Care and treatment of the animals were based on the Guide for the Care and Use of Laboratory Animals (27). Specific-pathogen-free (SPF) male and female, Balb/C, C57BL/6 wild-type (Wt), and C57BL/6-TLR2-deficient (TLR2^−/−) (8–10 weeks old) mice were housed under SPF conditions at the Animal Facility of the Department of Microbiology, Immunology and Parasitology from Federal University of Santa Catarina (UFSC). A total of 641 mice were used in this study. C57BL/6 Wt and Balb/C mice were purchased from Jackson Laboratory. TLR2^−/− mice generated by Dr. Shizuo Akira (Osaka University, Japan) (28) were a kind gift from Ricardo T. Gazzinelli (Federal University of Minas Gerais, Brazil). Mice were housed in cages at 21 ± 2°C with free access to water and food. All animal experiments were approved by the Animal Welfare Committee of the UFSC (PP00880).

Acute pancreatitis was induced by BDO, as described (26). Briefly, mice were anesthetized with xylazine (2 mg/kg, i.p.) followed by isoflurane (3–5 vol%). A 1-cm midline incision was made in the anterior abdomen, and the biliopancreatic duct was exposed at its orifice in the duodenal wall. The duct was dissected, and a tourniquet consisting of a cotton sewing and 2 mm of a P10 cannula was looped around the biliopancreatic duct adjacent to the duodenal wall. In the sham-operated (Sham) animals, the biliopancreatic duct was dissected but not obstructed. All animals received 1 mL of saline subcutaneously immediately after surgery. The obstructive model employed in the present study mimics the mechanical obstruction of the bile and pancreatitic ducts induced by biliar gallstones in AP patients, one of the main etiology of AP (6). The obstructive model of AP leads to disease development by self-mediators, which is a better representation of the natural course of the disease when compared to other models of severe AP, when chemical agents are intraductally administrated (24, 25).

Colorimetric assay kits were used to determine the serum activity of alanine aminotransferase, aspartate aminotransferase, amylase, lipase, alkaline phosphatase, and lactate dehydrogenase and the serum concentrations of bilirubin, urea, and creatinine (Labtest Diagnóstica, Brazil).

Mice were pretreated or treated intraperitoneally (i.p.) with saline or meropenem (Eurofarma, SP, Brazil, 100 mg/kg, 12/12 h) for 3 days and then subjected to surgery. The surgical procedure was performed 12 h after the last dose in the pretreated groups, and posttreatment started 12 h after surgery. Mice were also pretreated with meropenem for 3 or 5 days (30 or 100 mg/kg, 12/12 h) to evaluate the tissue injury induced by the antibiotic treatment.

Mice were pretreated i.p. (12/12 h) with saline or clindamycin (Hipolabor, MG, Brazil, 25 mg/kg) plus ceftriaxone (União Química, SP, Brazil, 30 mg/kg) for 3 days and then subjected to surgery.

To reduce the intestinal bacterial community, mice were treated with a cocktail of broad-spectrum antibiotics, as described previously (29). Animals were first exposed to amphotericin-B (Sigma-Aldrich, St. Louis, MO, USA, 1 mg/kg body weight, BW) delivered by gavage during 3 days, 12/12 h. From days 3 to 5, ampicillin (Sigma-Aldrich, 1 mg/mL) was added into the drinking water, and the mice were gavaged 12/12 h with a cocktail of antibiotics consisting of vancomycin (Sigma-Aldrich, 50 mg/kg BW), neomycin (Sigma-Aldrich, 100 mg/kg BW), metronidazole (Sigma-Aldrich, 100 mg/kg BW), and amphotericin-B (1 mg/kg BW). Additionally, from days 3 to 15, mice were i.p.-treated daily, with ciprofloxacin (Sigma-Aldrich, 50 mg/kg BW). From days 16 to 27, mice were gavaged daily with 100 µL (100 mg/mL) of a feces suspension or with 100 µL (1 × 10⁸ colony-forming units, CFUs) of an E. gallinarum suspension. Fresh feces were collected from naïve mice, homogenized in sterile PBS and decanted for 30 min at RT. Animals were then exposed to supernatants of the feces suspension. E. gallinarum was cultivated in brain–heart infusion liquid medium at 37°C until an optical density of 0.60 at 600 nm was reached, and the bacteria were washed three times with PBS prior to animal administration. On day 28, mice were subjected to Sham or AP induction.

