Lactobacillus rhamnosus GR-1 Ameliorates Escherichia coli-Induced Activation of NLRP3 and NLRC4 Inflammasomes With Differential Requirement for ASC

Escherichia coli is a common cause of mastitis in dairy cows. The adaptor protein apoptosis-associated speck-like protein containing a caspase recruitment domain (ASC) synergizes with caspase-1 to regulate inflammasome activation during pathogen infection. Here, the ASC gene was knocked out in bovine mammary epithelial (MAC-T) cells using clustered, regularly interspaced, short palindromic repeat (CRISPR)/CRISPR-associated (Cas)-9 technology. MAC-T cells were pre-incubated with and without Lactobacillus rhamnosus GR-1 and then exposed to E. coli. Western blot analysis demonstrated increased expression of NLRP3 and NLRC4 following E. coli infection, but this increase was attenuated by pre-incubation with L. rhamnosus GR-1, regardless of ASC knockout. Western blot and immunofluorescence analyses revealed that pre-incubation with L. rhamnosus GR-1 decreased E. coli-induced caspase-1 activation at 6 h after E. coli infection, as also observed in ASC-knockout MAC-T cells. The E. coli-induced increase in caspase-4 mRNA expression was inhibited by pre-incubation with L. rhamnosus GR-1. ASC knockout diminished, but did not completely prevent, increased production of IL-1β and IL-18 and cell pyroptosis associated with E. coli infection, whereas pre-incubation with L. rhamnosus GR-1 inhibited this increase. Our data indicate that L. rhamnosus GR-1 suppresses activation of ASC-dependent NLRP3 and NLRC4 inflammasomes and production of downstream IL-lβ and IL-18 during E. coli infection. L. rhamnosus GR-1 also inhibited E. coli-induced cell pyroptosis, in part through attenuation of NLRC4 and non-canonical caspase-4 activation independently of ASC.


INTRODUCTION
Escherichia coli is a frequent cause of bovine mastitis and a leading cause of clinical mastitis in bovine (Shaheen et al., 2015). The NLR family member pyrin domain-containing protein 3 (NLRP3) inflammasome is considered a suitable target for new alternatives to antibiotics to treat bovine mastitis (Thacker et al., 2012). Our previous study showed that probiotic Lactobacillus rhamnosus GR-1 ameliorates E. coli-induced inflammatory damage via attenuation of apoptosis-associated speck-like protein containing a caspase recruitment domain (ASC)-independent NLRP3 inflammasome activation in primary bovine mammary epithelial cells (PBMCs) (Wu et al., 2016). Therefore, L. rhamnosus GR-1 represents a potentially promising therapeutic agent targeting inflammasome activity in E. coli-associated bovine mastitis.
Binding of lipopolysaccharide (LPS) from gram-negative bacteria to toll-like receptor (TLR) 4 increases cellular expression of NLRP3 protein through nuclear factor-κB (NF-κB) signaling, leading to rapidly NLRP3 activation (Afonina et al., 2017). Upon activation, NLRP3 nucleates the adaptor protein ASC through interaction with the pyrin domain (PYD). Procaspase-1 is subsequently autoproteolytically processed through CARD-CARD (caspase recruitment domain) interactions in the NLRP3/ASC complex scaffold and cleaves precursors of the proinflammatory interleukin (IL)-1 family into their bioactive forms, IL-1β and IL-18. We found that L. rhamnosus GR-1 reduces E. coli-induced caspase-1 activation and production of IL-1β and IL-18. However, in contrast to increases in the expression of NLRP3 and caspase-1, expression of the adaptor protein ASC is decreased in PBMCs infected with E. coli, even in cells pretreated with L. rhamnosus GR-1 (Wu et al., 2016).
In contrast to the multiple stimuli that activate NLRP3, NLRC4 is activated by flagellin and the rod protein EscI of the E. coli type III secretion system (T3SS) apparatus (Miao et al., 2010). NLRC4 contains a CARD motif, through which it directly oligomerizes with caspase-1 independent of ASC; this complex activates caspase-1 without autoproteolysis, triggering pyroptosis, an inflammatory form of cell death (Broz et al., 2010b). However, ASC greatly enhances the efficiency of NLRC4-mediated maturation of IL-1β and IL-18 by inducing caspase-1 autoproteolysis (Lamkanfi and Dixit, 2014). NLRC4-dependent production of IL-1β is induced by pathogenic Salmonella or Pseudomonas but not commensal Lactobacillus plantarum, indicating that the NLRC4 inflammasome specifically discriminates pathogens and probiotic bacteria (Franchi et al., 2012). However, the contributions of the NLRC4 inflammasome to inflammatory responses that control E. coli infections are less clear in relation to L. rhamnosus GR-1.
NLRP3 and NLRC4 inflammasomes play a crucial role in potentiating the host antimicrobial response (Guo et al., 2015). Studies using ASC-deficient cells from ASC −/− mice demonstrated the dual role of ASC in bridging NLRP3 and NLRC4 inflammasomes and caspase-1 via PYD and CARD and regulating the result of inflammasome activation (Broz et al., 2010a;Gueya et al., 2014). ASC-dependent inflammasome activation results in the production of proinflammatory IL-1 family cytokines, whereas ASC-independent inflammasome activation induces cell pyroptosis. Given the significant potential of IL-1 family cytokines to cause detrimental inflammation and pyroptosis to control the spread of intracellular pathogens (Jorgensen et al., 2016;Lannitti et al., 2016), the role of ASC in regulating inflammasome activity during E. coli infection must be examined in detail to determine and how L. rhamnosus GR-1 regulates the immune response to prevent E. coli-associated bovine mastitis.
In the present study, we knocked out the ASC gene in bovine mammary epithelial (MAC-T) cells using the RNAguided clustered regularly interspaced short palindrome repeats (CRISPR)-CRISPR-associated nuclease 9 (Cas9) system. We hypothesized that during E. coli infection, the activity of NLRP3 and NLRC4 inflammasomes is differentially regulated by L. rhamnosus GR-1, inducing maturation of IL-1β and IL-18 or cell pyroptosis, depending on ASC. We provide evidence that L. rhamnosus GR-1 suppresses E. coli-induced ASCdependent activation of NLRP3 and NLRC4 inflammasomes and thus decreases production of IL-lβ and IL-18 during E. coli infection. In addition, L. rhamnosus GR-1 suppresses E. coli-induced cell pyroptosis, in part through attenuation of NLRC4 inflammasome and non-canonical caspase-4 activation, independent of ASC.

