Correspondence: David R. Brown, Department of Veterinary & Biomedical Sciences, University of Minnesota, 1988 Fitch Avenue, St Paul, MN 55108-6010, USA. Tel.: +1 612 624 0713; fax: +1 612 625 0204; e-mail: firstname.lastname@example.org
Peyer's patches of the small intestine serve as inductive sites for mucosal immunity as well as targets for invasive enteropathogens, including Salmonella. Because they are innervated by catecholamine-containing enteric nerves, the hypothesis that the endogenous catecholamines dopamine and norepinephrine or sympathomimetic drugs alter Salmonella Typhimurium uptake into Peyer's patches was tested. Porcine jejunal Peyer's patch explants were mounted in Ussing chambers and inoculated with a porcine field isolate of Salmonella Typhimurium DT104. Salmonella recovery from gentamicin-treated tissues increased significantly between 30 and 90 min of bacterial exposure to the mucosal surface. Addition of the neuronal conduction blocker saxitoxin (0.1 μmol L−1) or dopamine (30 μmol L−1) to the contraluminal aspect of explants decreased bacterial recovery after 60 min of Salmonella exposure. The effects of dopamine were mimicked by cocaine and methamphetamine (30 μmol L−1), which act on catecholaminergic nerve terminals to increase synaptic neurotransmitter concentrations. These results suggest that enteric catecholaminergic nerves modulate Salmonella colonization of Peyer's patches at the earliest stages of infection, in part by altering epithelial uptake of bacteria.
Peyer's patches in the small intestine represent a major inductive site for mucosal immunity and a target for enteroinvasive pathogens, including Salmonella (Jepson & Clark, 2001), Yersinia (Autenrieth & Firsching, 1996) and Shigella (Sansonetti et al., 1996). The specialized epithelium covering these discrete lymphoid follicles contains cuboidal epithelial cells referred to as microfold or ‘M’ cells. These cells, unlike columnar enterocytes, lack an organized brush-border, a thick glycocalyx or an abundance of hydrolytic enzymes (Owen, 1999). M cells, together with columnar epithelial cells, constitute the follicle-associated epithelium. M cells endocytose and process particulate matter including luminal antigens and other macromolecules for subsequent exposure to underlying lymphocytes and antigen-presenting cells located in the ‘intracellular pocket’ of the M cell (Didierlaurent et al., 2002). The follicle-associated epithelium appears to have bacterial and antigen uptake properties that differ from the nonfollicular intestinal epithelium (Keita et al., 2006).
Nontyphoidal Salmonella enterica infections are a major cause of food-borne gastroenteritis, and strains of S. enterica manifesting resistance to multiple classes of antibiotics have emerged over the past several years (Butaye et al., 2006). The worldwide spread of the multidrug-resistant DT104 strain of S. enterica serovar Typhimurium in humans as well as in swine and other food-producing animals has been of particular public health concern (Helms et al., 2005; Gebreyes et al., 2006). Accumulating evidence indicates that Peyer's patches in swine and other mammalian species are innervated by sympathetic, cholinergic and peptidergic nerves (Kulkarni-Narla et al., 1999; Defaweux et al., 2005, 2007; Ma et al., 2007; Vulchanova et al., 2007). Stimulation of peripheral sympathetic nerves results in the release of norepinephrine, which alters innate immune responses to infection (Maestroni, 2006). Furthermore, chemical sympathectomy has been shown to increase the resistance of mice to infection by the food-borne pathogen Listeria monocytogenes (Miura et al., 2001; Cao et al., 2002). Although the role of enteric nerves in Peyer's patch function has not been precisely defined, it has been reported that internalization of a vaccine strain of Salmonella choleraesuis is increased in porcine jejunal Peyer's patches that were treated with either norepinephrine or the neural conduction blocker, saxitoxin (Green et al., 2003).
In the present study, the hypothesis that norepinephrine likewise alters the uptake of Salmonella Typhimurium DT104 in Peyer's patch explants from the porcine jejunum, a biomedical model of the human small intestine was tested (Kararli, 1995). Moreover, it was hypothesized that the effects of norepinephrine would be mimicked by the related catecholamine dopamine and the sympathomimetic drugs of abuse, cocaine and methamphetamine, which act to increase norepinephrine and dopamine concentrations at catecholaminergic synapses (Fleckenstein et al., 2000). The results of this investigation may have implications relating to enteric disease susceptibility and oral vaccine efficacy.
