Vibrio vulnicus Hemolysin Induces TNF-α Independent ROS Production by Primary Murine Macrophages

Background Vibrio vulnicus (V. vulnicus) is a gram-negative opportunistic pathogen that causes lethal infections in humans. Vibrio vulnicus hemolysin (VVH) is a key virulence factor that exhibits strong hemolytic and cytolytic activities and contributes to the pathogen's invasion, vasodilatation, and septic shock. Most of the studies so far have focused on VVH's cytolytic activity against cell lines derived from host cells. However, the cytolytic activity of VVH on primary macrophages is still unclear. In addition, although it is known that VVH induces host cell apoptosis via triggering ROS production, the impact of VVH on host immune response has not been fully understood. This study aimed to investigate the role of VVH-induced TNF-α expression and ROS production in the absence of apoptosis of murine primary macrophages and related signaling pathways using FACS, DCFH-DA, real-time PCR, and western blotting. The results showed that murine primary macrophages from different organs displayed differential sensitivities towards VVH-induced cell death. Liver Kupffer cells, splenic macrophages, and BMMfs were more sensitive to VVH-induced cytotoxicity, while alveolar macrophages, lung interstitial macrophages, and lung neutrophils were resistant to VVH-induced cell death. Besides, we found that a low dose of VVH, which did not induce apoptosis in murine primary macrophages, could induce apoptosis independent TNF-α expression and ROS generation. Such ROS production in macrophages could be further blocked by inhibiting p38-MAPKs or NFκB activation but was not affected by knockout of TNF-α. macrophages, an apoptosis-independent TNF-α expression, and ROS production, provides insight into the mechanism underlying crosstalk between VVH-induced inammation


Introduction
Vibrio vulni cus (V. vulni cus) is a halophilic, gram-negative marine bacterium [1,2]. As an opportunistic human pathogen that exsits in the envrioment, such as warm sea water. V. vulni cus can cause severe life-threatening food poisoning, wound infection, and primary septicemia, especially in patients with hepatitis, hereditary hemochromatosis (HCC), or compromised immune system [3,4]. Also, V. vulni cus infection can cause gastroenteritis, pneumonia, and keratitis. Infections can result from exposing an open wound to contaminated seawater or consuming raw seafood contaminated with V. vulni cus, such as oysters. V. vulni cus infection's mortality usually exceeds 50%. It can increase to more than 90% in patients who develop sepsis and sepsis-induced organ failure [1]. This hyper-in ammatory response is characterized by massive increases in reactive oxygen species (ROS), nitric oxide (NO), and in ammatory cytokines [5].
V. vulni cus hemolysin (VVH) is a hemolytic and cytotoxic exotoxin that contributes to the invasion of bacteria into the bloodstream [6,7]. It has been reported that VVH secretion during V. vulni cus infection may be involved in increasing vascular permeability, mobilizing P-selectin to the pulmonary endothelial cell surface for adhering to neutrophils, and subsequently causing lung damage [6,8,9]. Most of the studies have been focused on the cytotoxic function of VVH, but its immunological functions have not been well elucidated. An early study suggested several clinical strains of V. vulni cus induced apoptosis in human and murine macrophage cell lines [10][11][12]. However, the mechanism and the possibility of V. vulni cus-induced death in primary macrophages are unclear. VVH can induce necrosis and apoptosis of multiple cell lineages, such as macrophages and epithelial cells, via different mechanisms [13][14][15]. It can also induce the continuous production of serum pro-in ammatory cytokines in animals and humans, such as TNF-α and interleukin 10 (IL-10) [16].
ROS plays a signi cant role in the pathophysiology of bacterial sepsis [17,18]. Excessive ROS production is harmful to the host cells [19]. However, low ROS levels are also important for cell functioning, particularly for immune cell function against infection [20]. ROS can promote pathogen elimination by direct oxidative damage or by a variety of host defense mechanisms through the immune cells. Immune cells depend on ROS to kill the phagocytosed microorganisms and mediate in ammatory and cellular signaling cascades [21,22]. Tumor necrosis factor α (TNF-α) is a well-known pro-in ammatory cytokine, important for immunity and cellular homeostasis. Numerous studies have shown that TNF-α and ROS are interconnected, and both play important roles in necrosis and apoptosis. ROS can be induced by cytokines and stimulate pro-in ammatory cytokine production by activating NFκB signaling [23]. Thus, the crosstalk between TNF-α and ROS in in ammation appears complicated and not fully understood. It is known that VVH can induce apoptosis via ROS production in several host cells [13-15, 24, 25].
However, it is unclear whether VVH can induce ROS production and cytokine production in the absence of apoptosis in the innate immune cells. In this study, we investigated the role of VVH-induced TNF-α expression and ROS production in the absence of apoptosis in murine primary macrophages and related signaling pathways.