Biosecurity Statement
All bacterial strains were treated in strict accordance with the Regulations on Biological Safety Management of Pathogen Microbiology Laboratory (000014349/2004-00195) from the State Council of the People's Republic of China. The E. coli CVCC1450 was subjected to all necessary safety procedures to avoid pathogen transmission and infection.

Cell Culture and Transfection
MAC-T cells transferred with the SV40 T antigen (Huynh et al., 1991) was a gift from Dr. Ying Yu (China Agricultural University). MAC-T cells were cultured in Dulbecco's Modified Eagle medium/Ham's F-12 medium (1:1) supplemented with 10% heat-inactivated fetal calf serum, 100 U/mL of penicillin, and 1 g/mL of streptomycin (Invitrogen, Carlsbad, CA, United States) at 37 • C in an atmosphere of 5% CO 2 and 95% air at 95% relative humidity.

Sequencing and Protein Analysis of the Gene Target Site
Genomic DNA samples were extracted using a TIANamp Genomic DNA Kit (Tiangen, Beijing, China) according to the manufacturer's instructions, and 50 ng of DNA template was used to amplify the 630-bp fragment encompassing the gene inactivation locus in 25 µl of PCR buffer (Takara, Shiga, Japan) using primer pairs listed in Table 1. The resulting PCR products were purified and subsequently sequenced to identify deletions.
In addition, clonal cell population whole-cell extracts were analyzed by Western blotting.

Immunocytochemistry
The epithelial origin of MAC-T cells was tested by staining for cytokeratin 18. MAC-T cells (6 × 10 4 cells/well) were seeded into a 24-well culture plate with glass coverslips. After 24 h, cells were washed three times with phosphate-buffered saline (PBS) and fixed with 4% paraformaldehyde for 15 min on ice. The cells were then permeabilized with 0.2% (v/v) Triton X-100 (Sigma-Aldrich) and blocked with 1% bovine serum albumin. Subsequently, cells were incubated with mouse anti-cytokeratin-18 primary monoclonal antibody at a dilution of 1:200 (Ab668; Abcam, Cambridge, United Kingdom) for 45 min at 4 • C, following by incubation with secondary antibody, goat anti-mouse fluorescein isothiocyanate (FITC)conjugated IgG (F4143; Sigma-Aldrich). Cell nuclei were stained using 4 ,6 -diamidino-2-phenylindole (DAPI; Sigma-Aldrich). Coverslips were imaged on an FV1000 confocal laser scanning biological microscope (Olympus, Tokyo, Japan). Escherichia coli CVCC1450 (serotype O111:K58) was purchased from the China Institute of Veterinary Drug Center (Beijing, China) and grown in Luria-Bertani (LB) broth (Oxoid). After overnight incubation at 37 • C with vigorous shaking, bacteria were diluted 1:100 in fresh LB and grown for approximately 3 h until reaching mid-log phase (OD 600 of 0.5).

