Vibrio vulnificus induces mTOR activation and inflammatory responses in macrophages

Vibrio vulnificus (V. vulnificus), a Gram-negative marine bacterium, can cause life-threatening primary septicemia, especially in patients with liver diseases. How V. vulnificus affects the liver and how it acts on macrophages are not well understood. In this report, we demonstrated that V. vulnificus infection causes a strong inflammatory response, marked expansion of liver-resident macrophages, and liver damage in mice. We demonstrated further that V. vulnificus activates mTOR in macrophages and inhibition of mTOR differentially regulates V. vulnificus induced inflammatory responses, suggesting the possibility of targeting mTOR as a strategy to modulate V. vulnificus induced inflammatory responses.


Introduction
Vibrio vulnificus (V. vulnificus) is a halophilic, Gram-negative and life-threatening marine bacterium [1,2]. V. vulnificus infection can cause severe primary septicemia in patients with liver diseases such as viral hepatitis, advanced cirrhosis, and chronic alcoholic liver disease [3]. Despite previous studies have significantly advanced our understanding of the pathogenesis of V. vulnificus infection, mechanisms that regulate V. vulnificus-mediated responses remain unclear.
In this report, we demonstrated that V. vulnificus could directly inflict liver damage and induce strong innate immune responses and Kupffer cell expansion. Moreover, V. vulnificus can activate mTOR in macrophages and inhibition of mTOR differentially regulates V. vulnificus induced inflammatory responses.

Results
Acute infection by V. vulnificus CGMCC1.1758 strain induced liver inflammation in C57BL/6J mice Patients with liver diseases are susceptible to severe V. vulnificus infection. One of the liver's key functions is to clear blood-borne pathogens and to prevent bacteria from spreading to other organs to cause sepsis [28]. Several strains of V. vulnificus have been found to cause systemic infection in mice [29][30][31]. However, whether the V. vulnificus 1.1758 strain can cause liver damage has been unclear. To examine whether V. vulnificus is capable of causing liver damage, we intraperitoneally (i.p.) injected V. vulnificus CGMCC1.1758 into wild-type C57BL/6J mice. Mice injected with 1×10 8 CFU V. vulnificus showed higher mortality than those injected with 5×10 7 CFU (Fig 1A). V. vulnificus infection resulted in liver damage, indicated by elevated serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels ( Fig 1B). H&E staining showed piecemeal necrosis, focal inflammation, and portal inflammation in the liver (Fig 1C and 1D). Additionally, we detected V. vulnificus in liver homogenates ( Fig 1E). These observations indicated that V. vulnificus CGMCC1.1758 is capable of inducing liver inflammatory damage in wild-type C57BL/6J mice.

V. vulnificus acute infection-induced Kupffer cell and neutrophil accumulation in the liver
To determine the infiltrating cell types in the liver after V. vulnificus infection, we performed immunofluorescence analysis of frozen liver sections of V. vulnificus-infected and control mice.
As Fig 2A and 2B shows, both CD68 + CD11b + Kupffer cells and CD11b + Ly6G + neutrophils accumulated in the livers of infected mice. To further confirm Kupffer cell accumulation, we performed FACS analysis of liver mononuclear cells. As shown in Fig 2C and 2D, CD11b + F4/ 80 int Kupffer cells were considerably increased in V. vulnificus infected mice. Thus, acute V. vulnificus infection caused Kupffer cell accumulation in the liver.

V. vulnificus acute infection-induced Kupffer cell proliferation and proinflammatory responses in vivo
To determine whether Kupffer cell accumulation in the liver after V. vulnificus infection was associated with increased proliferation, we injected BrdU into mice immediately before V. incorporated more BrdU than those in uninfected mice, suggesting that these cells proliferated in infected mice ( Fig 3A). Kupffer cells from V. vulnificus-infected mice upregulated CD71 (transferrin receptor protein 1, required for iron delivery from transferrin to cells) and CD98 (an amino acid transporter) ( Fig 3B). In vitro, V. vulnificus infection of BMMϕs also upregulated CD71 and CD98 (Fig 3C), suggesting that V. vulnificus directly acted on macrophages to induce expression of these molecules. Elevated expression of CD71 and CD98 might ensure sufficient nutrients for macrophage metabolism, proliferation, and function. Together, these To determine if V. vulnificus induced inflammatory responses, we measured cytokines concentrations in the serum and liver homogenates after V. vulnificus infection using a multiplex flow assay. In V. vulnificus-infected mice, multiple cytokines, such as IL-27, IL-6, IL-10, and MCP-1, obviously increased in both blood and liver (Fig 3D). Several other cytokines, such as IFN-β, IL-23, IL-12 p40, and IL-1α, increased only in the liver, while TNF-αwas increased in serum, but not in the liver. The differences between liver and blood could result from differences in bacterial burdens and in local immune cell compositions in these organs/tissues. Thus, V. vulnificus infection triggered strong innate immune responses in vivo.