The bacterial count was determined as previously described (30). Briefly, samples were collected under sterile conditions, and 10 µL of serial dilutions were seeded on agar dishes containing Muller-Hinton, blood agar Muller-Hinton, or bile esculin azide agar with 6 µg/mL of vancomycin (Difco Laboratories, Detroit, MI, USA) and grown at 37°C aerobically for 24 h. The lumina of the intestinal samples were washed with PBS for intestinal CFU quantification. The results were expressed as the mean ± SEM log of CFU per milliliter, grams of feces, or milligrams of tissue.

We identified bacterial species using biochemical and microbiological methods, as described (31) and using the Vitek2 automated system (bioMérieux, see Tables S1–S8 in Supplementary Material for details regarding the bacterial identification).

Individual colonies of E. coli and E. gallinarum were resuspended in sterile PBS and boiled for 10 min. These samples were used for PCR analysis with the following primers for the vanC1 gene: F_5′GGTATCAAGGAAACCTC3′ and R_5′CTTCCGCCATCATAGCT3′ (32). PCR was performed on a Mastercycler Gradient™ thermocycler (Eppendorf, Germany) with the following conditions: 95°C (10 min) followed by 35 cycles of 95°C (1 min), annealing at 54°C (30 s), 72°C (33 s), with a final extension step at 72°C (5 min). The concentration of magnesium used was 2.0 mM.

Enterococcus gallinarum susceptibility to vancomycin, clindamycin, streptomycin, ampicillin, and meropenem was determined using a disk diffusion or Etest assay. E. gallinarum was grown in brain–heart infusion liquid medium at 37°C until an optical density of 0.60 at 600 nm was reached, and aliquots (including a 0.5 McFarland standard) of the cultures were spread onto Muller-Hinton agar plates; 6-mm Whatman disks loaded with the appropriate antibiotic were added to the agar. Plates were incubated overnight at 37°C, and the zone of bacterial growth inhibition was measured.

Enterococcus gallinarum (FMC012174) was isolated from the blood of the AP mice that were pre-exposed to meropenem. The clinical strain of E. gallinarum (UH-FMC001) was isolated from the blood of a patient with sepsis and was a gift from the Microbiology Laboratory, University Hospital, UFSC. The E. gallinarum strains were grown in brain–heart infusion liquid medium at 37°C until an optical density of 0.60 at 600 nm was reached, and the strains were washed three times and resuspended with PBS. Six hours later, the AP induction mice were intravenously (i.v.) treated with 100 µL of saline, 100 µL of 1 × 10⁸ CFU of live E. gallinarum or 100 µL of 1 × 10⁸ CFU of heat-killed E. gallinarum. Samples of E. gallinarum were boiled for 15 min, and no bacterial growth was observed in these samples on the blood agar dishes (37°C, 24 h). In another experimental set, Wt or TLR2^−/− naïve mice were i.p.-treated with 200 µL of 1 × 10⁹ CFU of the live mouse or clinical strains of E. gallinarum.

Mice were euthanized 24 h after surgery, and the pancreases were fixed by immersion in 5% paraformaldehyde, dehydrated, and embedded in paraffin wax. Three-micron-thick sections were stained with hematoxylin and eosin for histological examination. The scoring criteria for edema, inflammatory cell infiltration, and necrosis were as follows: 0, absent; 1, mild (5–10% of tissue); 2, moderate (10–30% of tissue); and 3, severe (30% of the tissue). The mean ± SEM represented the sum of scores (range from 0 to 9) of the overall damage to the pancreas. Blind histological assessments were performed by a pathologist.

Leukocyte migration was assessed 1, 4, and 24 h after surgery, as described (33).

Lungs were perfused with PBS/EDTA (1 mM) before harvest from mice. Samples were collected 16 h after the last treatment. Lungs were passed through 40-µm nylon cell strainers and single cell suspensions were centrifuged in 35% Percoll^® solution (315 mOsm/kg, Sigma-Aldrich) for 15 min at 700 × g to enrich leukocytes populations. Pelleted cells were then collected, and erythrocytes were lysed. Single cell suspensions from individual mice were stained for flow cytometry analysis, as described (34). Cells were stained with anti-mouse antibodies (BD Pharmingen) to GR1 (clone RB6-8C5), CD11b (clone M1/70) and F4/80 (clone BM8). Data were collected using a FACSCanto II (BD Immunocytometry Systems) and analyzed using FlowJo software.