Adhesion Assay
Wild-type (WT) and ASC −/− MAC-T cells (3 × 10 5 cells/well) were seeded onto a six-well transwell collagencoated polytetrafluoroethylene (PTFE) filter. Confluent cell monolayers were pretreated with L. rhamnosus GR-1 (3 × 10 7 CFU) for 3 h, and then were washed three times with PBS and exposed to E. coli (3 × 10 7 CFU). At 1.5, 3, and 6 h after E. coli challenge, the cell monolayers were washed four times with PBS to remove non-adherent bacteria and treated with 0.05% trypsin for 10 min at 37 • C. Cells were harvested by centrifugation for 10 min at 4000 g and lysed using 100 µl of 0.2% Triton X-100 (Sigma-Aldrich) in sterile water. The populations of E. coli and L. rhamnosus GR-1 were determined on LB and MRS agar plates, respectively. The adhesion rate of E. coli was defined as the adhered E. coli population on the cells pretreated with L. rhamnosus GR-1 relative to the adhered E. coli population in the adhesion assay of E. coli infection alone.

Lactate Dehydrogenase (LDH) Assay
The death of WT or ASC −/− MAC-T cells under the different conditions was assessed using the CytoTox 96 Non-Radioactive Cytotoxicity Assay (Promega, Madison, WI, United States) according to the manufacturer's instructions. The assay measures the release of LDH into the supernatant, calculated as the percentage of total LDH content as determined from cell lysates (100%). LDH released by uninfected cells was used as a maximum-lysis control. The percentage of LDH released was calculated using the following equation: [(LDH infected-LDH uninfected)/(LDH total lysis-LDH uninfected)] × 100.

Enzyme-Linked Immunosorbent Assay (ELISA)
The concentrations of IL-1β and IL-18 in cell-free supernatants of WT or ASC −/− cells were determined at 1.5, 3, and 6 h after E. coli infection using commercially available ELISA kits specific for bovine IL-1β (DG90995Q) and bovine IL-18 (DG91524Q; Beijing Dongge Biotechnology Co., Beijing, China).

Quantification of Lactate Content
Cell culture supernatants were collected. Lactate content in the supernatants were quantified using the enzymatic kit K-DLATE (Megazyme, Bray, Ireland) that allows the measurement of both D-lactate and L-lactate.

Statistical Analysis
Statistical analysis was performed using the SAS statistical software package, version 9.1 (SAS Institute Inc., Cary, NC, United States). With regard to small sample sizes, normal distribution and homogeneity of variance were assumed using the UNIVARIATE (Shapiro-Wilk test) and HOVTEST procedures. Natural logarithm transformation was performed prior to analysis for IL-1β and IL-18 data to yield a normal distribution. Statistical significance of differences was tested using ANOVA procedures, following Tukey's honestly significant difference post hoc test. Data of adhesion assay were compared by an unpaired two-tailed Student's t-test. Data were visualized using GraphPad Prism5 software (GraphPad Software Inc., San Diego, CA, United States). Data from adhesion assay are presented as the mean ± standard deviation (SD) and data from Western blotting, LDH, ELISA and lactate quantification assays are presented as the mean ± standard error of the mean (SEM). Results are representative of three independent experiments, each performed in triplicate. P-values: * P < 0.05; * * P < 0.01; * * * P < 0.001.

CRISPR/Cas9 Mediates Knockout of ASC in MAC-T Cells
To demonstrate the role of ASC in L. rhamnosus GR-1 modulation of inflammasome activation during E. coli infection, we attempted knockout of the ASC gene in MAC-T cells using the CRISPR/Cas9 system. Upon immunocytochemistry analysis, MAC-T cells showed intense positive staining for epithelial cell-specific cytokeratin-18 in the cytoplasmic meshwork of cytokeratin fibrils (Supplementary Figure S1). Compared with program W-001, after electroporation with program T-020, MAC-T cells exhibited higher transfection efficiency ( Figure 1A). Thus, the ASC gene knockout experiment was subsequently performed in MAC-T cells using program FIGURE 2 | Pre-incubation with Lactobacillus rhamnosus GR-1 reduced the adhesion of E. coli to MAC-T cell monolayers. An adhesion assay was performed. Cells and supernatants were collected. The number of adherent E. coli and L. rhamnosus GR-1 was determined (A). The adhesion rate of E. coli was defined as the adhered E. coli population on the cells pretreated with L. rhamnosus GR-1 relative to the adhered E. coli population in the adhesion assay of E. coli infection alone (B). The number of E. coli recovered was determined in adhesion assay supernatants (C). Data are presented as the mean ± SD of three independent experiments. Asterisks indicate the significances among different treatments in the same cell type. Pound signs indicate the significance among WT and ASC −/− MAC-T cells infected with E. coli alone at 3 h. * P < 0.05, * * P < 0.01, ## P < 0.01.
T-020. Among the three designed sgRNA sequences, the specific 20-nucleotide sgRNA1 sequence targeted the exon 1 regions of the ASC gene and directed Cas9 nuclease to precisely introduce a DNA double-strand break in front of a protospacer adjacent motif (PAM). After cloning and sequencing of the DNA fragment, an 8-bp deletion was observed ( Figure 1B). Western blot analysis did not show expression of ASC protein in sgRNA1 sequence-targeted MAC-T cells ( Figure 1C). Furthermore, E. coli infection triggered assembly of ASC specks in WT cells, whereas pre-incubation with L. rhamnosus GR-1 attenuated E. coli-induced ASC speck assembly. No ASC specks were observed in ASC −/− cells, regardless of E. coli infection ( Figure 1D). These results demonstrated that knockout of the ASC gene in MAC-T cells was successful.
Pre-incubation With L. rhamnosus GR-1 Reduces the Adhesion of E. coli to

