All animal studies were carried out in strict accordance with the Guidelines for Animal Experimentation of the Japanese Association for Laboratory Animal Science (JALAS; Science Council of Japan, 2006). The animal experimentation protocol was approved by the president of Kitasato University based on the judgment of the Institutional Animal Care and Use Committee of Kitasato University (Approval No. 17-268).

Vibrio vulnificus CMCP6 and E4 was cultured aerobically in Luria-Bertani (LB) broth or on LB agar at 37°C. When required, the medium was supplemented with rifampicin (50 μg/ml) or ampicillin (100 μg/ml) for V. vulnificus.

The gene deletion mutant was constructed as described previously (Yamazaki et al., 2020). Briefly, the franking regions, 1,005 bp of upstream and the 1,040 bp of downstream, of rtxA gene were amplified with primers: RTX up Fw (5'-GGA TCC TTG CCT TTT AAT TGA GCT TGC-3'), RTX up Rev (5'-TCT AGA TCT CAT TTA CCC TGT GAA GAA GTA TTC AAC-3'), RTX down Fw (5'-TCT AGA TGC TAA CGC ATG TTG GTG ATG-3'), and RTX down Rev (5'-GCA TGC ATG AGT AAT GAT GTT GGC TTT-3'). The underline shows the recognition sites of the restriction enzymes, BamHI, XbaI, and sphI, respectively. The amplicons were confirmed by Sanger sequencing, then digested with BamHI-XbaI and XbaI-SphI, and cloned into the suicide vector pYAK1, retaining sacB. E. coli BWλpir were used as a conjugal donor to V. vulnificus CMCP6 with spontaneous rifampicin resistance. pYAK1-rtxAKO was introduced into V. vulnificus CMCP6. Vibrio vulnificus retaining pYAK1-rtxAKO was cultured in LB broth containing 20% sucrose following the standard sacB-assisted allelic exchange method. Mutants were confirmed by PCR to detect expected changes in size at the rtxA locus. The plasmid pXen-13 containing a bacterial luminescent gene cluster (luxCDABE) was transformed into V. vulnificus via electroporation. Vibrio vulnificus was grown in LB medium supplemented with 100 μg/ml ampicillin with agitation (163 rpm) at 37°C.

Five-week-old female C57BL/6 and BALB/c mice were purchased from Charles River Laboratories Japan (Atsugi, Japan). C57BL/6 mice and BALB/c mice were bred and maintained under specific pathogen–free conditions at Kitasato University. The mice were housed in plastic cages in a group and were maintained on a standard laboratory diet (rat chow MF, Oriental Yeast Co., Ltd. Tokyo, Japan) and tap water under a 12 h light and dark cycle. The ambient temperature during the study was maintained at about 21°C.

Thirty minutes prior to euthanasia, infected mice were injected intravenously (IV) with 0.1 ml of 5 mg/ml Evans blue dye (Wako, Osaka, Japan). Whole thighs of the sacrificed mice were transected at the ankle and at the hip joint, peeled, divided into SC tissue with skin and muscle tissue, and incubated at 60°C for 24 h in three volumes of formamide. Quantitative analysis of extracts containing Evans blue dye was determined by measuring optical density of blue at 600 nm (OD₆₀₀) with a microplate reader (Sunrise/TECAN Japan, Kanagawa, Japan; Nakamura et al., 2013; Radu and Chernoff, 2013).

Infection sites excised from sacrificed mice were demineralized by immersing them in buffer solution containing 0.2 M EDTA-4Na and 1% formalin for 1 week, fixed in 10% buffered formalin for 1 day, embedded in paraffin, sliced into 3-μm sections, and stained with Hematoxylin–eosin (H&E). For immunofluorescence staining, the paraffin-embedded sections were used. After deparaffinization, heat-based antigen retrieval and endogenous peroxidase blockade using 3% hydrogen peroxide were performed. Rabbit anti-neutrophil elastase (1:100, ab68672; Abcam) was added and incubated overnight at 4°C. Following a rinse in PBS, peroxidase (PO)-conjugated goat anti-rabbit IgG (Histofine Simple Stain Mouse MAX-PO, Nichirei Bioscience, Tokyo, Japan) was added and incubated for 1 h at room temperature. The site of the immunoreaction was visualized by the diaminobenzidine (DAB) reaction. Sections were counterstained with Mayer’s haematoxylin and mounted under coverslips. Image acquisition of the muscle tissue was performed using an inverted microscope (DM2500/Leica Microsystems, Tokyo, Japan) equipped with 5×/0.40, 10×/0.40, 20×/0.40, and 40×/0.40 objective lenses.

