Background
Invasive
Salmonellosis is a global burden with more than 100 million cases per year resulting in over 350,000 deaths [
1]–[
3].
Salmonella enterica infections, which are the Gram-negative intracellular bacteria responsible for this burden, can result in diverse clinical manifestations. Food-borne non-typhoidal
Salmonella (NTS), caused by serovar Typhimurium or Enteriditis, has recently emerged as a prominent cause of bloodstream infection, primarily in African adults and children, with an associated case fatality of up to 25%. HIV, malaria, and malnutrition being important risk factors [
2]. Typhoid or enteric fever, caused by the exclusively human serovar Typhi or Paratyphi A, is a bacteremic disease that can result into intestinal perforation, peritonitis, encephalopathy myocarditis and hemodynamic shock [
3]–[
5]. Antimicrobial resistance for all
Salmonella enterica infections is widespread [
6]. The variations in the clinical features of infection with these intracellular pathogens relate to differences in the interaction between different
Salmonella serovars and the host [
3]. A better understanding of host-pathogen interactions in invasive Salmonellosis could explain these diverse clinical manifestations and potentially lead to new therapeutic strategies in order to decrease the considerable morbidity and mortality.
Salmonella spp. are recognized by pattern recognition receptors (PRR), via conserved motifs termed ‘pathogen-associated-molecular-patterns’, resulting in activation of signaling pathways that initiate the inflammatory response [
3],[
7]–[
9]. The two most important families of PRRs are the membrane-bound Toll-like receptors (TLRs) and the cytosolic Nod-like receptors (NLRs). Pro-inflammatory cytokines such as interleukin (IL)-6, IL-1β (also called IL-1 F2), IL-18 and tumor necrosis factor (TNF)-α are released during early
Salmonella infection. By contrast, interferon (IFN)-γ secretion (triggered by IL-18) plays a central role in the control of persistent infection by affecting the extent of macrophage activation [
10]–[
16].
Salmonella spp. expresses multiple pathogen-associated-molecular-patterns, most notably type 3 secretion systems (T3SS), flagella, fimbriae, lipopolysaccharide (LPS) and bacterial DNA [
3],[
4]. TLR activation induces the synthesis of pro-IL-1β and pro-IL-18, upon which a second signal is provided by the activation of the intracellular inflammasome and caspase-1 leading to IL-1β and IL-18 processing [
8],[
17].
The role of the inflammasome in the recognition of
Salmonella spp. has been studied extensively using in vitro and in vivo models [
9]. Best studied NLRs in the recognition of
Salmonella are the pyrin domain containing-3 (NRLP3) and the CARD domain-containing protein-4 (NLRC4), which form inflammasome complexes consisting of caspase-1 and the adaptor protein apoptosis-associated speck-like protein containing a CARD (called PYCARD or ASC) [
18]. Both the inflammasome receptors activate caspase-1 in response to
S. Typhimurium infection together with endogenous signals, and recruit ASC and caspase-1 into a single cytoplasmatic focus, which subsequently serves as the site for pro-IL-1β processing [
18]. Knockout mice lacking functional
casp1,
casp1-casp11 double knockout mice, or mice deficient in the end product of inflammasome activation (viz., IL-1β and IL-18), do have higher bacterial loads and succumb earlier upon infection with
S. Typhimurium [
19],[
20]. It was therefore hypothesized that
Nlrp3
−/−
(and
Asc
−/−
mice especially), having lower levels of inflammasome-processed cytokines, would succumb earlier to infection with
S. Typhimurium. On contrary, mice deficient in NLRC4, NLRP3 or ASC were still able to clear
S. Typhimurium as efficiently as control mice [
18],[
21]. An explanation given for this discrepancy in literature was that intracellular
Salmonella can also induce a
Salmonella pathogenicity island (SPI)-1 independent form of lytic cell death via caspase-11 [
20]. However,
casp11 deficient mice have a phenotype comparable to wild-type (WT) mice, showing a role for caspase-11 in the absence of caspase-1-mediated immunity [
20]. Furthermore, pre-growth conditions could have also led to this unexpected outcome, where conditions that favor
S PI-1 expression (log-phase bacteria) show minimal roles for the inflammasome, while
S PI-2 growth conditions (stationary phase bacteria) show a more significant role for caspase-1 during
Salmonella infection [
9]. Additionally, NLRC4-dependent bacterial clearance is independent of NLRP3, ASC, IL-18, and IL-1R, which could perhaps be an explanation for the unexpected phenotype of ASC-deficient animals [
17]. By contrast, cytokine processing that correlates with the formation of this ASC/caspase-1 focus, ASC-independent inflammasomes containing un-processed but active caspase-1 is also able to initiate rapid cell death [
22]. Yet, the relative importance of inflammasome-mediated cytokine maturation and pyroptosis during
S. Typhimurium infection still remains to be investigated.
In this study, we examined the in vivo relevance of ASC and NLRP3 during Salmonella infection using two clinically relevant experimental models. All mice received S. Typhimurium orally and either developed a typhoid-like disease (typhoid model) or, when pretreated with streptomycin, developed a non-typhoid salmonellosis (colitis model). We focused on bacterial growth, systemic cytokine release and organ pathology caused by S. Typhimurium using log-phase bacteria.