Materials and methods
Bacteria and growth conditions
A porcine field isolate of S. enterica ssp. enterica serovar Typhimurium var. Copenhagen (a strain resistant to ampicillin, chloramphenicol, streptomycin, sulfisoxazole and tetracycline, i.e. R-type ACSSuT; Minnesota Department of Health isolate number #E02-000392) was provided by Dr Jeffrey Bender (Center for Animal Health and Food Safety, University of Minnesota, St Paul, MN) and subsequently identified as definitive type 104 (DT104) by the Centers for Disease Control and Prevention (Atlanta, GA). Salmonella Typhimurium DT104 and S. choleraesuis strain SC-54 (Enterisol SC-54, Boehringer-Ingleheim Vetmedica Inc., St Joseph, MO; Roof & Doitchinoff, 1995) were stored in 4% (v/v) glycerol/phosphate-buffered saline (PBS) at −80 °C until the time of culture. They were grown overnight in Luria–Bertani (LB) medium at 37 °C in a humidified 5% CO2 atmosphere. Spectrophotometric evaluations indicated that overnight incubation was sufficient for all bacteria to reach the stationary growth phase. Three to four hours before the experiments, bacteria in the mid-logarithmic phase were obtained by inoculating 300 μL of stationary-phase bacteria into 30 mL LB broth. Inocula (100 μL) were added to the luminal-bathing medium; aliquots were reserved for quantification of the bacterial inoculum by spread-plating.
Cytochalasin D was purchased from EMD Biosciences-Calbiochem (San Diego, CA). Morphine sulfate was provided by Dr Michael Murtaugh (Department of Veterinary & Biomedical Sciences, University of Minnesota) and other drugs were obtained from Sigma Chemical Co. (St Louis, MO). Cytochalasin D was solubilized in ethanol and all other drugs were dissolved in distilled water.
Animals and tissue isolation
Individual Peyer's patches were isolated from the distal jejunum of outbred Yorkshire–Landrace pigs of either sex that were 5–9 weeks old and weighed between 10 and 18 kg; the animals were obtained from a national specific pathogen-free accredited herd. Animals had continuous access to water and nonmedicated pig feed and were not fasted before sacrifice. They were anesthetized by an intramuscular injection of tiletamine hydrochloride-zolazepam (Telazol®; 8 mg kg−1; Fort Dodge Laboratories, Fort Dodge, IA) in combination with xylazine (3 mg kg−1; Lloyd Laboratories, Shenandoah, IA). They were subsequently euthanized by barbiturate overdose in accordance with approved University of Minnesota Institutional Animal Care and Use Committee protocols.
Sheets of follicle-associated epithelium were mounted between two lucite Ussing-type half chambers (2 cm2 flux area). Mucosal sheets were bathed on both luminal and contraluminal aspects in 10 mL of a buffered, physiological saline solution (composition in mmol L−1: 130 NaCl, 6 KCl, 3 CaCl2, 0.7 MgCl2, 20 NaHCO3, 0.29 NaH2PO4 and 1.3 Na2HPO4) that was continuously oxygenated on the serosal side with 95% O2–5% CO2 delivered by gas lift and maintained at pH 7.4 and 39 °C (porcine core temperature). d-Glucose and mannitol (10 mmol L−1) were added to the contraluminal and luminal bathing media, respectively.
Bacterial uptake and gentamicin resistance assay
Bacterial invasion into Peyer's patch explants was determined by a gentamicin resistance assay (Elsinghorst, 1994). Drugs were added to the contraluminal medium bathing Peyer's patch mucosa either at the same time (0 min) or 15 min before Salmonella Typhimurium DT104 addition to the luminal bathing fluid. The catecholamines (30 μmol L−1) were administered at time 0 as they may be subject to rapid removal processes in tissues. Saxitoxin (0.1 μmol L−1), cocaine and methamphetamine (30 μmol L−1) were added to the contraluminal bathing medium 15 min before Salmonella inoculation of the luminal medium. Cytochalasin D (10 μmol L−1) or its vehicle ethanol was added to the luminal bathing medium 15 min before inoculation of Salmonella. Mucosal sheets were removed from Ussing chambers after 30–90 min of luminal exposure to Salmonella Typhimurium DT104 (mean inoculum: 6.77 ± 0.06 log10 CFUs per 10 mL reservoir, n=28) or S. choleraesuis SC-54 (mean inoculum: 6.31 ± 0.141 log10 CFUs per 10 mL reservoir, n=21). Each tissue was weighed, washed three times in PBS (pH 7.4), and subsequently incubated at 37 °C in a humidified 5% CO2 atmosphere in a gentamicin solution (100 μg mL−1 in PBS; Sigma) for 80 min to eliminate extracellular bacteria. All tissues were subsequently homogenized using a Brinkmann Polytron and spread-plated on differential and selective medium for Salmonella (Difco XLD agar; Becton, Dickinson Co., Sparks, MD).