Vibrio vulnificus was grown in LB medium supplemented with 100 μg/ml of rifampicin with agitation (163 rpm) at 37°C. Overnight cultures (100 μl) were inoculated into 2 ml of fresh LB medium and incubated for 2 h. Bacteria were harvested, washed with PBS (pH 7.2) containing 0.1% gelatine, and resuspended in fresh LB medium. Then, 10⁶ colony-forming units (CFU)/mouse were SC inoculated in mice. Infected mice were sacrificed at 9 h post-infection. Whole-blood samples were collected in a syringe by cardiac puncture at multiple time points after infection and centrifuged at 1,200 g for 30 min, and sera were collected. Serum creatine kinase (CK), lactate dehydrogenase (LDH), and aspartate aminotransferase (AST) concentrations were evaluated using Dimension EXL with the LM Integrated Chemistry System (Siemens, Tokyo, Japan).

Vibrio vulnificus pXen-13 was grown in LB medium supplemented with 100 μg/ml ampicillin with agitation (163 rpm) at 37°C. Overnight cultures (100 μl) were inoculated into 2 ml of fresh LB medium supplemented with 100 μg/ml ampicillin and incubated for 2 h. Bacteria were harvested, washed with PBS (pH 7.2) containing 0.1% gelatine, and resuspended in fresh LB medium. Then, 10⁶ CFU/mouse were SC inoculated in BALB/c mice. Luminescence signals emanating from V. vulnificus were imaged at defined time points using an in vivo imaging system (IVIS) 200 imaging system (Xenogen/PerkinElmer, MA, United States) with a 1 min exposure time. The total photons emitted were acquired using the Living Image software package. Mice were anesthetized in chambers containing 2.0% isoflurane inhalant (Pfizer, Tokyo, Japan).

Vibrio vulnificus was grown in LB medium supplemented with 100 μg/ml of rifampicin with agitation (163 rpm) at 37°C. Overnight cultures (100 μl) were inoculated into 2 ml of fresh LB medium and incubated for 2 h. Bacteria were harvested, washed with PBS (pH 7.2) containing 0.1% gelatine, and resuspended in fresh LB medium. Then, the bacteria were SC, intramuscularly (IM), or intravenously (IV) inoculated in mice. Infected mice were carefully monitored and sacrificed at endpoints to measure survival time or at defined time points to collect tissues. The collected muscles or spleen were suspended in cold PBS containing 0.1% gelatine, homogenized for 5 s with a lab mixer IKA EUROSTAR digital (IKA Werke, Germany; 1,300 rpm), and centrifuged at 800 rpm for 5 min. The supernatants were plated at 10-fold serial dilutions in duplicate on LB agar containing 50 μg/ml rifampicin and incubated for 12 h at 37°C. Vibrio vulnificus colonies were counted, and bacterial burden was determined by calculating the number of CFU/g.

Statistical analysis was performed using GraphPad Prism (GraphPad Software, CA, United States). Statistical differences between the two groups were analyzed using the Mann–Whitney U test or Fisher’s exact test. Survival curves were analyzed using the log-rank test. A p value less than 0.05 was considered significant, and significance values are indicated as follow: ^*p < 0.05; ^**p < 0.01; and ^***p < 0.001.

Vibrio vulnificus NSTI is a fulminant infection that causes death within a few days in humans. The consistent assessment of the process of host death due to the pathogenicity of V. vulnificus is a prerequisite for an infection model. First, we investigated whether the wound infection model could assess the lethality of a clinical isolate strain. Clinical strain CMCP6 wild type (WT) and environmental strain E4 were SC inoculated into mice and observed for 72 h. Mice infected with CMCP6 WT caused dose-dependent death (Figure 1). Some mice infected with CMCP6 WT survived in the 10⁴ CFU/head-inoculation group (Figure 1). However, in the 10⁵ CFU/head-inoculation or higher, the survival time was shortened in a dose-dependent manner, and all mice died (Figure 1). On the other hand, mice infected with the environmental strain E4 were asymptomatic and did not die (Figure 1). These results indicate that our murine model can determine the dose-dependent lethality of clinical strains. Besides, some mice infected with 10⁶ CFU/head of CMCP6 ΔrtxA survived (Figure 1). Data showed that the murine model has the resistance against V. vulnificus proliferation depending on the number of infected bacteria and virulence factors.