MAC-T Cell Monolayers
Escherichia coli or L. rhamnosus GR-1 exhibited similar adhesion capacity in WT and ASC −/− MAC-T cells. The number of adherent E. coli was about 1.29 × 10 4 ± 0.67 × 10 2 CFU (means ± standard deviation) at 1.5 h after E. coli infection, Results are presented as the ratio of NLRP3 band intensity to that of GAPDH (Right). Data are presented as the mean ± SEM of three independent experiments. * P < 0.05, * * P < 0.01, * * * P < 0.001. and increased to 1.95 × 10 4 ± 1.17 × 10 3 CFU at 6 h. At 3 h after E. coli challenge, the number of adherent E. coli in ASC −/− cells was lower (P = 0.007) than in WT MAC-T cells (Figure 2A). L. rhamnosus GR-1 had a lower adhesion capacity and the number of adherent L. rhamnosus GR-1 was about 8.75 × 10 3 ± 1.56 × 10 3 CFU, regardless of infection time. Preincubation with L. rhamnosus GR-1 resulted in a reduction in the E. coli adhesion rate to 57% of that observed in MAC-T cells infected with E. coli alone (Figure 2B). No differences were observed in the number of E. coli recovered from the supernatant fraction among two groups ( Figure 2C).

Pre-incubation With L. rhamnosus GR-1 Attenuates E. coli-Induced NLRP3 Expression and Increases Lactate Content in the Supernatants
Compared with untreated control WT cells, Western blot analysis showed an increase in NLRP3 protein expression at 1.5, 3, and 6 h after E. coli challenge in cells only infected with E. coli, but not in cells incubated with L. rhamnosus GR-1 alone (P = 0.046, P < 0.001, and P < 0.001, respectively; Figures 3A-C). In contrast, WT cells pre-incubated with L. rhamnosus GR-1 had a lower expression of NLRP3 protein than did WT cells only infected with E. coli at 3 and 6 h (P < 0.001 for both).
The lactate content (D-lactate plus L-lactate) in the supernatants was quantified. Compared with untreated control cells, the lactate content was increased in the supernatants from both WT and ASC −/− cells incubated with L. rhamnosus GR-1 alone or pre-incubated with L. rhamnosus GR-1, but not the cells only infected with E. coli at 1.5, 3, and 6 h after infection, regardless of ASC knockout (P < 0.05; Supplementary Figure S2A). Western blot analysis showed that compared with untreated control cells, at 6 h after E. coli challenge, NLRP3 protein expression was elevated in WT or ASC −/− cells infected with E. coli, but not in cells only treated with lactate alone (P < 0.05; Supplementary Figure S2B). Lactate addition did not attenuate E. coli-induced increase in NLRP3 protein expression in either WT or ASC −/− cells. Data are presented as the mean ± SEM of three independent experiments. * P < 0.05, * * P < 0.01, * * * P < 0.001.

Effect of L. rhamnosus GR-1 on NLRC4 Activation During E. coli Infection
Compared with untreated control cells, expression of NLRC4 protein was elevated at 3 h after E. coli infection in WT cells only infected with E. coli (P = 0.027; Figure 4B) but not in WT cells incubated with L. rhamnosus GR-1 alone or pre-incubated with L. rhamnosus GR-1. Compared with WT cells only infected with E. coli, expression of NLRC4 protein at 3 h was decreased in WT cells incubated with L. rhamnosus GR-1 alone and pre-incubated with L. rhamnosus GR-1 (P = 0.019 and P = 0.001, respectively). No differences were observed at 3 h in ASC −/− cells, regardless of treatment.
Compared with untreated control cells, E. coli challenge led to increased expression of NLRC4 protein at 6 h after E. coli challenge in WT cells (P = 0.019) and ASC −/− cells (P < 0.001; Figure 4C). Expression of NLRC4 protein was lower in ASC −/− cells incubated with L. rhamnosus GR-1 alone and pre-incubated with L. rhamnosus GR-1 (P = 0.001 for both) than in ASC −/− cells only infected with E. coli. No changes were observed at 1.5 h after E. coli infection in WT cells or ASC −/− cells, regardless of treatment ( Figure 4A).