Data are reported as the means ± SEM of the values obtained from three independent experiments. The mean values for the different groups were compared by analysis of variance (ANOVA). If significance was determined, individual comparisons were subsequently tested with Bonferroni’s t-test for unpaired values. Bacterial counts were analyzed by the Mann–Whitney U-test or unpaired t-test using a parametric test with Welch’s correction. Survival curves were plotted using the Kaplan–Meier method and then compared using the log-rank method and Gehan-Wilcoxon test. Data were analyzed using GraphPad Prism version 6.00 for Mac (GraphPad Software, USA). A P < 0.05 was considered significant.

Gallstone disease is one of the most common etiologies of AP (6). Therefore, we modeled obstructive AP with BDO. All mice subjected to BDO died within 6 days after surgery, whereas all sham-operated (Sham) mice survived (Figure 1A). Compared to Sham animals, BDO animals displayed a significant increase in serum amylase/lipase and alkaline phosphatase/direct bilirubin, which are biomarkers associated with AP and biliary obstruction, respectively (Figures 1B–E). Additionally, we detected a significant augmentation of leukocytes in the peritoneal cavity of the AP mice compared to the Sham mice (Figure 1F). Histological examination of the pancreas indicated that the obstructive experimental AP model resembled major features of severe AP in patients (35). The pancreas from the Sham group showed a mean of histological scores of 1.75 (with a range from 0 to 2) with edema and inflammatory cell infiltration as the predominant characteristics (Figure 1G). Pancreatic tissue sections of the AP group exhibited a mean histological score of 5.0 (with a range from 3 to 7), characterized by diffuse acinar cell necrosis, inflammatory cell infiltration, and edema. These results confirm that the obstructive mouse model employed induces severe AP.

As meropenem is a broad-spectrum antibiotic used as a prophylactic treatment in severe AP (36), the effects of meropenem on the outcome of experimental AP were investigated. Posttreatment of AP C57BL/6 mice with meropenem significantly accelerated animal mortality (Figure 2A). As expected, all Sham mice treated with saline or meropenem survived (Figure 2A). To investigate the factor(s) underlying the observed mortality induced by meropenem in AP C57BL/6 mice, animals were pretreated with the antibiotic for 3 days, and AP was induced by BDO 12 h after the last dose. Meropenem-pretreated AP mice displayed a strong reduction in survival rate (median survival of 2 days) compared to saline-pretreated AP animals (median survival of 7 days, Figure 2B). In the two meropenem-exposed AP groups (posttreatment and pretreatment), the serum levels of amylase and lipase were equal to those in the saline-exposed AP mice (Figures 2C–F), suggesting that meropenem did not directly affect AP development.

As reported in previous studies (37, 38), treatment of naïve mice for 3 or 5 days did not increase the serum activities of alanine aminotransferase, aspartate aminotransferase, urea, creatinine, or lactate dehydrogenase (Figures S1A–E in Supplementary Material), indicating the faster mortality observed in the meropenem-treated animals was not due to acute toxicity mediated by this antibiotic. In addition, the effect of antibiotic on immune system activation was evaluated. Leukocyte infiltration in the lungs is a sensitive hallmark of systemic inflammation and a key step to lung injury in AP (39–42). Figure S2 in Supplementary Material shows that antibiotic treatment did not change the numbers of neutrophils (GR1^highCD11b⁺) and monocytes (F4/80^highCD11b^high) in lungs of naïve or Sham mice, however, significantly increased the F4/80^highCD11b^high cells in the lungs of AP mice. These results suggest whether antibiotic induced an activation of the immune system, it is not enough to generate a systemic inflammatory response in naïve or Sham mice. However, in the context of experimental AP, the antibiotic pretreatment could worsen the outcome of systemic inflammation.

Since antibiotic-induced bacterial dissemination has been associated with a poor inflammatory disease outcome (3, 5), the effect of the meropenem pretreatment on the CFU number in tissues and blood was investigated. The CFU numbers in the pancreas were similar in all groups (Figure 3A). In the spleen, the number of CFUs increased in the AP and meropenem-pretreated AP mice compared with the Sham groups but were similar between the AP and meropenem-pretreated AP mice (Figure 3B). However, 24 h after surgery, a significant increase in the CFU content in the blood was observed in meropenem-pretreated AP mice compared to AP mice (Figure 3C). The pictures provided in Figure 3C show an example of the high CFU content in the blood from meropenem-pretreated AP mice.