Methods
Ethics statement
Experiments were carried out in accordance with the Dutch Experiment on Animals Act and approved by the Animal Care and Use Committee of the University of Amsterdam (Permit number: DIX102113, DIX102114).
Mice
Eight- to 12-week-old male
Asc
−/−
and
Nlrp3
−/−
mice, backcrossed 9 times to a C57BL/6 genetic background, were generated as described [
23] and bred in the animal facility of the Academic Medical Center (Amsterdam, the Netherlands). Pathogen-free C57BL/6 WT mice were purchased from Charles River Laboratories Inc. (Maastricht, the Netherlands). Age- and sex-matched animals were used in all experiments and were housed in rooms with a controlled temperature a 12-h light–dark cycle, and were acclimatized for 2 weeks before the experiments. All experimental procedures and animal handling were done during the light cycle.
Experimental infection and design
For preparation of the inoculum
Salmonella Typhimurium strain 14028 was used (ATCC, LGC Standards GmbH, Wesel, Germany). Stock bacteria were streaked from frozen aliquots into 50 ml Luria broth for overnight incubation at 37°C in a 5% CO
2 incubator. Thereafter, a 1 ml portion for 1:50 dilution was transferred to fresh Luria broth and grown for approximately 2 h to mid-logarithmic phase (OD 0.4). Bacteria were harvested by centrifugation at 1500 ×
g for 10 min, washed and resuspended in sterile isotonic Hank’s Buffered Salt Solution (HBSS) at a concentration of 10
7 cfu/ml
S. Typhimurium. Inoculum size was verified retrospectively by serial 10-fold dilutions on blood agar (BA). Two different
Salmonella infection models were used: mice were starved for 12 h (typhoid model) or pre-treated with streptomycin (Rotexmedica GmbH, Trittau, Germany) 7.5 mg/100 μl HBSS per os (colitis model), and inoculated orally with 10
6 cfu/ml
S. Typhimurium in 100 μL HBSS as described previously [
24]–[
26]. At 2 and 5 days after infection (typhoid model) and 4 days (colitis model), mice were anesthetized with Hypnorm (Janssen Pharmaceutica, Beerse, Belgium) and midazolam (Roche, Basel, Switzerland), and then sacrificed by bleeding from the inferior vena cava. Blood was drawn into heparinized tubes and organs (Mesenterial lymph nodes (MLN), liver, spleen) were harvested and homogenized at 4°C in four volumes of sterile saline using a tissue homogenizer (Biospec Products, Bartlesville, UK). Death was confirmed by cutting the diaphragm. Bacterial loads were determined by serial ten-fold dilutions on BA incubated at 37°C for 16 h. Homogenates were centrifuged at 1500 ×
g at 4°C for 10 min, and supernatants were stored at −20°C pending assay.
Assays
IL-18 (Medical and Biological Laboratories, Nagoya, Japan), IL-1β and TNF-α (R&D, Minneapolis, MN) concentrations were measured by enzyme-linked immunosorbent assays. IFN-γ, MCP-1, IL-6, IL-10 and IL-12 were measured by cytometric bead array (BD Biosciences, San Jose, CA) in accordance with the manufacturer’s recommendations. AST, ALT and LDH were measured in plasma with spectrophotometry (Roche Diagnostics).
Histology
Liver, spleens, MLN, terminal ileum and cecum were harvested after sacrifice, fixed in 10% formalin and embedded in paraffin for histology. Sections of 4 μm were stained with haematoxylin–eosin, and read by a pathologist who was blinded to the groups. To score liver inflammation and damage, the entire slide surface was analyzed with respect to the following parameters: area of liver with parenchymal inflammation, necrosis and/or abscess formation, portal inflammation and thrombus formation. Each parameter was graded on a scale of 0 to 4 (0: absent; 1: mild; 2: moderate; 3: severe; 4: very severe). Thrombi were scored as follows: 0: no thrombi; 1: 1–4 thrombi; 2: 5–9 thrombi; 3: 10–15 thrombi; 4: more than 15. The total liver inflammation score was expressed as the sum of the scores for each parameter, the maximum being 16. Spleen sections were scored for inflammation, necrosis/abscess formation, and thrombus formation using the scales given above. The maximum total spleen inflammation score was 12. In both mouse models the cecum, colon and terminal ileum were assessed for neutrophil infiltration, edema and disruption of the epithelium to determine whether there was a colitis or gut infection present as is expected in the colitis model (streptomycin pre-treatment group).