Pre-incubation With L. rhamnosus GR-1 Attenuates E. coli-Induced Caspase-1 Maturation
Immunofluorescence staining showed that compared with untreated control cells, E. coli challenge triggered assembly of ASC specks in WT cells at 6 h after E. coli challenge, and this was attenuated in WT cells pre-incubated with L. rhamnosus GR-1 (Figure 5A). No ASC specks were observed in ASC −/− cells, regardless of treatment. Compared with untreated control cells, bright foci indicative of increased caspase-1 expression was observed in WT cells only infected with E. coli at 6 h after challenge; this increase was attenuated by incubation with L. rhamnosus GR-1 (Figure 5B). Compared with WT cells, ASC deletion attenuated, but did not abolish, caspase-1 staining in response to E. coli infection and preincubation with L. rhamnosus GR-1. A punctate staining pattern for caspase-1 was observed in ASC −/− cells only infected with E. coli, and pre-incubation with L. rhamnosus GR-1 inhibited the E. coli-induced punctate caspase-1 staining pattern. Compared with untreated control cells, incubation with L. rhamnosus GR-1 only did not result in increase in caspase-1 activation either in WT or ASC −/− cells ( Figure 5B).
Compared with untreated control cells, increased maturation of procaspase-1 into its catalytic 10-kDa subunit was observed at 6 h after E. coli challenge in WT and ASC −/− cells only infected with E. coli (P < 0.001 for both; Figure 6C). However, maturation of caspase-1 declined in WT and ASC −/− cells pre-incubated with L. rhamnosus GR-1 (P = 0.014 and P = 0.006, respectively) or incubated with L. rhamnosus GR-1 alone (P < 0.001 for both), compared with WT cells infected with E. coli only. No differences were observed in caspase-1 maturation at 1.5 and 3 h after E. coli infection in WT cells or ASC −/− cells, regardless of treatment (Figures 6A,B).

Pre-incubation With L. rhamnosus GR-1 Attenuates E. coli-Induced Caspase-4 Activation
At 3 h after E. coli infection, expression of cleaved caspase-4 (26 kDa) in WT and ASC −/− cells pre-incubated with L. rhamnosus GR-1 was lower than in untreated control cells and cells only infected with E. coli (Figure 7B). Infection with E. coli resulted in increased expression of cleaved caspase-4 in WT and ASC −/− cells (P = 0.001 and P = 0.004, respectively; Figure 7C) at 6 h. However, expression of caspase-4 decreased in WT and ASC −/− cells pre-incubated with L. rhamnosus GR-1, compared with cells only infected with E. coli (P < 0.001 and P = 0.001, respectively). No differences were observed in the expression of cleaved caspase-4 at 1.5 h after E. coli infection in WT or ASC −/− cells, regardless of treatment ( Figure 7A).

Pre-incubation With L. rhamnosus GR-1 Attenuates E. coli-Induced Production of IL-1β and IL-18 and Cell Pyroptosis
Compared with untreated control cells, IL-1β production was increased at 3 and 6 h after E. coli infection in WT cells only infected with E. coli (P < 0.001 for both; Figure 8A). IL-1β production was lower at 3 and 6 h both in WT cells incubated with L. rhamnosus GR-1 alone (P = 0.002 and P = 0.004, respectively) and cells pre-incubated with L. rhamnosus GR-1 (P = 0.013 and P = 0.011, respectively) than in WT cells only infected with E. coli. Compared with WT cells, ASC deletion led to a decrease in production of IL-1β in ASC −/− cells in response to different treatments. Challenge with E. coli resulted in elevated production of IL-1β at 1.5, 3, and 6 h after E. coli Results are presented as the ratio of caspase-1 p10 subunit band intensity to that of GAPDH (Right). The detection of NLRP3 and caspase-1 proteins was performed in the same gel. Caspase-1 blots shared the same GAPDH blots with NLRP3 protein. Data are presented as the mean ± SEM of three independent experiments. * P < 0.05, * * P < 0.01, * * * P < 0.001. infection in ASC −/− cells only infected with E. coli (P = 0.009, P = 0.021, and P = 0.001, respectively; Figure 8A) compared with untreated control cells; however, the E. coli-induced increase in IL-1β production was attenuated at 3 and 6 h by incubation with L. rhamnosus GR-1 alone (P = 0.014 and P = 0.007, respectively) and pre-incubation with L. rhamnosus GR-1 (P = 0.009 and P = 0.003, respectively).
IL-18 exhibited similar differential production as IL-1β. Compared with untreated control cells, production of IL-18 was increased at 3 and 6 h after E. coli challenge both in WT (P = 0.001 for both) and ASC −/− (P < 0.001 for both) cells only infected with E. coli ( Figure 8A). However, incubation with L. rhamnosus GR-1 alone and pre-incubation with L. rhamnosus GR-1 attenuated the E. coli-induced elevation in the concentration of IL-18 at 3 and 6 h after E. coli challenge both in WT and ASC −/− cells. Compared with untreated control ASC −/− cells, the concentration of IL-18 was elevated at 3 and 6 h in ASC −/− cells pre-incubated with L. rhamnosus GR-1 (P = 0.008 and P = 0.002, respectively).
Cell pyroptosis was quantified by monitoring the release of LDH into the supernatants after E. coli challenge. Compared with untreated control cells, the percentage of pyroptotic cells at 3 and 6 h was increased in WT cells only infected with E. coli (P < 0.001 for both; Figure 8B) but not in WT cells incubated with L. rhamnosus GR-1 alone and pre-incubated with L. rhamnosus GR-1. Compared with WT cells, ASC deletion let to a similarly differential but attenuated cell pyroptosis. Compared with untreated control cells, the percentage of pyroptotic cells was increased at 3 and 6 h in ASC −/− cells only infected with E. coli (P = 0.020 and P < 0.001, respectively), whereas incubation with L. rhamnosus GR-1 alone and pre-incubation with L. rhamnosus GR-1 ameliorated the E. coli-induced increase in pyroptotic cell death at 6 h (P < 0.001 and P = 0.002, respectively). No changes were observed in WT and ASC −/− cells, regardless of treatment.

