4*,-mCitrine)Ute}), and ROSA-DTA (B6.129P2-Gt(ROSA)26Sortm1(DTA)Lky/J) were purchased from The Jackson Laboratories and maintained in our facilities. SNS^cre (Tg(Scn10a^cre)1Rkun were a gift of R. Kuhner, Nestin^GFP (Tg(Nes-EGFP)33Enik were generously provided by P. Frenette and G. Enikolopov, Adrb2^flox/flox (Adrb2^tm1Kry) by G. Karsenty, Sox10^CreERT2 (Tg(Sox10-icre/ERT2)26Vpa) were generously provided by B. Gulbransen and V. Pachnis, Nlrp6^flox/flox by P. Rosenstiel (targeting of Exon 1 of Nlrp6, frozen sperm generously provided by T. Kanneganti), and Casp11^flox/flox by KOMP and A. Wullaert. Mouse lines were interbred in our facilities to obtain the final strains described in the text. Genotyping was performed according to the protocols established for the respective strains by The Jackson Laboratories, or personal communication with the donating investigators. Mice were maintained at the Rockefeller University animal facilities under specific pathogen-free conditions. Mice were fed a standard chow diet and used at 7–12 weeks of age for most experiments. Animal care and experimentation were consistent with NIH guidelines and were approved by the Institutional Animal Care and Use Committee at the Rockefeller University.

Microorganisms

Salmonella enterica serovar Typhimurium (SL1344) and its mutant spiB were used for infection experiments and cultured prior to infection as described below. Yersinia pseudotuberculosis (IP32777) was cultured prior to infection as described below. S. venezuelensis was maintained in our facility by periodically infecting NSG mice and culturing larvae as described below. Toxoplasma gondii was maintained in our lab by periodically infecting C57BL/6 mice with 5 cysts administrated intraperitoneally (i.p.). Cysts were isolated from brain tissue 30 days after infection. Trypanosoma cruzi. The T. cruzi Y strain was maintained by serial passage from mouse to mouse.

Method Details

Infections

Salmonella enterica Typhimurium. For infections with Salmonella spiB, mice were pre-treated with a single dose of Streptomycin (20 mg/mouse dissolved in 100 μl of DPBS) administered by oral gavage 18–24 hours prior to infection Mice were then orally inoculated with 10⁹ CFU of spiB. For Salmonella re-infection experiments, mice were subjected to spiB infection as described above. 1 week post-clearance of spiB from the feces, mice were fasted for 4 hours and infected with 10⁶ CFU of wild-type Salmonella enterica Typhimurium (SL1344). For all Salmonella infections, a single aliquot of either strain of Salmonella was grown in 3 ml of LB overnight at 37 ºC with agitation. Bacteria were then sub-cultured (1:300) into 3 ml of LB for 3.5 hours at 37 ºC with agitation, and diluted to final concentration in 1 ml of DPBS. Bacteria were inoculated by gavage into recipient mice in a total volume of 100 μl. For experiments using heat-killed spiB, samples subjected to heat treatment (95 °C for 10 min in water bath), prior to gavage and successful inactivation was confirmed by plating a serial dilution made from the suspension and then 5 μL onto Salmonella-Shigella plates.

Yersinia pseudotuberculosis. Y. pseudotuberculosis (strain IP32777) was grown as previously described with some adjustments (Fonseca et al., 2015). Briefly a single aliquot of the strain was grown in 3 ml of 2xYT media overnight at 28°C with vigorous agitation. Mice were fasted for 12 hours prior to infection with 10⁸ CFU by oral gavage.

For heterologous re-infection experiments, mice infected with spiB or Y. pseudotuberculosis as described above were subjected to either 10⁸ of Y. pseudotuberculosis or 10⁹ CFU of Salmonella spiB on day 21 days post-infection with spiB or Y. pseudotuberculosis infection, and sacrificed 10 days post-secondary infection.

Strongyloides. venezuelensis. S. venezuelensis was maintained in our facility in NSG mice by subcutaneous infection with 700 larvae, resulting in chronic infection of this strain. For each experiment, feces of infected NSG mice were collected and spread on Whatman paper, which was placed into a beaker with water and incubated at 28 °C. The hatching larvae were collected over 2–3 days. Mice were infected subcutaneously with 700 larvae/100 μl water per mouse. S. venezuelensis was passaged by periodically infecting naïve adult NSG mice.

Toxoplasma gondii. T. gondii was maintained in the lab by periodically infecting mice administered by intraperitoneal injection of 5 cysts/mouse in a total volume of 100 μl of DPBS.

Trypanosoma cruzi (T. cruzi, Y strain). For each experiment, blood from an infected mouse was collected, parasites were quantified and naïve recipient mice infected by intraperitoneal injection of 10⁴ parasites. Infected mice were sacrificed for tissue analysis at day 22 post-infection, when the parasite load reaches a plateau (Arantes et al., 2004).

Antibodies and flow cytometry

Antibodies used for whole-mount immunofluorescence imaging

The following primary antibodies were used, and unless otherwise indicated concentrations apply to all staining techniques: Somatostatin (1:200, Millipore Sigma, MAB354), ANNA-1 (1:200,000, Gift of Dr. Vanda A. Lennon), HA (1:400, Cell Signaling Technologies, 3724S), nNOS (1:200, ABCAM, ab76067), MHC II (1:400, Millipore Sigma, MABF33), S100 beta (1:200, Abcam, ab52642), ASC (1:200, Adipogen, AL177), GFP (1:400, Nacalai, GF090R). Fluorophore-conjugated secondary antibodies were either H&L or Fab (Thermo Fisher Scientific) at a consistent concentration of 1:400 in the following species and colors: Goat anti-rabbit (AF488/568/647), goat anti-rat (AF488/647), goat anti-chicken (AF488/568/647), goat anti-human (AF568/647),

Antibodies used for cell sorting of intestinal macrophages

Fluorescent-dye-conjugated antibodies were purchased from BD-Pharmingen (USA) (anti-CD45.2, 104; anti-CD45R, RA3–6B2); eBioscience (USA) (anti-CD103, 2E7; anti-MHC II, M5; anti-F4/80, BM8; anti-CD11b, M1/70; anti-CD11c, N418; anti-Siglec F, E50–2440; anti-CD3e,145–2C11; anti-Ly6G, RB6–8C5) or BioLegend (USA) (anti-CD64 X54–5/7.1). Live/Dead staining was performed using Aqua fixable dead cell stain (Invitrogen). Macrophages were sorted as Aqua–CD45+Lin–(CD3–B220–Siglec F–LY6G–) MHCII+F4/80+CD11B+CD11C+CD103–) using a FACS Aria cell sorter flow cytometer (Becton Dickinson).