Statistical analysis
Values are expressed as mean with standard error of the mean unless stated otherwise. Differences between groups were analyzed by Mann–Whitney U test. These analyses were performed using GraphPad Prism version 5.0, GraphPad Software (San Diego, CA). Values of p < 0.05 were considered statistically significant.
Discussion
In the present study we aimed to characterize the in vivo relevance of the central inflammasome molecules ASC and NLRP3 in two different murine models of systemic
Salmonella infection. Although in vitro studies confirmed that the inflammasome is activated during
S. Typhimurium infection, we here show that ASC and NLRP3 individually are dispensable for the development of innate immunity against the pathogen. In the murine typhoid model, in which mice were not pretreated with streptomycin prior to oral inoculation with
S. Typhimurium, equal bacterial counts were seen in WT,
Asc
−/−
and
Nlrp3
−/−
mice after infection in all organs (MLNs, liver, spleen, blood) at all time points. In line with the bacterial loads, proinflammatory cytokine levels, markers for tissue damage and organ pathology did not differ between groups. These results were replicated in a colitis model of
S. Typhimurium infection, in which mice were pretreated with streptomycin prior to oral inoculation, were we saw an equally limited role for both ASC and NLRP3 in systemic host defense against
S. Typhimurium. Taken together, the present results reveal a surprisingly limited role for ASC and NLRP3 during in vivo
S. Typhimurium infection when dissecting different local and systemic compartments, especially given the fact that
Salmonella infection is generally believed to be intracellular and that the inflammasome has previously been described to be crucial for the host defense against other intracellular pathogens [
9].
NLRP3 can be triggered via terminal signals such as lysosomal rupture and the release of cathepsins and potassium, and the production of reactive oxygen species. The precise bacterial trigger for NLRP3 in
S. Typhimurium infection remains unclear, since it has been demonstrated that
Salmonella pathogenicity island-2 T3SS mutants (which lack the capacity to replicate intracellularly) are still detected by NLRP3 [
9],[
18],[
28]. More recently, it was shown that NLRP3 inflammasome activation is bacterial RNA-mediated [
29], which might explain the NLRP3-dependent sensing of these mutants
. Viable
S. Typhimurium can trigger both the NRLP3 and NLRC4 inflammasomes in vitro, resulting in activation of caspase-1, without involvement of other inflammasomes. The release of IL-1β was shown to be completely dependent on the combined function of the NLRP3 and NLRC4 inflammasomes and blocking either one lead to partially decreased IL-1β release [
18]. The adaptor molecule ASC is a crucial component of both the NLRP3 and NLRC4 inflammasomes as it can bridge the pyrin and CARD domain of these NRLs to caspase-1, which is necessary for its activation [
30].
Asc
−/−
bone-marrow derived macrophages stimulated with
S. Typhimurium in vitro had defects in both the NLRP3 and NLRC4 inflammasomes with regard to releasing mature IL-1β [
18]. Although these studies all underscore the potential role for the inflammasome during
Salmonella infection in vitro, previous in vivo studies were not as evident.