DISCUSSION
Bacterial adhesion to host epithelial cells is an essential step in the initiation of infection. Lactobacillus can reduce pathogen adhesion to epithelial cells and exert direct antimicrobial activity due to accumulation of antimicrobial substances (Gudina et al., 2015). We found that L. rhamnosus GR-1 did not directly kill E. coli, but did decrease the level of E. coli adhesion to 57% of that observed in MAC-T cells infected with E. coli alone. We previously revealed that live and ultravioletirradiated L. rhamnosus GR-1 rather than culture supernatant of L. rhamnosus GR-1 and medium acidified with lactate lead to a decrease in the E. coli adhesion rate in bovine mammary epithelial cells (Wu et al., 2016). The reduced E. coli adhesion level mediated by L. rhamnosus GR-1 may be attributed to steric hindrance due to competition for attachment sites (Ardita et al., 2014;Tytgat et al., 2016).
NLRP3 is activated by a wide variety of stimuli, including pore-forming toxins, extracellular adenosine triphosphate, RNA-DNA hybrid molecules, and pathogens (Jo et al., 2016). We found that E. coli infection also increased the expression of NLRP3 protein from 1.5 to 6 h, but L. rhamnosus GR-1 pretreatment inhibited this increase. NLRP3 must be primed before activation. Escherichia coli LPS binds to TLR4 to induce expression of NLRP3 protein via NF-κB signaling. Bacterial mRNA from viable E. coli cells that have been phagocytosed enters the cytosolic compartment, resulting in assembly of the NLRP3 inflammasome (Sander et al., 2011). Intake of L. plantarum CECT 7315/7316 downregulates expression of Nlrp3 in the ileum of rats (Vilahur et al., 2015). However, the NLRP3 inflammasome is also activated by L. rhamnosus GG and LC705 originating from dairy sources (Miettinen et al., 2012).
In a mouse immune hepatitis model, lactate treatment was shown to attenuate hepatic and pancreatic injury by negatively regulating TLR4-mediated activation of the NLRP3 inflammasome and production of IL-1β through arrestin β2 and G-protein-coupled receptor 81 (Hoque et al., 2014). In the present study, there was a higher lactate content in the supernatants of cells incubated with L. rhamnosus GR-1. However, additional lactate treatment did not attenuate the E. coli-induced increase in expression of NLRP3. This indicates that the elevated lactate content in the supernatant is a secondary effect of L. rhamnosus GR-1 treatment and cannot account for attenuation of E. coli-induced activation of NLRP3. Previously, we have shown that L. rhamnosus GR-1 attenuates E. coli-induced TLR4 expression in bovine mammary epithelial cells (Wu et al., 2016). This may contribute to attenuating the priming step and subsequent activation of NLRP3 through TLR4-mediated NF-κB signaling (Bauernfeind et al., 2009). The mitogen-activated protein kinase (MAPK) subfamilies, c-Jun N-terminal kinase (JNK) and extracellular regulated protein kinase (ERK) are essential for NLRP3 inflammasome activation and addition of JNK1/2 inhibitor SP600125 or upstream MAPK/ERK kinase inhibitor PD98059 of ERK inhibits the LPS-induced increase in NLRP3 protein expression (Liao et al., 2013). Culture medium of L. rhamnosus GR-1 inhibits LPS-induced JNK FIGURE 8 | Pre-incubation with L. rhamnosus GR-1 attenuated E. coli-induced production of IL-1β and IL-18 and cell pyroptosis. Secretion of IL-1β and IL-18 (A) into the indicated cell culture supernatants at 1.5, 3, and 6 h after E. coli challenge as determined by ELISA. Cell pyroptosis was determined by measurement of LDH release (B). Data are presented as the mean ± SEM of three independent experiments. * P < 0.05, * * P < 0.01, * * * P < 0.001. activation in macrophages or monocytic THP-1 cells (Kim et al., 2006). Histamine derived from Lactobacillus reuteri 6475 inhibits activation of ERK in THP-1 cells (Thomas et al., 2012). It is possible that L. rhamnosus GR-1 attenuates E. coliinduced NLRP3 activation through reducing the adhesion of E. coli to MAC-T cells and subsequently negatively regulating the functional synergy between NF-κB and JNK/ERK MAPK pathways mediated by some uncertain soluble factors. Further studies are required to determine active components derived from L. rhamnosus GR-1 and elucidate possible mechanisms underlying the antagonistic effects of L. rhamnosus GR-1 on NLRP3 activation during E. coli infection.
NLRP3 contains only a PYD, which engages the PYD of ASC, leaving the CARD of ASC to interact with the CARD-containing region of pro-caspase-1. Caspase-1 is thought to be activated by a proximity-induced dimerization and autoproteolytic process in the NLRP3/ASC complex platform (Shi, 2004). Active caspase-1 cleaves pro-IL-lβ and pro-IL-18 into mature IL-lβ and IL-18, which are essential for coordination of immune responses to pathogen infection through allograft neutrophil sequestration, mononuclear phagocyte recruitment, and T-cell activation (Samuel Weigt et al., 2017). In the present study, L. rhamnosus GR-1 attenuated E. coli-induced caspase-1 autoproteolysis and elevated production of mature IL-lβ and IL-18 at 6 h after challenge. Another study also showed that the NLRP3 inflammasome pathway plays a critical role in the host immune response to pathogen infection (Dikshit et al., 2018). However, inappropriate activation of the NLRP3 inflammasome is linked not only to local inflammation but also several autoimmune inflammatory disorders in humans (Seo et al., 2015). Indeed, activation of the NLRP3 inflammasome amplifies inflammation and promotes pathogen infection via a process involving triggering of T helper 2-biased adaptive immune responses (Gurung et al., 2015) or secretion of secondary danger-associated molecular pattern molecules (Bui et al., 2016). Our data suggest that L. rhamnosus GR-1 prevents E. coli-induced inflammation by suppressing activation of ASC-dependent NLRP3 inflammasomes.