Intestine dissection

Mice were sacrificed and duodenum (1 cm moving distal from the gastroduodenal junction), ileum (1 cm moving proximal from the ileocecal junction), or colon (4 cm moving proximal from the rectum) was removed. For dissection of the muscularis, following the above procedures, the intestinal tissue was placed on a chilled aluminum block with the serosa facing up (Gabanyi et al., 2016). Curved forceps were then used to carefully remove the muscularis.

Nodose ganglion dissection

Mice were sacrificed and the ventral neck surface was cut open. Associated muscle was removed by blunt dissection to expose the trachea and the nodose ganglion was then located by following the vagus nerve along the carotid artery to the base of the skull. Fine scissors were used to cut the vagus nerve below the nodose ganglion and superior to the jugular ganglion.

Colony forming unit counting

Fecal pellets from Salmonella spiB-infected mice were weighed and then disrupted in 400 μL of DPBS. Serial dilutions were made from the original suspension and then 5 μL of each dilution was plated onto Salmonella-Shigella plates. The plates were then incubated overnight and the number of black colonies were counted for the serial dilution with the clearest delineation of single units. This number was then multiplied by the dilution factor and by 80 to give the number of colony forming units (CFU) in the original suspension. CFU numbers were then divided by the original fecal pellet weight to give the number of CFU per mg of feces.

Intestine motility measurements

For measurement of total intestinal transit time, mice were given an oral gavage of 6% carmine red (Sigma-Aldrich) dissolved in 0.5% methylcellulose (prepared in with sterile 0.9% NaCl). Total intestinal transit time was measured as the time from oral gavage it took for mice to pass a fecal pellet that contained carmine. Experiments were ended after 450 minutes.

Ileal smooth muscle contractility

Ileal ring tension was assessed by organ bath (Radnoti) myography as previously described (Muller et al., 2014). Briefly, distal ileal rings were mounted and equilibrated for one hour and were distended to 0.5 g followed by 10 minutes of relaxation. The data was acquired using the PowerLab acquisition system and analyzed using LabChart Software (AD Instruments).

Lipocalin-2 ELISA

Lipocalin-2 levels in fecal samples were quantified using a Mouse Lipocalin-2 ELISA kit (R&D Systems) according to manufacturer’s instructions.

Drug Administration

Salbutamol

Salbutamol sulfate (Selleck Chemicals) was dissolved in sterile NaCl to a concentration of 56 mg/ml, loaded into osmotic pumps and administered by subcutaneous osmotic pumps implanted as described below at a final dose of 400 μg/day for 14 days.

Pan caspase inhibition

zVAD-FMK (Selleck Chemicals) was dissolved in DMSO and administered at a dose of 50 μg/mouse by daily intraperitoneal injections over the course of a 7-day infection starting 1 day prior to infection.

Polyamine administration

Spermine (Sigma-Aldrich) was administered at a concentration of 2% in drinking water. Spermine-substituted drinking water was changed daily and fluid intake was monitored. Treatment was started 3 days prior to infection with Salmonella spiB and continued over the course of experiments.

ODC1 inhibition

Difluoromethylornithine (DFMO, gift from P. Woster, MSSM) was dissolved in tap water, filter-sterilized, and administered at a concentration of 4% in drinking water. DFMO-substituted water was changed daily, and fluid intake was monitored. Treatment was started 4 days prior to infection with Salmonella spiB and continued over the course of experiments.

Isoflurane induction

A plexiglass induction chamber was placed on a heating pad set to 37°C. Isoflurane was set at 1% with an oxygen flow of 1 liter/minute. Mice were placed in the induction chamber and monitored for loss of movement, after which a period of 15 minutes was elapsed. Respiration was monitored continuously during this period. Following isoflurane treatment, mice were returned to their home cage for a period of 1 hour prior to receiving Streptomycin pre-treatment, followed by spiB infection as described above.

Sham intraperitoneal injections

Sham intraperitoneal injections of DPBS or IgG (isotype control from MM depletion experiments, see below) were performed by mimicking the anti-CSF1R MM depletion protocol, by intraperitoneal injection of 200 μl of DPBS or IgG/ mouse (100 μl/flank). Injections were performed 12 hours prior to spiB infection, and an additional injection was performed at day 3.5 post-infection. Mice were sacrificed and tissue was analyzed at 7 days post-infection.

Subcutaneous pump implantation

Mice were anesthetized using isoflurane and, under sterile conditions, a small incision was made on the dorsal side of the neck. An Alzet pump (Model 1002) filled with Salbutamol sulfate prepared as described above was placed under the skin and moved back towards the right flank. The incision was then closed with surgical wound clips (Kent Scientific). Mice were then left to rest for 7 days before infection with spiB.

Adrenalectomy

Mice were anesthetized using isoflurane and, under sterile conditions, bilateral incisions were made dorsally just below the rib cage. The adrenal glands were then grasped by forceps from the suprarenal blood vessels and extracted from the surrounding tissue. Care was taken to minimize any bleeding before the muscle incision was closed with absorbable sutures. Surgical wound clips were then used to close the skin. Mice were then kept on 0.9% saline drinking water for 9 days prior to infection with spiB.

Generation of Nlrp6^flox/flox mice

The Nlrp6^flox/flox strain was generated by Flp-Neo-mediated insertion of loxP sites flanking exon 1 of the Nlrp6 locus. ES cells were used to generate the conditional knockout allele. The strain was generated in collaboration with genoWay (Lyon, France).

16S sample processing

16s samples were processed utilizing a Promega Maxwell® RSC 48 Instrument. Following RNA extraction from all samples, samples were quantified using a ThermoFisher Quant-It dsDNA High-Sensitivity Kit on a microplate reader.

16S sequencing

16S sequencing was performed on either the Illumina HiSeq 2500 or the Illumina MiSeq depending on project-specific needs. Raw paired-end fastq files containing sequence reads were merged at the overlapping region to produce a single 16S contig. All merged sequences having more than 1 expected error per read were filtered. Operational taxonomic units (OTU) were generated by clustering sequences with a 99% correspondence and chimera sequences were removed using usearch (Edgar, 2010) (v11).