Activation of NFκB and mTOR in macrophages by V. vulnificus
Transcription of many cytokines depends on NFκB after stimulating pathogen pattern recognition receptors [5,32]. Bone marrow derived macrophages (BMMϕs) infected with live 0.2 MOI V. vulnificus or treated with heat inactivated V. vulnificus (In) or V. vulnificus lysates for 6 hours contained increased IKKα/β and IκBα phosphorylation (Fig 4A), suggesting that V. vulnificus was capable of triggering IKK-NFκB activation. Low dose (0.2 MOI) V. vulnificus treatment also induced phosphorylation of mTOR, S6K1, S6, 4EBP1 and Akt. However, 2 MOI of V. vulnificus-treated cells did not induce phosphorylation of these proteins (Fig 4A), which was likely caused by negative feedback mechanisms triggered by over-stimulation or by death of BMMϕs after 6 hours of infection. Indeed, after 3 hours treatment, 2 MOI V. vulnificus induced stronger S6K1 and S6 phosphorylation than 0.2 MOI V. vulnificus (Fig 4B). Moreover, Kupffer cells from V. vulnificus-infected mice also contained increased 4E-BP1 and Akt S473 phosphorylation (Fig 4C). These observations suggested that V. vulnificus induced both mTORC1 and mTORC2 activation in macrophages in vitro and in vivo.
To examine if V. vulnificus could directly induce innate immune responses in macrophages, we treated BMMϕs with live V. vulnificus or heat inactivated V. vulnificus, or V. vulnificus culture supernatant (Sup) in vitro for 6 hours. As Fig 4D shows, both live and heat-inactivated V. vulnificus induced IL-6, IL-10, TNF-α, and IL-1α production in these cells. In addition, live but not inactivated V. vulnificus induced GM-CSF, IFNβ, IL-27, and IL-1β production. Interestingly, treatment of BMMϕs with rapamycin during V. vulnificus stimulation resulted in  decreased expression of GM-CSF, IFNβ, IL-27, and MCP-1 without obviously affecting the production of other cytokines. In rapamycin treated BMMϕs, V. vulnificus-induced S6K1 and S6 phosphorylation was abolished and Akt phosphorylation was deminished (Fig 4E), consistent with the notion that rapamycin is able to inhibit not only mTORC1 but also mTORC2. Interestingly, V. vulnificus CFUs were lower in rapamycin treated BMMϕ compared with untreated controls (Fig 4F), which could also contribute to decreased expression of cytokines induced after treatment with rapamycin. Together, these observations suggested that V. vulnificus contain heat resistant and labile components that can induce multiple cytokine production in macrophages and that mTOR is important for V. vulnificus proliferation in BMMϕs and may selectively regulate expression of certain but not all cytokines induced by V. vulnificus.

V. vulnificus infection causes sepsis in the host and induces inflammation via recruiting and
activating immune cells such as neutrophils, monocytes, and macrophages. Extensive work has linked various virulence factors of V. vulnificus to cell death and modulation of innate immune responses to this pathogen [30,[33][34][35][36][37]. Patients with liver diseases such as cirrhosis are more susceptible to developing V. vulnificus infection, which can lead to sepsis, sepsisinduced organ failure, and even death [38]. Surprisingly, little is known about the liver phase during V. vulnificus infection. In this report, we showed that V. vulnificus CGMCC1.1758 was capable of inducing inflammatory responses in the liver and inflicting liver damage in C57BL/ 6J mice, manifested by induction of multiple inflammatory cytokines and elevated serum AST and ALT levels. These observations are consistent with a previous report that another V. vulnificus strain, M06-24/O, is capable of causing liver inflammation and damage in BALB/c mice [31].
As tissue-resident macrophages, Kupffer cells are strategically located to clear bacteria in the liver during infection. Kupffer cells do not obviously proliferate in the steady state [39]. We have found that Kupffer cell numbers drastically increased in mice after V. vulnificus CGMCC1.1758 infection, which is at least partially caused by proliferation. Like V. vulnificus infection, systemic Listeria monocytogenes infection also results in Kupffer cell proliferation [40]. Thus, Kupffer cell expansion after bacterial infection may represent a means to increase the capacity of tissue-resident macrophages for clearance of invading pathogens. In addition to proliferation, V. vulnificus infection could increase Kupffer cells via other mechanisms. Our study provides the first evidence that V. vulnificus can cause mTOR activation. In BMMϕ, either infection with live V. vulnificus or treatment with heat-inactivated V. vulnificus or V. vulnificus lysates induces mTORC1 and mTORC2 activation reflected by 4EBP1 and Akt S473 phosphorylation, respectively. Thus, V. vulnificus contains component(s) that can induce mTOR activation, although these components and the pattern recognition receptors that lead to mTOR activation remain to be identified. As mentioned earlier, mTOR and its tight regulation play important roles in innate immune responses [11,15,16,[20][21][22]. We have found that V. vulnificus infection triggered strong inflammatory responses in vivo and V. vulnificus treatment induced production of multiple cytokines in macrophages. Among these cytokines, GM-CSF, IFNβ, IL-27, and IL-1β production but not IL-6, IL-10, TNF-α, and IL-1α production is dependent on mTOR activity. It would be interesting to further determine the underlying mechanisms by which mTOR controls production of selective cytokines during V. vulnificus infection.