In mice, non-canonical caspase-11 was identified as a key regulator of NLRP3 inflammasome-associated caspase-1 activation in response to E. coli infection. Caspase-11 is activated via NLRP3-independent mechanisms, but it is essential for NLRP3-dependent and ASC-dependent caspase-1 processing and IL-1β maturation in response to E. coli infection (Kayagaki et al., 2011). In addition, binding of LPS to human caspase-4 or murine caspase-11 via the CARD directly induces cell pyroptosis, independently of NLRP3 and ASC (Shi et al., 2014). A recent study revealed that outer-membrane-vesicle-mediated cytoplasmic delivery of extracellular E. coli LPS activates murine caspase-11 to induce pyroptosis and IL-1β maturation (Vanaja et al., 2016). Bovine caspase-4 is a homolog of human caspase-4 and mouse caspase-11 and plays a role in the processing of IL-1β and IL-18 precursors (Koenig et al., 2001;Martinon and Tschopp, 2004). In the present study, caspase-4 was activated by E. coli at 6 h; however, L. rhamnosus GR-1 pretreatment attenuated this activation.
FIGURE 9 | Lactobacillus rhamnosus GR-1 attenuated activation of ASC-dependent and ASC-independent inflammasomes during E. coli infection. L. rhamnosus GR-1 ameliorates E. coli-induced inflammatory damage via attenuation of both ASC-dependent and ASC-independent inflammasome activation in MAC-T cells. L. rhamnosus GR-1 inhibits activation of ASC-dependent NLRP3 and NLRC4 inflammasomes and production of the downstream proinflammatory cytokines IL-lβ and IL-18 during E. coli infection. L. rhamnosus GR-1 suppresses E. coli-induced cell pyroptosis, in part through attenuation of NLRC4 inflammasome activation, independently of ASC. In addition, L. rhamnosus GR-1 inhibits non-canonical caspase-4 activation, which subsequently synergizes with the NLRP3 inflammasome to attenuate caspase-1 activation and potentially inhibit caspase-1-independent cell pyroptosis and IL-lβ and IL-18 production. Full lines represent the results of the present study, and dashed lines represent the conclusions drawn in other studies.
Interestingly, there was a lower number of adherent E. coli in ASC −/− MAC-T cells at 3 h after E. coli challenge compared with WT MAC-T cells. It must be noted that ASC deficiency reduced, but did not abolish, caspase-1 processing, IL-1β and IL-18 maturation, and cell pyroptosis during E. coli infection. Decreased E. coli adhesion may delay the expression of NLRC4 receptor and the downstream activation of caspase-1, maturation of proinflammatory cytokines and cell pyroptosis in ASC −/− MAC-T cells. Caspase-1 activation by E. coli requires NLRP3 and ASC, but caspase-11 processing and cell pyroptosis do not (Kayagaki et al., 2011). A previous study reported weaker oligomerization of both ASC and caspase-1 in macrophages infected with E. coli compared with the canonical NLRP3 inflammasome activator nigericin, but with comparable production of IL-1β (Rathinam et al., 2012). Caspase-11 interacts with caspase-1 in infected cells, forming a heterodimeric complex (Kayagaki et al., 2011). These data suggest that caspase-4/-11 amplify caspase-1 activation independently of ASC by enabling caspase-1 autoprocessing through heterodimerization. Indeed, in the present study, caspase-4 activation was enhanced in ASC −/− cells compared with WT cells, which could have compensated for the loss of caspase-1 activation due to ASC-dependent NLRP3 inflammasome activation. Our findings indicate that E. coli infection activates caspase-4, subsequently resulting in cell pyroptosis and maturation of IL-1β and IL-18 via an NLRP3 inflammasome-dependent and ASC-independent pathway. Lactobacillus rhamnosus GR-1 suppresses ASC-dependent NLRP3 inflammasome activation and ASC-independent caspase-1 processing by inhibiting caspase-4 activation, thereby attenuating cell pyroptosis and cytokine production and thus preventing establishment of E. coli infection.
In contrast to NLRP3 activation in response to diverse stimuli, upon E. coli infection, NLRC4 responds to bacterial rod protein of the T3SS apparatus and flagellin (Zhao et al., 2011). We found that L. rhamnosus GR-1 inhibited E. coliinduced NLRC4 expression, as was also observed in ASC −/− cells. NLRC4 contains a CARD motif through which it directly interacts with caspase-1 to induce pyroptosis, independently of ASC. This NLRC4-dependent/ASC-independent cell death pathway proceeds in the absence of caspase-1 autoproteolysis. Interestingly, we observed weaker staining for caspase-1 in ASC −/− cells. Caspase-1 autoproteolysis is often used as an indicator of caspase-1 activation. However, it was also reported that uncleaved caspase-1 is enzymatically active in ASC −/− cells and can induce pyroptosis. In contrast to the formation of a single large ASC/caspase-1 focus for efficient IL-lβ and IL-18 processing, pro-caspase-1 could be recruited to NLRC4, with which it forms a smaller complex that induces pyroptosis (Broz et al., 2010b). Although NLRC4 contains a CARD, ASC amplifies NLRC4 inflammasome activity because ASC is essential for NLRC4-induced caspase-1 autoprocessing and maturation of IL-lβ and IL-18 (Brubaker et al., 2015). Our data suggest that L. rhamnosus GR-1 inhibits E. coli-induced cell pyroptosis via suppression of ASC-independent NLRC4 inflammasome activation. During E. coli infection in the present study, L. rhamnosus GR-1 decreased the secretion of IL-lβ and IL-18, in part due to suppression of ASC-dependent NLRC4 inflammasome activation.
This MAC-T cell model of E. coli and L. rhamnosus GR-1 co-incubation presents an in vitro framework for assessing bovine mammary immune response to pathogen infection, and evaluating the efficiency of Lactobacillus-based intervention in preventing bovine mastitis. The results require further confirmation in other cell lines and in vivo studies. However, several concerns need to be addressed before the clinical application of Lactobacillus in bovine mastitis. Oral ingestion of probiotics promotes mucosal immune response to pathogen infection in the gut. The mechanism underlying how the immunomodulatory effect extends to the mammary glands remains unclear. The means of probiotic supplementation (e.g., mixed into the feed, oral capsules or intra-mammary infection) and dose effect also needs to be studied in more details. The molecular mechanism underlying regulation of inflammasome activity by Lactobacillus requires further investigation. Our findings identify NLRP3 and NLRC4 inflammasomes as potential targets for bovine mastitis therapy and could strengthen the development for other inflammasome-targeted therapies in E. coli-associated mastitis.