Chemogenetics

Mice expressing hM4Di in relevant neuronal populations were injected with 10mg/kg Compound 21 (HelloBio) 24 hours prior to oral gavage with spiB.

Antibiotic-mediated microbiota depletion and recolonization

7 days post-oral gavage with spiB mice were given drinking water supplemented with 10g/L Splenda (artificial sweetener, control), or broad-spectrum antibiotics (1g/L ampicillin, 0.5g/L vancomycin, 1g/L neomycin, 0.5g/L metronidazole) in the drinking water supplemented with 10g/L Splenda for 2 weeks. Mice on antibiotics were then put onto Splenda-containing water and all mice were given feces from an uninfected cage of age-matched C57BL/6J mice. All tissues were then harvested 2 weeks after recolonization.

Anti-CSF1R (ASF98) antibody production

Anti-CSF1R was produced by suspension culture of the ASF98 hybridoma cell line (gift from Miriam Merad). Cells were thawed and passaged twice (1:10) in PFHM-II (Thermo Fisher) media supplemented with P/S (Thermo Fisher). Cells were then seeded at 5×10⁶ in 15 mL of PFHM-II in the cellular compartment of a 1000 mL bioreactor (Wheaton) and allowed to grow for 7–10 days. The cell compartment was then harvested, spun down and the supernatant collected. Supernatant was filtered and antibody was purified from the supernatant by affinity purification using protein G sepharose (GE Healthcare) in a gravity column (Bio-Rad). Briefly, the protein G was equilibrated with 100 mL binding buffer (Thermo Fisher Scientific) and supernatant was loaded onto and run through the column. The column was then washed with 200 mL of binding buffer and eluted with 10 mL elution buffer (Thermo Fisher Scientific). Antibody was eluted in fractions directly into 100 μl 1 M Tris-HCL pH 8 (Invitrogen) and the concentrations were measured on a Nanodrop 2000 spectrophotometer (Thermo Fisher Scientific). Fractions were combined and dialyzed in 2 L DPBS for 2 days at 4°C using 14,000 MWCO dialysis tubing (Spectrum Labs) changing once. The antibody was then concentrated using centrifugal filters (30,000 MW, Amicon) and stored at 4°C until use.

MM depletion using anti-CSF1R

Anti-CSF1R was diluted in sterile DPBS to a final concentration of 6.25 mg/ml and 50 mg/kg of anti-CSF1R or isotype control (IgG from rat serum, Sigma-Aldrich) were administered by i.p. injections. Mice were previously habituated to i.p. injections for at least 5 days prior to the start of depletion and spiB infection. For all experiments, depletion was performed twice over the course of 7 days (day 0 and day 3.5). For infection experiments, depletion was performed 12 hours prior to infection. Animals were then subjected to the Salmonella spiB infection protocol as described above, and received and additional dose of either anti-CSF1R or isotype control 3.5 days after the initial dose. Animals were sacrificed at 7 dpi and intestinal tissue dissected for analysis.

Immunocytochemistry

Cuprolinic blue staining for visualization of enteric neurons was performed as previously described (Holst and Powley, 1995). The method was utilized for quantification of enteric neurons in T. gondii and T. cruzi infection experiments (Figure S1).

Whole-mount intestine immunofluorescence

Briefly, mice were sacrificed by cervical dislocation and the small intestine was removed and placed in HBSS Mg²⁺Ca²⁺(Gibco) + 5% FCS. The intestine was cut open longitudinally and the luminal contents washed away in DPBS. The muscularis was then carefully dissected away from the underlying mucosa in one intact sheet. The tissue was pinned down in a plate coated with Sylgard and then fixed for O/N with 4 % PFA with gentle agitation. After washing in DPBS whole mount samples were then permeabilized first in 0.5 % Triton X-100/0.05 % Tween-20/4 μg heparin (PTxwH) for 2 hours at room temperature (RT) with gentle shaking. Samples were then blocked for 2 hours in blocking buffer (PTxwH with 5% bovine serum albumin/5 % donkey/goat serum) for 2 hours at RT with gentle agitation. Antibodies were added to the blocking buffer at appropriate concentrations and incubated for 2 days at 4 °C. After primary incubation the tissue was washed 4 times in PTxwH and then incubated in blocking buffer with secondary antibody at concentrations within the primary antibody range for 2 hours at RT. Samples were again washed washed 4 times in PTxwH and then mounted with FluoroMount G on slides with 1 1/2 coverslips. Slides were kept in the dark at 4 °C until they were imaged.

Confocal imaging

Whole-mount intestine samples were imaged on an inverted LSM 880 NLO laser scanning confocal and multiphoton microscope (Zeiss) and on an inverted TCS SP8 laser scanning confocal microscope (Leica).

Neuronal quantification

A minimum of 10 images were randomly acquired across a piece of whole-mount muscularis. These images were then opened in ImageJ, and the cell counter was used to count the number of ANNA-1+ cells in a given field. This number was then multiplied by a factor of 2.95 (20x objective) or 3.125 (25x objective), to calculate the number of counted neurons per square millimeter (mm²). The average of 10 (or more) images was then calculated and plotted. Thus, every point on a given graph corresponds to a single animal. For determining iEAN subtype changes, the number of nNOS+, SST+, and tdTomato+ neurons were also counted. These numbers were then reported as both number per mm² and percent of ANNA-1+ neurons.

RiboTag

Heterozygous or homozygous Snap25^RiboTag were used for TRAP-seq analysis as no differences were found between either genotype. For intestine immunoprecipitation (IP) mice were sacrificed and tissue remove and divided as above. Samples were washed of fecal contents in PBS with cycloheximide (0.2 mg/mL) (PBS/CHX). Mesenteric fat was removed and the muscularis was separated from the mucosa as described above and samples were washed 5 times in PBS/CHX. For nodose ganglion IP, tissues were isolated as described above. For Snap25^{ΔNlrp6:RiboTag} animals, ileum muscularis was isolated as above and flash frozen in liquid nitrogen prior to immunoprecipitation. The RiboTag immunoprecipitation protocol (http://depts.washington.edu/mcklab/RiboTagIPprotocol2014.pdf) was then followed with the following modifications: All samples were homogenized by hand with a dounce homogenizer in 2.5 mL supplemented homogenization buffer (changes per 2.5 mL: 50 μl Protease Inhibitor, 75 μl heparin (100 mg/mL stock), 25 μl SUPERase In RNase Inhibitor). Samples were then centrifuged for 10 minutes at 10,000 G, after which 800 μl of supernatant was removed and 5 μL of anti-HA antibody (Abcam, ab9110) was added. For input samples, 40 μl of this supernatant was added to 60uL of PicoPure lysis buffer and stored at −80 °C. Samples were kept rotating at 4 °C with antibody for 1 hour. 200 μl of Thermo Protein magnetic A/G beads were washed with homogenization buffer, added to the sample, and kept rotating for 30 min at 4 °C. The beads were washed four times with high-salt buffer and samples were eluted with 100 μl of PicoPure lysis buffer. RNA was extracted using the Arcturus PicoPure RNA isolation kit (Applied Biosystems) according to the manufacturer’s instructions.