Materials And Methods
Mice and bacterial strains C57BL/6 mice were purchased from the Wenzhou Medical University Laboratory Animal Facility. TNF-α −/− mice were generated from the Jackson Laboratory and purchased from the National Animal Resource Information Platform of China. Mice were maintained under an SPF environment at the Central Animal Laboratory of the Wenzhou Medical University. L-929 cells were purchased from the Cell Bank of the Chinese Academy of Science in Shanghai and cultured in RPMI-1640 containing 10% heat-inactivated fetal bovine serum (Ausvin) and penicillin-streptomycin (50 IU/mL and 50 mg/mL; Beyotime). V. vulni cus CGMCC 1.1758 strain was provided by the China General Microbiological Culture Collection Center. V. vulni cus was cultured at 37°C in the brain heart infusion broth (BHI) or on the BHI rabbit blood agar plate. E. coli BL21 (DE3) was provided by the Key Lab of Laboratory Medicine, Ministry of Education of China. The pMD18T-VvhA plasmid was kindly provided by Dr. Jie Yan (School of Medicine, Zhejiang University, China).
The heat-inactivation of VVH was performed by incubating VVH at 65°C for 10 min. One hemolytic unit (HU) level is de ned as the amount of VVH present when OD545 RBC suspension (VVH treated)/OD545 RBC suspension = 0.5.
Isolation of total lung single-cell suspension, splenocytes, and liver monocytes Murine lung was removed from mice with perfusion with 1× PBS (twice), cut into small pieces, incubated in RPMI 1640 containing 20 µg/mL DNase (Beyotime) and 30 µg/mL collagenase (Sigma) for 40 min at 37°C, and passed through a 200 mesh nylon cell lter. Then the cells were washed with RPMI 1640 containing 10% FBS two times. The erythrocytes were lysed with Ammonium-Chloride-Potassium (ACK) lysis buffer. The isolation of liver monocytes and splenocytes was performed as previously described [27].

Generation of bone marrow-derived macrophages (BMMφs)
Bone marrow cells were isolated from femurs and tibiae and then cultured with 10% L-929 cell culture medium (conditioned medium), which was described previously. After 3 days of culture at 37°C in the CO 2 incubator, non-adherent cells were collected and transferred into a new petri dish with a fresh conditioned medium for another 4 d to grow in culture before used for the experiment. The purity of F4/80 + CD11b + cells was analyzed by ow cytometry. More than 90% of the cells were F4/80 + CD11b + cells.

Stimulation of cells with VVH in vitro
A total of 1.0 × 10 6 /well total lung cell, splenocytes, or liver monocytes were equally seeded in 48-well plates in a 0.5 mL RPMI-1640 medium containing 10% FBS at various concentrations VVH from 0 to 0.75 HU/mL for 6 h at 37°C in the CO 2 incubator.
For stimulation of BMMφ with VVH, a total of 0.5 × 10 6 /well BMMφ were equally seeded in 48-well plates in 0.5 mL RPMI-1640 medium containing 10% FBS. After overnight incubation at 37°C in the CO 2 incubator, BMMφs were treated with VVH at the indicated concentration and time. For the inhibitor suppression assay, the cells were left without inhibitors (VVH only), with 10 M SB203580 (p38 inhibitor) plus 0.25 HU/mL VVH at 37 ℃ for 6 h, or with 10M BAY11-7082 (NFB inhibitor) plus 0.25 HU/mL at 37 ℃ for 6 h.