AUTHOR CONTRIBUTIONS
QW, Y-HZ, and J-FW conceived and designed the experiments, analyzed the data, and wrote the manuscript. QW, JX, XL, CD, and M-JW performed the experiments.

ACKNOWLEDGMENTS
We would like to thank Professor Sen Wu for the pCRISPR-sg5 plasmid and Dr. Ying Yu for providing MAC-T cells.

SUPPLEMENTARY MATERIAL
The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fmicb. 2018.01661/full#supplementary-material FIGURE S1 | Immunodetection of cytokeratin-18 in MAC-T cells. Representative confocal immunocytochemistry images showing typical morphology of MAC-T cells. MAC-T cells were immunostained for cytokeratin-18 (red). DAPI (blue) was used to localize nuclei. Scale bar, 20 µm. Data are representative of three independent experiments. FIGURE S2 | The lactate content in the cell supernatants and the effect of lactate on NLRP3 expression during E. coli infection. Lactate content in the supernatants was determined (A). Cells were also simultaneously treated with lactate (LACT) at a concentration of 0.6 g/L (equivalent to 7 mM) and E. coli at a MOI of 100:1. Western blot detection of NLRP3 in WT and ASC −/− cells collected from the indicated cell cultures at 6 h after E. coli challenge (B). Representative panels showing expression of NLRP3 protein (Left). NLRP3 band intensity was determined using Quantity One software. Results are presented as the ratio of NLRP3 band intensity to that of GAPDH (Right). Data are presented as the mean ± SEM of three independent experiments. * P < 0.05, * * P < 0.01, * * * P < 0.001.