The number of CFU per gram of intestinal tissue was determined. Because of a relatively high inter-animal variability in Salmonella recovery from mucosal explants, the effects of drugs in each experiment with mucosal explants were compared in a drug-treated tissue and an adjacent, untreated tissue serving as the control from the same animal. Fractional recovery was calculated as the number of salmonellae recovered relative to the total number of bacteria added to the luminal medium in each experiment. Data are expressed as means ± SEs. Statistical analyses of data were performed using the prism computer software program (Version 4.0; graphpad Software, San Diego, CA). Comparisons between a control mean and a single treatment mean were made by paired or unpaired t-tests. Comparisons of a control mean with multiple treatment means were made by one-way anova and Tukey's multiple comparisons test where appropriate. In all cases, the limit for statistical significance was set at P<0.05.
Internalization of S. enterica serovars Typhimurium DT104 and choleraesuis SC-54 in Peyer's patch explants
Salmonella Typhimurium DT104 internalized into Peyer's patch explants in a time-dependent manner (Fig. 1). The mean fractional recovery of internalized bacteria was significantly greater (P<0.01) after 60 and 90 min than after 30 min of luminal bacterial exposure. In comparison, Peyer's patch recovery of the SC-54 strain of S. choleraesuis was c. 10-fold less than that of DT104 at the 30- and 60-min time points (mean log10± SE of fractional recoveries for SC-54 at 30, 60 and 90 min of luminal exposure were, respectively, −4.65 ± 0.281, −3.40 ± 0.258 and −2.60 ± 0.225, n=4, 5 and 21 tissues, P<0.01 for 30- vs. 90 min time point, Tukey's test). As with Salmonella Typhimurium, there was a significant time-related increase in the number of S. choleraesuis organisms recovered from gentamicin-treated tissues (F=8.64, df=29; P=0.0013, one-way anova).
As reported previously for S. choleraesuis strain SC-54 (Green & Brown, 2006), the internalization of Salmonella Typhimurium DT104 into Peyer's patches was unaffected by the actin-disrupting agent, cytochalasin D, at a luminal concentration of 10 μmol L−1 (mean log10± SE fractional recoveries of DT104 after a 60-min exposure from ethanol- or cytochalasin-pretreated tissues were −3.40 ± 0.204 and −2.88 ± 0.296, respectively; P=0.10, paired t test, n=one tissue pair from each of five pigs).
Effects of saxitoxin and catecholamines on Salmonella Typhimurium DT104 internalization
In Peyer's patch explants pretreated for 15 min with saxitoxin (0.1 μmol L−1 in contraluminal medium), the recovery of Salmonella Typhimurium was significantly decreased relative to control tissues untreated with the neurotoxin (mean log10 fractional recoveries of DT104 after 60 min from control and saxitoxin-pretreated tissues incubated in gentamicin were −2.36 ± 0.117 and −2.68 ± 0.077, respectively; P=0.02, paired t-test, n=one tissue pair from each of seven pigs).
Dopamine, added to the contraluminal bathing medium at 30 μmol L−1 at the same time as DT104 addition to the luminal medium, decreased Salmonella Typhimurium recovery from Peyer's patch explants after 60 min of bacterial exposure (Table 1). In explants treated similarly with 30 μmol L−1 norepinephrine, Salmonella recovery was not significantly different from control tissues (Table 1).