Flow cytometry analysis
Flow cytometry analysis was performed as previously described [28]. Cells in single-cell suspension were brie y incubated at 4°C with PBS containing 2% FBS (staining buffer) and uorescent-conjugated antibodies. After 30 min of incubation, the cells were washed at least two times with staining buffer and resuspended in PBS containing 2% paraformaldehyde. Intracellular staining of cytokines was performed using xation and permeabilization buffer (eBioscience). All ow cytometry data were collected using the BD FACS Aria (BD Biosciences) and analyzed with the FlowJo software (Treestar).

Measurement of cellular ROS
DCFH-DA (Beyotime) was used to detect cellular ROS production. DCFH-DA became DCF with strong uorescence once oxidized by cellular ROS. VVH treated cells were loaded with DCFH-DA probes according to the manufacture's protocol. After VVH treatment, cells were brie y incubated with DCFH-DA at 37°C for 30 min in the CO 2 incubator and later stained with other ow antibodies.

Real-time PCR
At each concentration of VVH, RNA was extracted from BMMφs using TRIzol reagent (Invitrogen). cDNA was synthesized using SuperScript III and random primers, and quantitative real time-PCR was performed as described previously. The expression of mRNAs was normalized with β-actin and calculated using the 2 −ΔΔCT method. Primers used include TNF-α (forward, 5 -TGTCTCAGCCTCTTCTCATT-3 ; reverse, 5 -AGATGATCTGAGTGTGAGGG-3 ).

Results
Production and puri cation of VVH We rst generated prokaryotic VVH expression vector pET28a-VvhA for high-level VVH expression in E.coli BL21 (DE3). The recombinant VVH was expressed in an insoluble form within inclusion bodies (Fig. 1A).
High purity VVH was obtained after NTA-Ni 2+ column a nity puri cation, re ected in a single protein band having an expected molecular weight of 54 kDa on an SDS-PAGE gel (Fig. 1B). The puri ed VVH was renatured, and it showed a single band of 54 kDa by Coomassie staining (Fig. 1C) and western blot with an anti-His Tag antibody (Fig. 1D). Refolded VVH demonstrated hemolytic activity against rabbit erythrocytes in a dose-dependent manner (Fig. 1E). After heat inactivation at 65 ℃ for 10 min, VVH lost the hemolytic activity, supporting the assumption that the hemolytic activity was due to VVH.

Analysis of cell death in murine macrophages from various organs under VVH treatment
To investigate the cytotoxicity of VVH towards primary macrophages from different tissues, murine lung single-cell suspension, liver mononuclear cells (MNCs), splenocytes, and bone marrow-derived macrophages (BMMφs) were treated with 0.25, 0.50, or 0.75 HU/mL VVH for 6 h followed by Annexin V or 7-AAD staining to detect cell death. Macrophages of different origins displayed different sensitivities to VVH-induced death. Compared to splenic macrophages, Kupffer cells had a much lower basal death rate but displayed a much higher death rate following VVH treatment ( Fig. 2A and 2B). Previous studies showed that VVH could induce murine lung vascular structures' dysfunction by increasing vascular permeability and neutrophil sequestration [29]. Although a majority of the alveolar Mφ (AM) and interstitial Mφ (IM) died during in vitro culture without VVH treatment, VVH treatment did not induce additional cell death. Moreover, VVH did not display obvious cytotoxicity to neutrophils ( Fig. 2C and 2D).
Low dose VVH triggers reactive oxygen species production without triggering cell death Similar to the liver and splenic myeloid cells, CD11b + F4/80 + BMMφs were sensitive to high dose (> 0.25 HU/mL) VVH induced death, but cells were resistant to low dose (0.25 HU/mL) VVH treatment ( Fig. 3A  and 3B). As a low dose is more relevant to the early phase of infection, we sought to investigate BMMφs response to treatment with a low dose of VVH. As ROS production is associated with VVH-induced cell death, we rst investigated whether low dose VVH treatment induces ROS production activity. To our surprise, we detected elevated ROS levels in BMMφs treated with 0.25 HU/mL VVH using DCF staining and ow cytometry (Fig. 3C). Besides, in contrast to activated VVH, the heat-inactivated VVH treatment also showed the ROS production was slightly increased, indicating that VVH cytotoxicity's independence was probably involved in VVH-triggered ROS production.
To investigate how VVH induces ROS production in BMMφs, we examined MAPK and NFκB signaling pathways, two important pathways for ROS production. As shown in (Fig. 3D), 0.25 HU/mL VVH treatment-induced IKKβ, NFκB, and p38-MAPK, but not ERK1/2 or AKT phosphorylation in BMMφs. Moreover, inhibition of p38 with a chemical inhibitor SB203580 abolished VVH-induced ROS production, while NFκB inhibitor BAY 11-7082 partially inhibited VVH-induced ROS production (Fig. 3E). These data suggest that low dose VVH can induce cellular oxidative stress by inducing ROS production through activation of p38-MAPK and NFκB pathways.