Serum or plasma filling swells and blisters are known signs of V. vulnificus infection (Tacket et al., 1984; Klontz, 1988; Finkelstein et al., 2002; Ito et al., 2012; Oliver, 2015), and acute inflammation causes vasodilation and increased vascular permeability. In addition, it has been reported that a protease VvpE of V. vulnificus enhances vascular permeability in the guinea pig skin infection model (Jeong et al., 2000; Miyoshi, 2006). We examined the clinical signs and the risk factor during the V. vulnificus infection in our experiment, the thighs of mice infected with CMCP6 WT showed severe swelling. Therefore, we investigated whether the swelling was caused by the increased vascular permeability of Evans blue dye that binds to serum albumin immediately following its IV injection into the bloodstream. Under normal physiologic conditions, serum albumin does not cross the vascular endothelial cells, and the Evans blue dye remains restricted within blood vessels. Therefore, only conditions that promote increased vascular permeability allow for extravasation of Evans blue dye at the local tissues. We injected the Evans blue dye into the tail vein of the mice 30 min before harvesting the tissue for observation and measurement of the Evans blue dye extravasation. The time-course analysis by measurement of the leakage of Evans blue dye showed that increased vascular permeability began at 2 h post-infection and peaked at 4 h post-infection in the skin and SC tissue (Figure 2A). In the muscle, it gradually increased and peaked at 10 h post-infection (Figure 2B). Consistently with these results, leakage of Evans blue dye was observed both in the SC tissue and superficial muscle of the local infection site at 6 h post-infection, but not at 10 h post-infection (Figures 2A–D). These results suggested that the Evans blue dye leaked at the inside of deep muscle tissues. Thus, we clearly show that vascular permeability increases in the SC tissue in the early and deeper muscle tissue in the late stages of infection.

With inflammation, increased vascular permeability promotes the effusion of cellular components and plasma, such as albumin, immunoglobulins, and fibrinogen (Abtin et al., 2014; Pasparakis et al., 2014). In addition, vasodilation slows blood flow and causes congestion. As a result of the filling of red blood cells by the congestion, leukocytes are leaked. Especially, neutrophils infiltrate the site of infection to eliminate bacteria (Abtin et al., 2014; Pasparakis et al., 2014). In our experiment, histopathological analysis of SC tissue was performed to clarify the details of increased vascular permeability and the pathology dynamics during the infection. The congestion and red blood cells were not found immediately after infection at 0 h post-infection (Figure 3A), but the blood vessels were filled with red blood cells and became congested as the infection progresses (Figure 3A). Finally, red blood cells and leukocytes were decreased at the site of infection of end-stage mice at 12 h post-infection (Figure 3A). During the progression of this congestion, leukocytes appeared around the blood vessels at 3 h post-infection and peaked at 6 h, and then these were widely dispersed (Figure 3A). The leukocytes were positive for anti-neutrophil elastase antibody by the immunostaining from 3 to 12 h post-infection (Figure 3B). In comparison, no leukocytes were positive at 0 h post-infection (Figure 3B). Neutrophils, of which several parts were activated by the infection and labeled with anti-elastase, showed up at 3 h post-infection. They were increased, peaked at 6 h post-infection, and decreased after that in the SC tissue (Figure 3B). Thus, elastase release by infiltrated neutrophils at the infection site was shown to begin in the initial stage of infection. These aspects are known as cellulitis in V. vulnificus infection.

In NSTIs, necrosis is an irreversible pathological condition, occurred in fascia and muscle, and it progressed dramatically. In this study, the increased vascular permeability occurred along with the infiltration of neutrophils when blood vessels became congested (Figures 2A, 3B). Since the increased vascular permeability increased in the SC tissue, peaking in the muscle tissue, it was predicted that the increase resulted in pathological changes to muscle tissue. To investigate these, we performed a histopathological analysis. We observed the destruction of the fascial plane at the boundary between the SC tissue and the muscle, and muscle necrosis as reflected by enucleation and reduced dyeability with eosin at 6 h post-infection (Figure 4A). Moreover, these necrotic lesions (changes) of muscle tissue expanded from the surface to the deep muscle after that at 9–12 h post-infection (Figure 4A). Infiltration of neutrophils was also observed in muscle tissue at 9 h post-infection (Figure 4A).

We evaluated the severity of muscle necrosis by a biochemical test measuring levels of CK, LDH, and AST in the serum. These markers are released into the circulation when muscles are damaged. Levels of all markers began to increase at 6 h post-infection (Figure 4B). These results were consistent with the onset of muscle necrosis observed by histopathological analysis. After that, the levels increased in a time-dependent manner. These data, obtained from the murine wound infection model, could be used to help explain the fulminant progression of V. vulnificus NSTIs in patients.