RNA-sequencing

RNA libraries were prepared using SMARTer Ultra Low Input RNA (ClonTech Labs) and sequenced using 75 base-pair single end reads on a NextSeq 500 instrument (Illumina). Reads were aligned using Kallisto (Bray et al., 2016) to Mouse Ensembl v91. Transcript abundance files were then used in the DESeq2 R package, which was used for all downstream differential expression analysis and generation of volcano plots. Differentially expressed genes between samples were compared with a cutoff of log₂ Fold Change > 1.

Mining of published iEAN scRNA-seq

The published single-cell analysis of the mouse nervous system at mousebrain.org/genesearch.html (PMID: 30096314) was searched for Nrlp6 and filtered for clusters that express the gene at a trinarization score >= 0.95. Relative expression was then plotted as a heatmap in Prism 8. Vglut2+ subsets were defined based on the same single-cell analysis.

Cryosectioning of fresh-frozen tissue

Mice were given a lethal dose of isoflurane and then perfused with 30 ml ice cold DPBS. The ileum and colon were removed and flushed of luminal contents. The tissue was then transferred to OCT, positioned, flash frozen, moved to −20 for 2 hours, and finally moved to −80 for long term storage. Prior to cryosectioning, OCT blocks were equilibrated in the cryostat for 20 minutes. Sections were cut at 15 μm, put onto SuperfrostPlus slides, and left to dry for 10–15 minutes. Slides were then transferred to dry ice and finally to −80°C before RNAscope® processing.

RNA fluorescence in situ hybridization (FISH) using RNAscope® technology

RNAscope® in situ hybridization was performed using probes against Nlrp6 and Elavl4 on 15 μm sections of fresh frozen ileum or colon tissue isolated from C57BL/6J mice according to the manufacturer’s instruction with the following modification, tissue was pretreated with Protease IV for 20 min at RT. Samples were then mounted in Fluoromount-G with DAPI and 1 1/2 coverslips were applied.

Macrophage intercalation calculation

Raw data as imported from the microscope was used for all cell identification and subsequent analyses. Some post –analysis pseudo-color adjustment was performed for individual images to account for differences in auto-fluorescence and labeling. Imaris (Bitplane AG) software was used for cell identification, using the “Surfaces” algorithm as described by the manufacturer. A surface was created based on anti-ANNA-1 (neuron) staining to define neuronal ganglia. A masked channel was then created to capture fluorescence in other channels within the neuronal ganglia volume. A second surface was created based on anti-MHCII (macrophage) staining only within the masked channel. Volume statistics on both neuron (total) and macrophage (neuronal ganglia masked) surfaces were then exported. Percent macrophage intercalation per image was calculated as:

$$MacrophageIntercalation=\frac{\left(\Sigma \right(Macrophagesurfacevolumes)+\Sigma (Neuronalgangliasurfacevolumes)}{\Sigma \left(Neuronalgangliasurfacevolumes\right)}\times 100$$

The average of at least 10 random macrophage intercalation percentages was used to determine the percentage for each individual animal.

3D image reconstruction for MM ganglionic intercalation calculations

Images were adjusted post hoc using Imaris x64 software (version 9.1 Bitplane) and 3D reconstructions were recorded as mp4 video files. Optical slices were taken using the orthoslicer or oblique slicer tools.

Determination of caspase 11 mutation status by PCR and Sanger Sequencing

DNA was extracted from ear pieces of mice using QuickExtract™ DNA Extraction Solution (Lucigen). Target sequences were amplified using the following primers: Forward 5’-AGGCATATCTATAATCCCTTCACTG-3’; Reverse 5’-GAATATATCAAAGAGATGACAAGAGC- 3’. The following PCR conditions were used: 4 min 94°C, 1 min 94 °C, 0.5 min 58°C, 1 min 72 °C, 7 min 72 °C, and end at 12 °C. Samples were loaded on 3% agarose gels, and bands from all strains were compared to samples from C57BL/6J and 129S mice. The 5 bp deletion in Casp11 described in the 129S1 strain fragment runs slightly higher than the wild-type fragment at a height of approximately 220 bp (Vanden Berghe et al., 2015). Casp11 mutation status was further confirmed in C57BL/6J, CBA/J and 129S1 mice by Sanger sequencing of a gel-purified PCR product using primers 5′-CAGTATTATTATTGGTGATGCAAATG-3′ and 5′-GGAATATATCAAAGAGATGACAAGAGC-3′. Casp11 mutation status was also confirmed in further relevant mouse strains utilized in this study.

Single Cell Suspension of Intestinal Macrophages

Mice were euthanized, and the small intestine was carefully removed, cleaned, cut open longitudinally and washed 2X in HBSS Mg₂₊Ca₂Gibco) and 1X in HBSS Mg₂₊Ca₂ with 1 mM DTT (Sigma- Aldrich). The tissue was cut in two and the muscularis region was carefully dissected from the underlying mucosa. Muscularis tissue was then finely cut and digested in HBSS Mg₂₊Ca₂₊+ 5% FBS + 1x NaPyr + 25mM HEPES + 50 μg/ml DNase I (Roche) + 400 U/ml Collagenase D (Roche) + 2.5U/ml Dispase (Corning) at 37°C. The muscularis was digested for 40 min. The tissue was then homogenized with an 18-gauge needle, filtered through a 70 μm cell strainer and washed with HBSS Mg₂₊Ca₂. Cells were incubated with Fc block and antibodies against the indicated cell surface markers in FACS buffer (PBS, 1% BSA, 10 mM EDTA, 0.02% NaN₃).