During
S. Typhimurium infection IL-1β is reported to be critical for the intestinal phase of the disease, while IL-18 is important for resistance to systemic infection but not the early gastrointestinal phase of infection [
19]. Mice lacking both
Nlrp3 and
Nlrc4 genes showed an increased susceptibility to
S. Typhimurium infection, similar to
caspase-1
−/−
or
Il-18
−/−
mice, whereas mice lacking either NLRP3 or NLRC4 did not succumb earlier to infection, which was ascribed to the ability of NLRP3 and NLRC4 to be engaged by distinct signals [
18],[
19],[
21]. In previous studies using a typhoid infection model,
Asc
−/−
mice were shown to have comparable bacterial loads (MLNs, spleen and liver) relative to WT mice when challenged with a different
Salmonella strain (
S. Typhimurium SL1344), while their serum IL-18 levels remained low, albeit not significantly different [
18],[
21]. It was therefore speculated that ASC may play a role in vivo in the maturation of cytokines, whereas it is not essential for restricting bacterial growth, suggesting that additional ASC-independent pathways could be involved in activating caspase-1 in response to NLRC4 and NLRP3 activation [
18]. Indeed, it was shown that NLRC4 could also activate caspase-1 even in the absence of ASC, to induce pyroptosis, ASC however remains to be crucial for cytokine maturation during
Salmonella infection [
17],[
22]. Furthermore, intracellular
Salmonella could induce a form of lytic cell death via caspase-11 [
20], which was IFN-γ-mediated [
15],[
31]. To study the in vivo role of the ASC and NLRP3 molecules in more depth we here made use of two different lethal
S. Typhimurium models modeling typhoid and colitis disease. We showed that IL-18 release in plasma of
Asc
−/−
mice was significantly decreased in both models and previous studies have suggested that impaired IL-18 release could enhance
S. Typhimurium susceptibility in mice [
19]. Surprisingly, we were unable to find a role for either ASC or NLRP3 in the host defense against
S. Typhimurium infection in the typhoid model: no differences were found in bacterial counts in any organ (MLNs, liver, spleen and blood) at all time points. Furthermore, proinflammatory cytokine levels (TNF-α, IL-6, IL-10, MCP-1, IFN-γ), markers for hepatocellular injury (AST, ALT), cell damage in general (LDH) and organ pathology (liver, spleen) did not differ between groups. In a similar study using a
Burkholderia pseudomallei infection model, it was noted that IL-18 release was drastically reduced in
Asc
−/−
and
Nlrp3
−/−
mice, although it was still detectable in these mice at higher levels than uninfected mice, leading to the conclusion that inflammasome-independent production of IL-18 may be sufficient to provide some level of protection against infection with low dosages of
Burkholderia pseudomallei[
32]. This could be an explanation for the phenotype observed in the
Nlrp3
−/−
mice, which was comparable to WT mice, but not for the
Asc
−/−
mice, as in both our infection models the IL-18 levels of
Asc
−/−
mice remained below detection limit independent of the stage of infection.
Although no differences were observed between bacterial loads in WT and
Asc
−/−
or
Nlrp3
−/−
mice in the typhoid model, small differences in the streptomycin pre-treated group were seen with regard to bacterial loads (MLN) and splenic and hepatic inflammation between groups. A potential explanation for the different roles played by ASC in both models with regards to hepatocellular injury could be the marked difference in IFN-γ release: where IFN-γ was not reduced in the typhoid model, it was reduced in
Asc
−/−
mice in the colitis model. In
Salmonella- infected mice pre-treated with streptomycin, IFN-γ seems to coordinate T-cell responses and the anti-microbial activity of phagocytes [
33].
Il-18
−/−
mice show an increased susceptibility to
S. Typhimurium infection [
19]. Interestingly, in vivo IL-18 neutralization causes a marked decrease in circulating IFN-γ levels during
Salmonella infection [
16]. It could therefore be hypothesized that the increased susceptibility to
S. Typhimurium of
Il-18
−/−
mice could in part be caused by the low circulating IFN-γ levels and not by absence of IL-18 per se. The normal IFN-γ response in the typhoid model could potentially trigger the described ASC-independent pathway via caspase-11 and IFN-γ [
20]. It has to be noted, however, that this does not explain the observed differences between IFN-γ levels in the typhoid- and colitis model in
Asc
−/−
mice and the increased hepatocellular injury inflicted upon
Nlrp3
−/−
mice despite lower but not significantly different IFN-γ levels in the colitis model.
Our study has several limitations. Pre-growth conditions of
Salmonella could have influenced our results as it determines expression of
S PI-1 and
S PI-2 expression among other virulence factors necessary for inflammasome activation [
9],[
29]. We choose to use log-phase bacteria in our models according to standard methods in order to induce experimental infection in mice. Furthermore, the NLRC4 receptor has been described to play an important role during
Salmonella infection [
18],[
34]; adding a functional knockout mice to our experimental protocol could have given additional insights into the role of this inflammasome receptor in the typhoid and colitis model. It should be noted however that at present 22 human NLRs have been described and 34 mouse NLRs [
35]. A plausible alternative interpretation of our work might be the presence of notable redundancy in the mouse system of NLRs so that one or more NLRs may be able to take over the role of NLRP3 in knockout mice (but not necessarily in humans), and that a double (or triple) knockout may be necessary in order to see a phenotype (
Asc/Il-18 double KO for instance). Indeed, Broz et al
. recently showed that
Nlrp3/
Nlrc4 double knockout mice have a markedly increased susceptibility for invasive salmonellosis when compared to WT controls [
18].
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
Conceived and designed the experiments: HdJ, JvD, TvdP, and JW. Performed the experiments: HdJ, GK and MvL. Analysed the data: HdJ, MvL, GK, JR, JW. Wrote the paper: HdJ, TvdP, JW. All authors approved the final version for submission.