Our multilayer approach showed that clinical signs, such as cellulitis and necrotic fasciitis, emerge in the SC tissue in the early and evolve in deeper muscle tissue in the late stages of infection. We assess the proliferation associated with these clinical signs of CMCP6 WT and avirulent strains. To assess infection dynamics at the primary infection site, we infected mice with CMCP6 WT, ΔrtxA, and E4 expressing a luciferase operon and observed by IVIS (Figure 5A). The luminescence signals of CMCP6 WT were the most extensive at 6 h post-infection and were concentrated in the soft tissue of the thigh. On the other hand, CMCP6 ΔrtxA signals disappeared from 6 to 9 h post-infection, and E4 signals disappeared from 3 to 6 h post-infection from the soft tissue of the thigh (Figure 2A). These results indicate that CMCP6 WT colonizes and proliferates in the primary infection site, but CMCP6 ΔrtxA and E4 could not. Next, since the necrosis emerged in deeper muscle tissue in the late stages of infection, we skinned the mice to assess bacterial invasion into the muscle and observed by IVIS (Figure 5B). The luminescence signals of CMCP6 WT were extensive at 12 h post-infection in the muscle (Figure 5B). On the other hand, CMCP6 ΔrtxA signals completely disappeared at 9 h post-infection in the soft tissue, but some were detected from the muscle at 12 h post-infection (Figure 5B). We found that it was the CMCP6 WT strain infection in the skin and SC tissue that had spread, causing cellulitis and necrotizing fasciitis. Later, the WT strain invaded and proliferated in the muscle causing muscle necrosis. Almost all ΔrtxA strains were less virulent and infections were cleared from the skin and SC layers, suggesting that RtxA is essential for immune evasion and proliferation in the SC tissue.

In vivo imaging system analysis showed that V. vulnificus invades and proliferate the muscle tissue. We calculated the bacterial burden of the muscle beneath the infection site. CMCP6 WT were detected immediately after infection, exponentially increased until 6 h post-infection, and continued to increase at 12 h post-infection in the primary infection site (Figure 6A). Since the progression of clinical signs and proliferation of V. vulnificus in the muscle tissue were suggesting the fatal condition of the host, we investigate the correlation of bacterial proliferation between the local infection site and systemic circulation, an index of sepsis of the host, and lethality of V. vulnificus, we inoculated the 10⁶ CFU/head of CMCP6 WT or ΔrtxA and calculated the bacterial burden of the muscle and the spleen. The number of bacteria collected from the muscle of CMCP6 WT-infected mice was finally 4.20 × 10⁸ CFU/g (median) at 12 h post-infection when the mice began to die (Figures 1, 2C). The number of bacteria collected from the spleen of CMCP6 WT-infected mice was 2.33 × 10⁶ CFU/g (median; Figure 6B). At the same time point, the number of bacteria collected from the muscle of mice infected with CMCP6 ΔrtxA was 4.87 × 10⁵ CFU/g (median) and significantly lower than that of CMCP6 WT (Figure 6B). Mainly, CMCP6 ΔrtxA was frequently not detected from the spleen in the infected mice (n = 3 of 8), and the number of bacteria collected from the spleen was 4.59 × 10² CFU/g (median; Figure 6B). The correlation coefficient between the number of bacteria in muscle and the number of bacteria in the spleen was 0.8097 (Figure 6B). This strong positive correlation proved that the proliferation in the muscle was essential for increasing the number of bacteria in the systemic circulation, which is known as a representative aspect of septicemia in V. vulnificus infection, ultimately leading to host death.

We compared survival times of mice receiving SC, intramuscular (IM), and IV injections. IM-inoculated mice survived for the shortest amount of time, followed by SC- and IV-inoculated mice (Figure 7A). Similarly, we compared the survival times of mice receiving SC and IM injections of ΔrtxA. Mice IM-inoculated with 10⁶ CFU/head of ΔrtxA showed higher mortality than SC-inoculated mice (Figure 7B). The high dose of ΔrtxA (infection with 10⁷ CFU/head) seemed to have been the cause of death in mice via the SC routes of infection (Figure 7B), although the survival time was delayed compared with mice infected with WT (10⁶ CFU/head; Figures 7A,B). However, mice IM-inoculated with 10⁷ CFU/head of ΔrtxA died as early as did mice infected with WT (Figures 7A,B). Thus, allowing the invasion into the muscle by V. vulnificus, even attenuated strains, is at a potent risk for the fatal outcome. The high bacterial burden in muscle led to a fatal outcome caused by V. vulnificus infection.