Quantitative PCR

Total RNA was isolated using TRIzol™ (Invitrogen), from which cDNA libraries were reverse transcribed using Superscript II (Invitrogen) and random primers following the instructions provided by the manufacturer. Quantitative PCR (qPCR) was performed using SYBR green (Bio-Rad Laboratories). Data were collected and analyzed on a QuantStudio 3 (Thermo Scientific). The Rpl32 housekeeping gene was used to normalize samples. For TRAP-qPCR, RiboTag purified mRNA was reverse transcribed using iScript cDNA Synthesis Kit (BioRad) and qPCR was performed using SYBR green (Bio-Rad Laboratories). The following primers were used: Rpl32 forward 5’-ACAATGTCAAGGAGCTGGAG-3’, Rpl32 reverse 5’- TTGGGATTGGTGACTCTGATG-3’, Arg1 forward 5’-CTCCAAGCCAAAGTCCTTAGAG- 3’, Arg1 reverse 5’-AGGAGCTGTCATTAGGGACATC-3’, Ym1 forward 5’-AGACTTGCGTGACTATGAAGCATT-3’, Ym1 reverse 5’-GCAGGTCCAAACTTCCATCCTC- 3’, Nlrp6 exon 4 forward 5’-CAGACGCTGTGGACCTTGT-3’, Nlrp6 exon 4 reverse 5’- ACGTGCTCGCGGTACTTCTT- 3’, Nlrp6 exon 1 forward 5’-TTGACTGTCAGCAAGAGTCC-3’, Nlrp6 exon 1 reverse 5’-GGTGATCCTTTCTGGGCTAAA-3’, Elavl4 forward 5’-GATCAGGGATGCTAACCTGTATG-3’, Elavl4 reverse 5’-GGTGATGATGCGACCGTATT-3’. To calculate a fold change, the final qPCR cycle of 45 was chosen in cases in which there was no amplification. This was comparable to water, which was used as a control and did not amplify.

AAV9 Transduction of iEAN

Casp11^flox/flox mice were retro-orbitally injected with ~5 × 10¹¹ GC of AAV9-hSyn-HI-eGFP-Cre-WPRE-SV40 (105540-AAV9) or AAV9-hSyn-eGFP-WPRE-bGH (Addgene #105539-AAV9) diluted in 100 μl of sterile DPBS. Mice were then infected at least 2 weeks post-viral injection.

Induction of glia depletion

Plp1^CreERT;Rosa26^DTA ( Plp1^iDTA, Cre+) and Rosa26^DTA mice (Cre-) received a single oral gavage of 8 mg tamoxifen (Sigma) solubilized in 250 μl sterile corn oil (Sigma) at 6–7 weeks of age. Mice were infected with Salmonella spiB 5 days post-tamoxifen administration and sacrificed at day 6 post-infection (day 11 post-tamoxifen).

Statistical analysis

Significance levels indicated are as follows: * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001. All data are presented as mean ± SD or mean ± SEM. At least two independent experiments were performed throughout in this study. All statistical tests used were two-tailed. The experiments were not randomized and no statistical methods were used to predetermine sample size. Multivariate data were analyzed by one-way ANOVA and Tukey’s multiple comparisons post hoc test. Comparisons between two conditions were analyzed by unpaired Student’s t-test. GraphPad PRISM version 8.0 and R 3.4.3 were used for generation of graphs and statistics.

Data and Software Availability

Data generated by RNA sequencing are deposited in the NCBI Gene Expression Omnibus (GEO) database and are accessible under GSE140309. All software used is available online, either freely or from a commercial supplier and is summarized in the Key Resources table. Algorithms described above are purely mathematical calculations and can be performed using any software. No new software was written for this project.

KEY RESOURCES TABLE.