Ethics statement
This study was carried out in strict accordance with the guidelines of the Zhejiang Provincial Animal Care and Use Administration Office (SYXK-ZJ-2005-0061). We performed all the animal experiments in accordance with protocols approved by the Wenzhou Medical University Animal Care and Use Committee (reference: wydw2014-0009).

Bacterial strains, mice and cell culture
We purchased C57BL/6J mice from Wenzhou Medical University Laboratory Animal Facility and were maintained under a SPF environment at the Central Animal Laboratory of Wenzhou Medical University. The mice were kept for 5-7 days to acclimatize after transport from the supplier. In the survival experiment, mice were observed every 30 minutes for 24 hours, and mice that survived for 24 hours were euthanized. L-929 cells were purchased from the Cell Bank of the Chinese Academy of Science in Shanghai and cultured them in RPMI1640 containing 10% heat-inactivated fetal bovine serum (Ausvin) and penicillin-streptomycin (50 IU/ ml and 50 mg/ml; Beyotime). The China General Microbiological Culture Collection Center provided the Vibiro vulnificus CGMCC 1.1758 strain, which we grew at 37˚C in brain heart infusion broth (BHI) or on the BHI rabbit blood agar plate.

In vivo infection
We used i.p. injection to infect female C57BL/6J mice with V. vulnificus. For survival experiments, we i.p. injected mice with 5×10 7 or 1×10 8 CFU of V. vulnificus suspended in 200 μl PBS or with PBS alone. We euthanized and exsanguinated the mice 6 hours after infection for blood and tissue collection. The liver was fixed it in 4% paraformaldehyde for histological analysis, and froze it in OCT medium (Thermo) for immunofluorescence analysis. To measure V. vulnificus CFU in the liver, we perfused infected mice with PBS before harvesting their livers. We cultured liver homogenate on BHI blood agar plates at 37˚C for 24 hours. We measured serum AST and ALT levels using AST and ALT kits (NjjcBio) following the manufacturer's protocol.

Bacterial treatment of BMMϕs in vitro and Western blot
We isolated bone marrow cells and cultured them with 10% L-929 cell culture medium as described previously [15]. We plated BMMϕs in 35 mm dishes and cultured them for 12 hours. We then added live or heat inactivated V. vulnificus at the indicated multiplicity of infection (MOI) or V. vulnificus lysates to the cells. We collected supernatants at the indicated times for cytokine quantification. We then subjected the infected cells to lysate preparation and Western blot analysis by following a previous protocol [41]. Anti-phosho-mTOR, antiphosho-S6K1, anti-phosho-S6, anti-phosho-Akt, anti-phosho-4EBP1, anti-phosho-IKKα/β, anti-phosho-IκBα, and anti-β actin antibodies all came from Cell Signaling Technology.

Measurement of cytokines by multiplex flow assay
We measured multi-cytokines by using a LEGENDplex multi-analyte flow assay kit (13-plex) for mouse inflammation panel (Biolegend). We assessed the cytokine concentrations in sera, liver homogenates, and supernatants of cultured cells according to the manufacturer's instructions. We collected data using a BD FACSAria II and analyzed them with LEGENDplex data analysis software (Biolegend).

BrdU incorporation
We intraperitoneally injected the mice with 1 mg BrdU (BD Bioscience) in 200 μl PBS, then with1×10 8 CFU of V. vulnificus two hours later. We harvested the liver from each infected mouse six hours after infection, isolating liver monocytes by following the previous protocol [42]. Briefly, livers were mashed in IMDM with 10% FBS. The upper cell suspension was filtered by nylon mesh and spun at 4˚C, 2000 rpm for 5 min. The cell pellet was resuspended in 12 ml 35% Percoll (GE Healthcare), and carefully underlaid with an equal volume of 75% Percoll, and centrifuged at 1000 × g for 20 min at room temperature without break. Cells at the interface were collected, washed and resupended in IMDM (10% FBS) for further staining. We stained the cell surface with CD11b and F4/80 (Biolegend) and used a BrdU Flow Kit (BD Bioscience) for BrdU intracellular staining, following the manufacturer's protocol.

Histological and immunofluorescence analysis
For histology analysis, we imbedded the fixed livers in paraffin, then cut thin sections, and following standard procedures, stained them with hematoxylin and eosin. We quantified the liver histological score of inflammation according to Ishak inflammation score [43,44]. For immunofluorescence analysis, we cut frozen sections into 8μM slices in a cryostat microtome (Thermo) at -20˚C and permeabilized and blocked them. We probed the tissues with appropriate, fluorescently labelled antibodies, including APC-conjugated anti-CD11b, PE-conjugated anti-CD68, and PE-conjugated anti-Ly6G (Biolegend). Finally, we dried the sections and observed them with a Nikon A1 confocal microscope.

Statistical analysis
We analyzed statistical significance using the Student t test and determined survival difference by log-rank survival analysis, performing all statistics using Graphpad Prism 5.0 software. P values are defined as follows: P < 0.05, P < 0.01, and P < 0.001. microscopy Core Facility of Wenzhou Medical University, and the department of Pathology in the first affiliated hospital of Wenzhou Medical University for their services.