Low dose VVH induces pro-in ammatory cytokine production in murine BMMφs
To investigate whether low dose VVH treatment can induce pro-in ammatory cytokine production, we focused on TNF-α. We observed that 0.25 HU/mL VVH did not cause the death of primary murine BMMφs, but induction of TNF-α production did, which was detected by intracellular staining. The frequency of TNF-α + BMMφs was 18.8% in the 0.25 HU/mL VVH treatment group, 5.64% in the inactivated 0.25 HU/mL VVH treatment group, and 1.3% in the PBS control ( Fig. 4A and 4B). Consistent with the ow cytometry data, the TNF-α mRNA expression was also upregulated in murine BMMφs (Fig. 4C). These data indicate that low dose VVH can promote TNF-α production via inducing TNF-α transcription, which may trigger in ammatory responses at a very early stage of V. vulni cus infection.

ROS generation independent of TNF-α in BMMφs upon VVH treatment
Our data show that low dose VVH induces TNF-α production and ROS generation in BMMφs. As TNF-α has been reported to induce ROS production in macrophages [30], VVH-induced ROS generation could be indirectly caused by TNF-α. To determine the role of TNF-α in VVH-induced ROS production, we treated WT and TNFα −/− BMMφs with 0.25 HU/mL VVH. TNFα −/− mice did not produce TNFα (Fig. 5A). TNFα −/− BMMφs produced ROS levels similar to WT controls (Fig. 5B). These data indicate that VVH induces ROS independent of TNF-α.