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Human polyclonal anti-Hu (ANNA-1)	Gift of V. Lennon	N/A
Rat monoclonal anti-MHC class II	Millipore Sigma	Cat#: MABF33;
Rat monoclonal anti-Somatostatin	Merck	Cat#: MAB354; RRID: AB_2255365
Rabbit monoclonal anti-nNOS	Abcam	Cat#: ab76067, RRID: AB_2152469
Rabbit polycolonal anti-ASC	Adipogen	Cat#: AG-25B-0006; RRID: AB_2490440
Rabbit monoclonal anti-HA	Cell Signaling Technology	Cat#: 3724; RRID: AB_1549585
Rabbit monoclonal S100 beta		Cat#: ab52642
Rat monoclonal anti-CSF1R (AFS98)	Gift of M. Merad (MSSM)	AFS98
Anti-CD45.2	BD-Pharmingen	Cat#: 560693
Anti-CD45R	BD-Pharmingen	Cat#: 553088
Anti-CD103	eBioscience	Cat#: 11103185
Anti-CD11b	eBioscience	Cat#: 47011282
Anti-CD11c	Affymetrix	Cat#: 48011482
Anti-F4/80	Affymetrix	Cat#: 12480180
Anti-Siglec F	BD-Pharmingen	Cat#: 562680
Anti-CD3e	Thermo Fisher Scientific	Cat#: 45003180
Anti-Ly6G	BD-Pharmingen	Cat#: 560601
Anti-CD64	Biolegend	Cat#: 139309
Anti-MHC II	Thermo Fisher Scientific	Cat#: 17532380
Aqua fixable dead cell stain	Invitrogen	Cat#: L34966
Bacterial and Protozoan Strains
S. enterica serovar Typhimurium (SL1344)	UC Davis, Andreas Bäumler	Strain# SL1344; NCBI Taxonomy ID: 216597
S. enterica serovar Typhimurium (SL1344) ΔspiB	UC Davis, Andreas Bäumler	N/A
Yersinia pseudotuberculosis	NIH, Yasmine Belkaid
Toxoplasma gondii (me49)	UMass, Ricardo Gazzinelli	me49
Trypanosoma cruzi (Y strain)	UMass, Ricardo Gazzinelli	N/A
Strongyloides venezuelensis	Stanford, S. Galli	N/A
Biological Samples
N/A	N/A	N/A
Chemicals, Peptides, and Recombinant Proteins
Compound 21	Hello Bio	Cat#: HB6124
Tamoxifen	Sigma	Cat#: T5648
Ampicillin sodium salt	Sigma	Cat#: A0166
Neomycin trisulfate salt hydrate	Sigma	Cat#: N6386
Vancomycin hydrochloride	Sigma	Cat#: V2002
Metronidazole	Sigma	Cat#: M3761
Salbutamol sulfate	Selleck Chemicals	Cat#: S2507
Methylcellulose	Sigma-Aldrich	Cat#: 274429
Carmine Red	Sigma-Aldrich	Cat#: C1002
Spermine	Sigma-Aldrich	Cat#:85590
DFMO	Gift of P. Woster	N/A
zVAD-FMK	Selleck Chemicals	Cat#: S7023
TRIzol™ Reagent	Thermo Fisher Scientific	Cat#: 15596026
Streptomycin sulfate	Sigma-Aldrich	Cat#: S9137
HBSS, calcium, magnesium	GIBCO	Cat# 24020117
HBSS, no calcium, no magnesium	GIBCO	Cat# 14170112
Dulbecco’s Phosphate Buffered Saline	Corning	Cat# 21040CV
EDTA (0.5 M), pH 8.0	Ambion	Cat# AM9260G
RPMI 1640 Medium	GIBCO	Cat# 11875119
Collagenase VIII	Sigma-Aldrich	Cat# C2139
DNAse I	Roche	Cat# 04536282001
Isoflurane (IsoThesia)	Henry Schein	Cat# 029405
DL-Dithiothreitol (DTT)	Sigma-Aldrich	Cat# D0632
Triton X-100	Alfa Aesar	Cat# A16046
Dimethyl sulfoxide (DMSO)	Sigma-Aldrich	Cat# D8418
Tween 20	Promega	Cat# H5151
Heparin	Sigma-Aldrich	Cat# H3149
Cuprolinic Blue	Polysciences	Cat#: 17052-100
DNA extraction buffer	Lucigen	Cat#: QE09050
RNAscope ® probe Nlrp6	ACDbio	Cat#: 404561
RNAscope ® probe Elavl4	ACDBio	Cat#: 479581-C2
PFHM-II	Thermo Fisher Scientific	Cat#: 12040077
Dialysis Tubing	Spectrum Labs	Cat#: 132676
Penicillin/streptomycin 100x	Gibco	Cat#: 15140-122
Bioreactor Flask	Wheation	Cat#: WCL-1000-3
1M Tris-HCL pH 8	Invitrogen	Cat#: 15568-025
IgG Binding Buffer	Thermo Fisher Scientific	Cat#: 21011
IgG Elution Buffer	Thermo Fisher Scientific	Cat#: 21004
Protein G Sepharose	GE Healthcare	Cat#: 45-000-116
Gravity Column	Bio-Rad	Cat#: 7371512
30K - Centrifugal Filters	Amicon	Cat#: UFC903024
IgG from Rat Serum	Sigma-Aldrich	Cat#: I4131
Critical Commercial Assays
Mouse Lipocalin-2/NGAL DuoSet ELISA	R&D Systems	Cat#: DY1857
PicoPure RNA Isolation Kit	Thermo Fisher Scientific	Cat#: KIT0204
RNAscope® Kit	ACDBio	Cat#: 320850
Deposited Data
Data generated by RNA sequencing are deposited in the NCBI Gene Expression Omnibus (GEO) database.	NCBI	GSE140309
Experimental Models: Organisms/Strains
Mouse: C57BL6/J	Jackson Laboratory	#000664
Mouse: Lyz2Cre	Jackson Laboratory	#004781
Mouse: Rosa26tdTomato	Jackson Laboratory	#007914
Mouse: VGLUT2Cre	Jackson Laboratory	#016963
Mouse: 129S1/SvImJ	Jackson Laboratory	#002448
Mouse: Ccr2−/−	Jackson Laboratory	#004999
Mouse: Casp1−/− Casp11−/−	Jackson Laboratory	#016621
Mouse: CBA/J	Jackson Laboratory	#000656
Mouse: Rpl22HA	Jackson Laboratory	#011029
Mouse: Snap25Cre	Jackson Laboratory	#023525
Mouse: Adrb2flox	G. Karsenty	N/A
Mouse: Arg1flox	Jackson Laboratory	#008817
Mouse: Cx3cr1GFP
Mouse: NestinGFP	P. Frenette, G. Enikolopov	N/A
Mouse: Casp11−/−	Jackson Laboratory	#024698
Mouse: R26-CAG-ASC-citrine	Jackson Laboratory	#030744
Mouse: NSG	Jackson Laboratory	#005557
Mouse: Phox2bcre	Jackson Laboratory	# 016223
#Mouse: Plp1CreERT	Jackson Laboratory	# 005975
Mouse: Rosa26DTA	Jackson Laboratory	# 009669
Mouse: R26-hM4Di/mCitrine	Jackson Laboratory	# 026219
Mouse: Sox10CreERT2	B. Gulbransen, V. Pachnis	N/A
Mouse: SNScre	R. Kuhner	N/A
Mouse: Nlrp6flox	P. Rosenstiel	N/A
Mouse: Casp11flox	KOMP	N/A
Oligonucleotides
Casp11 forward 5’-AGGCATATCTATAATCCCTTCACTG-3’	IDT	N/A
Casp11 reverse 5’-GAATATATCAAAGAGATGACAAGAGC- 3’	IDT	N/A
Arg1 forward 5’-CTCCAAGCCAAAGTCCTTAGAG-3’	IDT	N/A
Arg1 reverse 5’-AGGAGCTGTCATTAGGGACATC-3’	IDT	N/A
Ym1 forward 5’-AGACTTGCGTGACTATGAAGCATT-3’	IDT	N/A
Ym1 reverse 5’-GCAGGTCCAAACTTCCATCCTC-3’	IDT	N/A
Rpl32 forward 5’-ACAATGTCAAGGAGCTGGAG-3’	IDT	N/A
Rpl32 reverse 5’-TTGGGATTGGTGACTCTGATG-3’	IDT	N/A
Nlrp6 E1 forward 5’-TTGACTGTCAGCAAGAGTCC-3’	IDT	N/A
Nlrp6 E1 reverse 5’-GGTGATCCTTTCTGGGCTAAA-3’	IDT	N/A
Nlrp6 E4 forward 5’-CAGACGCTGTGGACCTTGT-3’	IDT	N/A
Nlrp6 E4 reverse 5’- ACGTGCTCGCGGTACTTCTT-3’	IDT	N/A
Elavl4 forward 5’-GAT CAGGGATGCTAACCTGTATG-3’	IDT	N/A
Elavl4 reverse 5’- GGTGATGATGCGACCGTATT -3’	IDT	N/A
Recombinant DNA
AAV9-hSyn-eGFP-WPRE-bGH	Addgene	#105539-AAV9
AAV9-hSyn-HI-eGFP-Cre-WPRE-SV40	Addgene	#105540-AAV9
Software and Algorithms
GraphPad Prism version 8 for MacOS	Graphpad Software Inc.	https://www.graphpad.com
RStudio	RStudio®	https://www.rstudio.com
DEseq2	Bioconstructor	https://bioconductor.org/packages/release/bioc/html/DESeq2.html
Imaris (v. 8.4)	Bitplane
Fiji	ImageJ	https://imagej.net/Welcome
Microsoft Excel for MacOS	Microsoft	https://products.office.com/en-us/excel
Other
Salmonella Shigella Agar (SS Agar)	BD	Cat# 211597
Luria’s Broth base (LB)	Thermo-Fisher Scientific	Cat# 12795027
O.C.T. Compound (Tissue-Plus®)	Fisher Healthcare	Cat# 23-730-571