Table 1. Effects of drugs on fractional recovery (log10 mean ± SEM) of internalized Salmonella Typhimurium DT104 in Peyer's patch mucosa after 60 min luminal exposure
Effects of sympathomimetic drugs and morphine on Salmonella Typhimurium DT104 internalization
At a contraluminal concentration of 30 μM, the sympathomimetic drugs cocaine and methamphetamine each decreased Salmonella recovery from Peyer's patches relative to respective, untreated tissues serving as paired controls. Drug actions were manifested in tissues pretreated 15 min before bacterial exposure (Fig. 2).
At the same contraluminal concentration, an unrelated drug, morphine, had no significant effect on Salmonella internalization in Peyer's patch explants (Table 1).
Intracellular Salmonella Typhimurium in the log phase of growth was recovered from Peyer's patch explants. As reported previously in IPEC J2 epithelial cells derived from the porcine small intestine (Brown & Price, 2007), the internalization of this organism after luminal exposure to jejunal Peyer's patch explants was relatively rapid and increased in a time-dependent fashion. At early stages after its oral inoculation in swine, Salmonella Typhimurium has been found previously to colonize both Peyer's patches and the nonfollicular intestinal epithelium (Meyerholz et al., 2002; Schauser et al., 2004). These reports are in general agreement with the present data obtained in vitro. The early (30, 60 min) internalization of S. choleraesuis SC-54 into Peyer's patch explants was 10-fold less than that of Salmonella Typhimurium. This appears to parallel the differences observed by TEM between these two Salmonella serovars at similarly early stages of invasion of the porcine ileal epithelium (Meyerholz & Stabel, 2003). The ability of effector proteins translocated by the Salmonella type III secretion system to induce actin-dependent changes in the host epithelial cytoskeleton and facilitate bacterial internalization is well documented (Guiney & Lesnick, 2005). Salmonella Typhimurium DT104 invasion in IPEC J2 cell monolayers (Brown & Price, 2007) and ileal mucosa explants from swine (Schreiber et al., 2007) is reduced, but not abolished by the actin inhibitor, cytochalasin D. In contrast, the uptake of S. enterica serovars choleraesuis (Green & Brown, 2006) and Typhimurium (present results) into swine Peyer's patches appears to be relatively insensitive to cytochalasin, suggesting that the cellular mechanisms mediating Salmonella entry into the follicular and nonfollicular absorptive epithelium may differ.
Saxitoxin is a neuronal Na+ channel blocker that inhibits axonal impulse conduction in most populations of enteric neurons at low (<1 μmol L−1) concentrations. At higher concentrations (>1 μmol L−1), it has been found to alter Escherichia coli chemotaxis (Tisa et al., 2000) and cation fluxes in cyanobacteria as well (Pomati et al., 2003). The recovery of Salmonella Typhimurium from Peyer's patches treated with saxitoxin was decreased relative to tissues untreated with the toxin. As the toxin was applied to the contraluminal aspect of the Peyer's patch explants, limiting its contact with bacteria in the luminal medium, its effect was likely due to blockade of neuronal conduction in Peyer's patch explants rather than to a direct action on Salmonella Typhimurium. In another study, saxitoxin similarly inhibited Salmonella Typhimurium internalization in explants of the porcine ileal mucosa (Schreiber et al., 2007).
Norepinephrine is released from sympathetic nerve fibers innervating the gut mucosa that express the norepinephrine synthetic enzyme dopamine β-hydroxylase and originate in the prevertebral ganglia (Furness, 2006). Dopamine, on the other hand, is expressed in a subpopulation of nonsympathetic enteric neurons residing in the intestinal wall (Li et al., 2004). Nerves expressing immunoreactivity for the catecholamine synthetic enzyme tyrosine hydroxylase alone or in combination with that of dopamine β-hydroxylase have been found to innervate porcine Peyer's patches (Kulkarni-Narla et al., 1999; Green et al., 2003). There is additional immunohistochemical evidence for the neural expression of the type 2 vesicular monoamine transporter (Kulkarni-Narla et al., 1999) and plasmalemmal norepinephrine transporter (Green et al., 2003). By altering the action of these transporters, cocaine and methamphetamine increase the concentrations of dopamine and norepinephrine at catecholaminergic synapses. In Peyer's patch explants, dopamine decreased Salmonella recovery as did both sympathomimetic drugs. Salmonella Typhimurium possesses catecholate ‘receptor’ proteins in its outer membrane that appear to function in norepinephrine-facilitated iron acquisition (Williams et al., 2006), and both dopamine and norepinephrine have been found to enhance S. enterica growth (Freestone et al., 2007). However, as in the case of saxitoxin, the effect of dopamine is probably mediated by interactions with host cells rather than with salmonellae. First, dopamine was administered into the contraluminal medium bathing Peyer's patch explants and was therefore not in direct contact with luminally -applied bacteria. Second, the sympathomimetic drugs that act via enteric nerves mimicked the effects of dopamine. Peyer's patch M cells may respond directly to dopamine, as native dopamine receptors are known to be expressed in intestinal epithelial cells (Vieira-Coelho & Soares-da-Silva, 2001; Baglole et al., 2005). However, other eukaryotic cell types could also be affected by norepinephrine or dopamine released from nerves in Peyer's patches. These include dendritic cells, which take up salmonellae and appear to express catecholaminergic receptors (Shreedhar et al., 2003; Maestroni, 2006; Guo et al., 2007). Cocaine has recently been found to increase the expression of DC-SIGN in dendritic cells through a dopamine receptor-dependent mechanism (Nair et al., 2005).