Discussion
It has been reported that several clinical strains of V. vulni cus induce apoptosis in human and murine macrophage cell lines [10,31]. In this study, we showed that VVH could induce the death of murine primary macrophages. We further showed that murine macrophages from different organs display differential sensitivities to VVH-induced cell death. Liver Kupffer cells, splenic macrophages, and BMMφs are more sensitive to the cytotoxicity of VVH, while alveolar macrophages, lung interstitial macrophages, and lung neutrophils are resistant to VVH-induced cell death. The mechanisms that dictate the sensitivities of these cells towards VVH are unclear. One possibility is that VVH may partially inhibit innate immune responses in the spleen and liver by inducing macrophage death in these organs at an early stage of V. vulni cus infection.
To its hemolytic and cytolytic activities, the effects of VVH on immune cells have not been well de ned.
Previous studies have shown that pathogens' hemolysin could affect the host cells by triggering various cellular events, such as cell death or ROS production. Streptolysin O (SLO) triggers cell death through apoptosis in macrophages and neutrophils [32], and SLO also activates p38 via ASK1 and ROS [33]. Like SLO, Listeriolysin O(LLO)also triggers apoptosis in dendritic cells and T lymphocytes [34]. Moreover, nitric oxide (NO) production depends on high doses of pneumolysin (PLY) treatment and phagocytosis. However, the low dose of PLY mainly induces rapid ROS production [35]. Other hemolysins, such as hemolysin II (HlyII) of Bacillus cereus, have also been reported could induce apoptosis in host monocytes and macrophages in vivo [36]. During phagocytosis, the phagocytes internalized the microbes, such as L. interrogans that produce rSph2 hemolysin, which could directly target the mitochondria, and trigger cell apoptosis by induction of ROS and mitochondrial membrane damage [37]. This study found that the low dose of VVH could induce signi cant ROS production even without triggering apoptosis. To date, only the high dose of VVH has been reported to induce ROS production in epithelial cells and endothelial cells upon apoptosis [14,15,24,38,39]. The high dose of VVH induced cell death and ROS production through phosphorylation of a distinct ERK1/2 kinase [15,25], and ERK1/2 did not respond to treatment with the low dose of VVH in murine BMMφs. We found that the low dose of VVH treatment elevated phosphorylation of p38, Akt, NFκB, and IKKα/β in murine BMMφs, and that low dose VVH triggered both p38-MAPKs-and NFκB-dependent ROS generation. Thus, the mechanisms of ROS production are likely different between low dose of VVH and high dose of VVH treatment.
TNF-α and IL-1 are two important mediators leading to sepsis. Previous studies have reported that VVH can induce the release of IL-1β, but not TNF-α in macrophages [40]. In this study, we found that a low dose of VVH can directly induce TNF-α expression. Our result is in line with a previous observation that VVH treatment increases TNF-α and IL-10 levels in murine serum [16]. For other pathogens, previous reports show that Leptospiral hemolysin induces TNF-α, IL-1β, and IL-6 levels in murine macrophages via toll-like receptor 2 (TLR2)-or TLR4-mediated JNK and NFκB pathways [41]. Sub-cytocidal concentrations of α-hemolysin from E. coli could lead to increased IL-1β production, but not that of TNF-α in monocytes [42]. Another hemolysin study reported that it was associated with triggering in ammation. SLO inhibited TNF-α and IL-1β release from infected macrophages to blunt macrophages' immune response [32]. LLO causes IL-1β production by activating NLRP3 in ammasome in macrophages [43] and inducing cytokine gene expression, including IL-1, TNF-α, IFN-γ, and IL-12 [44]. However, the expression of IL-1β, IL-6, and IL-8 was not affected by PLY. Only the TNF-α level was enhanced by PLY in a phagocytosis-independent manner [35]. Vibrio parahaemolyticus plays a major role in triggering NLRP3 in ammasome activation with IL-1β secretion in macrophages [45]. Thus, in ammation cytokines production by VVH appears to be different from that of other hemolysins.
The results of this study also provide additional insights into the relationship between TNF-α and ROS. Both TNF-α and ROS are crucial for in ammation, but the crosstalk between ROS and TNF-α during Infection is still not fully understood. The lectin-like domain of TNF-α has been reported to decrease LLOinduced Nox4 mRNA expression and ROS generation [46]. TNF-α activates NFκB signaling pathway in apoptotic cells and subsequently inhibits intracellular ROS level [23]. However, ROS generation is also induced by cytokines. In phagocytes, activation of the TNF-NOX2 signaling leads to ROS production responsible for in ammation and associated tissue damage [47]. Pore-forming toxin from Serratia marcescens can cause TNF-α-dependent necroptosis and facilitate ROS generation [48]. Thus, TNF is also important for ROS production. In this study, we showed that knockout TNF-α in murine macrophages does not affect ROS generation by low dose VVH treatment, which differs from the classic interconnection between TNF and ROS. Future work will focus on the mechanism underlying VVHinduced in ammation and oxidation stress. CONSENT FOR PUBLICATION Not applicable.

AVAILABILITY OF DATA AND MATERIALS
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

COMPETING INTERESTS
All the authors declare there are no competing interests.   VVH-induced ROS production is independent of TNF-α (A) Absence of TNF-α in TNF-α de cient murine BMM∅s following 5 ng/mL LPS stimulation presence of 1 ng/mL GolgiPlug for 4 h in vitro. FACS plots show intracellular TNF-α staining in live-gated CD11b+ BMM∅s. (B) TNF-α is not essential for VVHinduced ROS generation in BMM∅s. WT and TNFα de cient BMM∅s were similarly stimulated, and examined as above in Figure 4C. Data shown are representative of at least three experiments.