Supplementary Material

FigS1

Figure S1, iEAN Loss is Pathogen-Specific, does not Impact Potential Precursors, and Correlates with Dysbiosis, Related to Figure 1.

(A) Neuronal quantification as assessed by immunofluorescence staining (ANNA-1) in the myenteric plexus of ileal segments from C57BL/6J mice on the indicated days post-exposure to 10⁹ CFU of heat-killed Salmonella spiB. (B) Neuronal quantification (ANNA-1 staining) in the myenteric plexus of ileal segments from C57BL/6J mice orally infected with 10⁸ CFU of Y. pseudotuberculosis. (C-F) C57BL/6J mice were orally infected with 5 cysts of Toxoplasma gondii. (C) Quantification of fecal Lcn-2 levels as assessed by ELISA on the depicted days; (D) Ileal ring myography assessed on day 18 post-infection; (E) Total gastrointestinal transit time as assessed on the indicated days post infection; experiments were ended at 450 min (dashed line); (F) Neuronal quantification as assessed by cuprolinic blue staining in the ileum myenteric plexus on the indicated days post infection. NI = non-infected. (G) Neuronal quantification as assessed by cuprolinic blue staining in the ileum myenteric plexus on the indicated days post-infection of C57BL/6J mice with 10⁴ parasites of T. cruzi. (H) Neuronal quantification (ANNA-1 staining) in the myenteric plexus of (left panel) duodenal and (right panel) ileal segments C57BL/6J mice subcutaneously injected with 700 L3 larvae of S. venezuelensis (S. ven.), analyzed on day 8 post-infection. (I) Representative confocal immunofluorescence images of ileum muscularis from Nestin^GFP reporter mice 7 days post-oral exposure to PBS or spiB. Scale bar = 50 μm. Images representative of at least n=5 animals per group. (J) Neuronal quantification and glial reporter overlap as assessed by immunofluorescence staining (ANNA-1, HA) in the ileal myenteric plexus from tamoxifen-treated Sox10^RiboTag reporter mice 126 days post oral-gavage with PBS or spiB. (K) Shannon diversity index of 16S sequencing from feces obtained from n=2 mice in representative groups in Figure 1L. (L) Top 11 most abundant genera from 16S sequencing of feces obtained from mice in representative groups in Figure 1L. Unless otherwise indicated, data are representative of at least 3 mice per condition. At least 2 independent experiments were performed. Data were analyzed by unpaired Student’s t-test or ANOVA with Tukey’s posthoc test and are shown as mean ± SD; *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001.

FigS2

Figure S2, Immunofluorescent Characterization of Inhibitory and Interneuron iEAN Populations post-spiB Infection, Related to Figure 2.

(A) Representative confocal immunofluorescence image of ileum myenteric plexus from C57BL/6J stained with anti-ANNA-1 (red) and anti-neuronal nitric oxide synthase (nNOS) (green) antibodies. (B) Representative confocal immunofluorescence image of ileum myenteric plexus from C57BL/6J stained with anti-ANNA-1 (blue) and anti-somatostatin (SST) (yellow) antibodies. Scale bars = 50 μm. Images representative of n= 5 mice.

FigS4

Figure S4, Targeting Validation of Antibody-Mediated MM Depletion and CCR2^−/− spiB Infection Characteristics, Related to Figure 4.

(A) Representative confocal immunofluorescence image of the ileum submucosal layer of Cx3cr1^GFP mice treated with anti-CSF1R or IgG isotype control antibody by intraperitoneal injections. Scale bar = 50 μm. Images representative of at least n=3 mice per condition. (B) Representative confocal immunofluorescence image of the ileum myenteric plexus of C57BL/6J mice on day 7 post-treatment with anti-CSF1R or isotype control. Scale bar = 50 μm. Images representative of at least n=5 mice per condition. (C) Quantification of MM intercalation into ileum myenteric plexus ganglia as assessed by confocal immunofluorescence imaging and calculations based on Imaris software surface functions. Samples obtained at 3 hours post-oral gavage of PBS or spiB; (D-F) CCR2^−/− mice were orally infected with Salmonella spiB or gavaged with PBS (D) Quantification of fecal CFU on the indicated days post infection; (E) Gastrointestinal motility activity as assessed by total intestinal transit time on the indicated days post-infection; experiments were ended at 450 min (dashed line); (F) Quantification of neurons of the ileum myenteric plexus as assessed by immunofluorescence staining (ANNA-1). Data are representative of at least n=4 per condition. At least 2 independent experiments were performed. Data were analyzed by unpaired Student’s t-test and are shown as mean ± SD (C) or mean ± SEM (D-F); *p ≤ 0.05, ***p ≤ 0.001.

FigS3

Figure S3, Characterization of RiboTag and Viral or Genetic Inflammasome Targeting, Related to Figure 3.