Compared to AP mice, meropenem-pretreated AP animals displayed approximately 10 times more CFUs in the small intestine but not in the colon or cecum (Figure S3 in Supplementary Material). Next, automated, classical biochemically based and microbiology-based methods were employed to identify the bacterial species. Approximately 80% of bacteria found in the small intestine were identified as E. gallinarum (Table S1 in Supplementary Material). The remaining 20% of the bacteria identified were Staphylococcus epidermidis, Staphylococcus lentus, and Bacillus spp. (Tables S2, S3, and S8 in Supplementary Material).

Importantly, E. gallinarum was also found in high abundance in the blood samples of 100% of meropenem-pretreated AP mice (9/9), indicating bacterial translocation following AP disease (Figure 3D). Although in lower frequency, Enterobacter cloacae complex strains (6/9) and Staphylococcus warneri (3/9) were observed in the blood of antibiotic-treated AP mice (Figure 3D; Tables S1, S4, and S5 in Supplementary Material). E. cloacae complex (3/9), Raoultella planticola (3/9), and S. lentus (6/9) were identified in saline-pretreated AP mice (Figure 3D; Tables S3, S4, and S6 in Supplementary Material). These results suggest a major change in the bacterial profile of the blood from meropenem-pretreated AP animals and reveal the emergence of E. gallinarum.

Since E. gallinarum was the most frequent species observed in the blood of meropenem-pretreated AP mice, this result was investigated further. E. gallinarum was positive for the VanC1 gene (Figure S4 in Supplementary Material) (43) and showed low resistance to vancomycin (MIC = 6–8 µg/mL), intermediate resistance to meropenem (zone of inhibition = 16 mm), and high resistance to clindamycin, ceftriaxone, and streptomycin (zone of inhibition = 0 mm), but E. gallinarum was sensitive to ampicillin (zone of inhibition = 30 mm). Notably, pretreatment (12/12 h, 3 days) of C57BL/6 and Balb/C mice with clindamycin and ceftriaxone also induced major reductions in the survival rates of AP mice (Figure S5 in Supplementary Material).

To investigate the existence of a possible link between E. gallinarum and the severity of AP, the intestinal microbiota of mice were suppressed with a cocktail of broad-spectrum antibiotics (29) and repopulated with a fecal or E. gallinarum suspension (Figure 4A). As expected (29), the cocktail of broad-spectrum antibiotics depleted all fecal cultivable aerobic bacteria. Oral administration of fecal or E. gallinarum suspensions to antibiotic-treated mice resulted in high levels of intestinal recolonization (Figure 4B). Moreover, fecal samples from mice treated orally with E. gallinarum, but not with feces, contained ~10⁸ CFU/g (Figure 4C) in the vancomycin-resistant enterococci (VRE) selective agar medium. All of these colonies were E. gallinarum as confirmed using the Vitek2 method. Strikingly, compared to mice recolonized with the fecal suspension, animals exposed to E. gallinarum displayed enhanced mortality during AP onset (Figure 4D). This effect was associated with E. gallinarum translocation into the blood (Figure 4E). Furthermore, intravenous administration of live E. gallinarum in AP mice induced significantly higher mortality rates compared to AP animals administered either saline or dead E. gallinarum (Figure 4F). Taken together, these results indicate that E. gallinarum is sufficient to exacerbate the severity of AP.

To determine if isolated E. gallinarum induces sepsis in naïve mice, Wt animals were challenged with these bacteria, and the survival rates were analyzed. All mice exposed to a systemic challenge with E. gallinarum died within 3 days of the infection (Figure 4G). Since E. gallinarum could induce cellular activation through a TLR2/6-dependent pathway (44) and since this receptor plays a pivotal role in mortality during severe sepsis (30), we challenged TLR2-deficient mice with E. gallinarum. Figure 4G shows that ~30% of TLR2-deficient animals survived after a lethal challenge with E. gallinarum. Similarly, TLR2-deficient mice were found to be partially resistant following exposure to a clinical strain of E. gallinarum isolated from a septic patient (Figure 4H). These data suggest that TLR2 is involved in the host pathogenesis of the E. gallinarum infection.