After oral Salmonella inoculation in rodents, opium or the major opiate alkaloid morphine have been shown to increase host mortality rate from Salmonella infection and augment the recovery of salmonellae from Peyer's patches (Kent et al., 1966; MacFarlane et al., 2000). In the present study, morphine treatment did not change the number of salmonellae recovered from Peyer's patch explants in the first hour of infection. Based on these results, it is concluded that it is unlikely that morphine increases Salmonella uptake into Peyer's patches, at least under short-term conditions in swine. It is more likely that the increased Peyer's patch colonization observed in vivo studies was attributable to an opioid-induced reduction in intestinal propulsion, which in turn would increase the contact time of salmonellae with the intestinal mucosa, including the follicle-associated epithelium (Feng et al., 2006). On the other hand, there is evidence for opioid receptor expression in absorptive epithelial cells of the intestine (Lang et al., 1996; Nano et al., 2000; Neudeck & Loeb, 2002), including those in the porcine small intestinal mucosa (Quito et al., 1991).
Saxitoxin and norepinephrine increased S. choleraesuis recovery from porcine Peyer's patch explants in a previous study by the authors' laboratory (Green et al., 2003), but both saxitoxin and dopamine decreased Salmonella Typhimurium DT104 recovery in the present investigation. Norepinephrine appeared to have no significant effect on Salmonella Typhimurium DT104 internalization into explants of either Peyer's patches or the porcine nonfollicular ileal mucosa (Schreiber et al., 2007). Although the cellular sites and mechanisms by which these drugs produce their effects on Salmonella internalization have yet to be defined, the differences observed might at least partly be explained by the different pathogenic profiles of these two Salmonella serovars. Salmonella choleraesuis is a pig-adapted pathogen that is capable of producing bacteremia, whereas Salmonella Typhimurium DT104 has a broader host range and predominantly produces a localized enteritis in immunocompetent hosts. Moreover, there are differences in the invasion and mucosal transcriptional responses of epithelial cells at the earliest stages of infection initiated by these two pathogens (Meyerholz & Stabel, 2003; Skjolaas et al., 2006, 2007; Uthe et al., 2007). Porcine jejunal Peyer's patches are highly innervated by substance P/calcitonin gene-related peptide-immunoreactive nerves, which are thought to subserve a sensory function (Vulchanova et al., 2007). After uptake across the M cells of PP domes, the two Salmonella serovars may differentially activate sensory nerves, which act via epithelial cells or immunocytes to rapidly modulate bacterial accumulation and control inflammation in this inductive site. The present results suggest that dopamine in Peyer's patches may be one neurotransmitter substance capable of modulating Salmonella colonization of the intestine at the earliest stages of infection. Further tests of the general hypothesis that Peyer's patch sampling function is regulated by enteric neurotransmitters or hormones may yield important transdisciplinary information pertinent to disease susceptibility, mucosal immunity and intestinal drug delivery.
This study was funded by National Institutes of Health grant RO1 DA-10200. The authors thank Dr Jeffrey Bender (Center for Animal Health and Food Safety, University of Minnesota, St Paul, MN) for kindly providing the field isolate of Salmonella Typhimurium DT104 used in this study.