(A) Representative confocal immunofluorescence images of the nodose ganglia from Snap25^RiboTag mice stained with anti-ANNA-1 (red) and anti-HA (green) antibodies. Scale bars = 50 μm. Images representative of n= 5 mice. (B) Immunofluorescence staining of the ileum myenteric plexus from C57BL/6J mice using anti-ANNA-1 (red) and anti-PYCARD (ASC, green) antibodies. Samples obtained 6 hours post-oral gavage of PBS or 10⁹ CFU of Salmonella spiB. Arrows indicate ASC “specks”; inset shows ImageJ zoom of the indicated panel in PBS-treated mice. Scale bars = 25 μm. Images representative of at least n=5 per group. (C) Immunofluorescence staining of ileum myenteric plexus from ASC^mCitrine mice using anti-ANNA-1 (red) antibody and native mCitrine fluorescence. Samples obtained 5 days post-oral gavage of PBS or 10⁹ CFU of Salmonella spiB. Arrows indicate ASC “specks” within neurons. Scale bars = 50 μm. Images representative of at least n=3 per group. (D) Heatmap representation of the relative expression of Nlrp6 in iEAN subsets as defined by single-cell analysis from the published dataset mousebrain.org/genesearch.html (Zeisel et al., 2018). Vglut2+ subsets as defined by the same analysis are highlighted. (E) Neuronal quantification (ANNA-1 staining) in the myenteric plexus of ileal segments on day 7 post-spiB infection of C57BL/6J mice receiving vehicle (dimethyl sulfoxide, DMSO) or zVAD-FMK (150 μg/day) delivered i.p. over the course of infection. Shaded area bounded by dashed lines indicates mean day 7 iEAN numbers +/- SEM of all control C57BL6/J mice analyzed in Figure 1F. (F-I) Analysis of Casp11 mutations in mouse strains used. (F) 3% agarose gel for Casp11 5bp deletion PCR product. Mouse strains used in thi work are indicated below each lane. Red dashed line indicates the size at which a potential 5bp deletion PCR product will run; (G) Sanger sequencing of Casp11 5bp deletion PCR product in C57BL/6J, CBA/J, and 129S1/SvImJ mice; (H) Sequence surrounding Casp11 5bp deletion region in C57BL/6J, CBA/J, and 129S1/SvImJ mice. Red line indicates 5bp deletion in 129S1/SvImJ mice; (I) Comparison of the length of intron sequences obtained from Casp11 5bp deletion PCR product in C57BL/6J, CBA/J, and 129S1/SvImJ mice.(J, K) Neuronal quantification (ANNA-1 staining) in the myenteric plexus of colonic segments on day 7 post-spiB infection of (J) 129S1/SvImJ mice and (K) CBA/J mice. (L) Representative confocal immunofluorescence image of the colon muscularis of Rosa26^lsl-tdTomato mice given i.v. injection of AAV9-hSyn-Cre-GFP virus. Image representative of at least n=3. Scale bar = 50 μm. (M) Log transformed relative expression to Rpl32 of TRAP-qPCR data of Elavl4 and Nlrp6 from Snap25^{ΔNlrp6:RiboTag} as compared to Snap25^{Nlr6flox/+:RiboTag}. Data are representative of 3–5 mice per condition. Data were analyzed by unpaired Student’s t-test and are shown as mean ± SD; *p ≤ 0.05, ***p ≤ 0.001.

FigS6

Figure S6, Stress or Anesthetic-mediated Activation of the Sympathetic Nervous System Can Rescue iEAN Loss, Related to Figure 6.

(A, B) Neuronal quantification as assessed by IF staining (ANNA-1) in the myenteric plexus of ileal segments on day 7 post-infection of C57BL/6 with spiB (A) receiving mock i.p. injections of PBS or IgG isotype control antibody or (B) following isoflurane anesthesia for 15 min 1 hour prior to infection. Shaded area bounded by dashed lines indicates mean day 7 iEAN numbers +/- SEM of all control C57BL6/J mice analyzed in Figure 1F; Data are representative of at least 3 mice per condition. At least 2 independent experiments were performed. Data were analyzed by unpaired Student’s t-test and are shown as mean ± SD; n.s. – not significant, ****p≤ 0.0001.

FigS7

Figure S7, Conditional Targeting of Arg1 in MM Does Not Alter Steady State Parameters, but Leads to Deficits post-Infection, Related to Figure 7.

(A) Neuronal quantification as assessed by immunofluorescence staining (ANNA-1) in the ileum myenteric plexus of naïve LysM^ΔArg1 and wild-type (WT) littermate control mice 7 days post–streptomycin treatment (no infection). (B) Gastrointestinal motility activity as assessed by total gastrointestinal transit time at the indicated time points of LysM^ΔArg1 and wild-type (WT) littermate control mice orally infected with 10⁹ CFU of spiB. Experiments were ended at 450 min (dashed line). Data are representative of at least 3 mice per condition. At least 2 independent experiments were performed. Data were analyzed by unpaired Student’s t-test and are shown as mean ± SD.

Suppl. Information

Supplementary Information, Related to Figures 2–7.

(A-L) Quantification of fecal Salmonella spiB CFU on the indicated days post-infection of the experiment displayed in (A) Figure 2A; (B) Figure 3D; (C) Figure 3E; (D) 3I–J; (E) Figure 3K; (F) Figure 4B; (G) Figure 4C; (H) Figure 4F; (I) Figure 4G; (J) 5F; (K) Figure 6B; (L) Figure 6D; (M) Figure 6E; (N) Figure 7B; (O) 7C; (P) Figure 7D; (Q) Figure S3K; (R) Figure S3L; (S) Figure S6A; (T) Figure S6B. Data are represented as mean (line), and individual values, with each point on a graph representing the value obtained from an individual animal.

FigS5

Figure S5, Conditional Targeting of Adrb2 in MM Does Not Alter Steady State Parameters, but Leads to Deficits Post-Infection, Related to Figure 5.

(A) Whole-mount IF images of the ileum myenteric plexus, nodose ganglion (NG), thoracic dorsal root ganglion (DRG), and celiac-superior mesenteric ganglion (CG-SMG) from LysM^{RiboTag:tdTomato} mice in the ileum stained for MHC II (green) and HA (blue),. Images representative of n=4. Scale bars = 50 μm. (B) Gastrointestinal motility as assessed by total intestinal transit time of naïve LysM^ΔAdrb2 and wild-type (WT) littermate control mice; experiments were ended at 450 min (dashed line); (C) Neuronal quantification as assessed by immunofluorescence staining (ANNA-1) in the ileum myenteric plexus of naïve LysM^ΔAdrb2 and wild-type (WT) littermate control mice; (D) Ileal ring myography assessed on day 53 post-infection of LysM^ΔAdrb2 and wild-type (WT) littermate control mice; Data are representative of at least 3 mice per condition. At least 2 independent experiments were performed. Data were analyzed by unpaired Student’s t-test and are shown as mean ± SD; n.s. – not significant, *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001.