Multidrug-resistant Acinetobacter baumannii resists reactive oxygen species and survives in macrophages

We investigated the intracellular survival of multidrug-resistant Acinetobacter baumannii (MDRAB) clinical isolates in macrophages, after phagocytosis, to determine their virulence characteristics. After ATCC 19606 and 5 clinical isolates of MDRAB were phagocytosed by mouse and human macrophages, the bacterial count of MDRAB strains, R4 and R5, increased in the mouse macrophages, 24 hours after phagocytosis. Bacterial count of the strains, R1 and R2, was almost equal 4 and 24 hours after phagocytosis. Intracellular reactive oxygen species was detected in the macrophages after phagocytosis of these bacteria. Further, the strains R1, R2, R4, and R5 showed higher catalase activity than ATCC 19606. Additionally, strains R1, R4, and R5 grew more efficiently than ATCC 19606 in the presence of H2O2, whereas growth of strains R2 and R3 was marginally more than that of ATCC 19606 in the presence of H2O2. The MDRAB clinical isolates altered the expression of TNF-α, IL-1β, IL-6, and MIP-2 mRNA induced in J774A.1 cells, 24 hours after phagocytosis. These results provide insights into the renewed virulence characteristics of MDRAB clinical isolates. Finally, tigecycline killed MDRAB phagocytosed by the macrophages more effectively than colistin, although colistin and tigecycline are both considered effective antibiotics for the treatment of MDRAB.

In A. baumannii infections, the production of ROS or NO appears to contribute to bactericidal function of neutrophils and macrophages and plays a crucial role in host defence and survival 24,25 . As a defence mechanism, S. aureus expresses the enzymes super oxide dismutases and catalase that protect it against ROS and enable its survival within the phagolysosome 20 . Likewise, A. baumannii is a catalase-positive bacterium, where in, catalase is encoded by the katE/katG genes. Additionally, the universal stress protein UspA protects it against H 2 O 2 stress 26,27 , suggesting that A. baumannii survives within phagolysosomes of macrophages through the degradation of H 2 O 2 by its catalase activities. Although, the uptake of A. baumannii by alveolar macrophages and murine macrophage cell line J774A.1 has been explored 25 , few studies have focused on the intracellular survival of A. baumannii in macrophages because it is regarded as an extracellular pathogen.
We have previously reported that the renewed virulence characteristics of A. baumannii clinical isolates depend on its ability to adhere to human epithelial cells, and on the expression level of omp mRNAs 17 . These results might imply that since the clinical isolates of A. baumannii may have been exposed to various environmental stress conditions in the hospital, numerus virulence factors in the clinical isolates may have been modulated. Therefore, in this study, we have focused on the intracellular survival of MDRAB clinical isolates in macrophages, and their catalase activity. We have further evaluated the expression levels of ROS and proinflammatory cytokines in macrophages after phagocytosis with the aim of exploring the influence of intracellular bacteria on the functioning of macrophages. Finally, colistin and tigecycline, which are considered effective antibiotics for the treatment of MDRAB, have been evaluated for their ability to kill intracellular MDRAB clinical isolates within macrophages.

MDRAB clinical isolates survive in macrophages.
Previous studies have shown that mouse macrophages can rapidly and efficiently phagocytose A. baumannii in vitro without the presence of antibody or complement opsonisation 25 . Therefore, we examined the MDRAB clinical isolate counts in J774A.1 and human macrophages at 4 and 24 hours after phagocytosis. As shown in Fig. 1a, E. coli, ATCC 19606, and 5 clinical isolates of MDRAB were detected in J774A.1 cells at 4 hours after phagocytosis. The bacterial count of E. coli and MDRAB strain R3, at 24 hours after phagocytosis, was significantly decreased compared with that at 4 hours after phagocytosis, whereas the bacterial counts of strains R4 and R5 at 24 hours after phagocytosis were increased compared with that at 4 hours after phagocytosis. Bacterial counts of the strains R1 and R2 were almost equal at 24 hours as well as 4 hours, after phagocytosis. Additionally, we examined the counts of E. coli, ATCC 19606, and a representative MDRAB strain, R1, in human macrophages at 4 and 24 hours after phagocytosis. As shown in Fig. 1b, the bacterial counts of E. coli and ATCC 19606 at 24 hours after phagocytosis were significantly decreased compared with that at 4 hours after phagocytosis, whereas a slight decrease in the count of strain R1 was observed at 24 hours compared with that at 4 hours, after phagocytosis. We next examined whether A. baumannii was phagocytosed by the macrophages or they invaded into the macrophages. J774A.1 cells were co-cultured with the bacteria in the presence of cytochalasin D (CytD), a potent inhibitor of micropinocytosis/phagocytosis. As shown in Fig. 1c, the bacterial counts of E. coli, ATCC 19606, and 3 representative MDRAB strains R1, R3, and R5 in CytD-treated J774A.1 cells were significantly lower than that in non-treated J774A.1 cells at 4 hours after phagocytosis. These results indicate that MDRAB clinical isolates were phagocytosed by the macrophages, following which, they managed to survive within the cells.
MDRAB clinical isolates induce, but not alter, RoS production in macrophages. On phagocytosis, the production of ROS in the phagolysosome plays a crucial role in destroying of microorganisms [20][21][22][23] . Therefore, we examined whether macrophages produced ROS in response to intracellular MDRAB. As shown in Fig. 2a, ROS production was detected in J774A.1 cells at 24 hours after phagocytosis of E. coli, ATCC 19606, and 3 representative MDRAB strains. However, the induction of ROS was of about the same level in E. coli, ATCC 19606, and MDRAB strains R3 and R5, whereas the ROS level in J774A.1 cells co-cultured with MDRAB strain R1 was slightly and significantly lower than that when J774A.1 cells were co-cultured with E. coli (Fig. 2b). These results suggest that macrophages produced ROS in response to intracellular MDRAB.
MDRAB clinical isolates have high catalase activity. The results mentioned above suggest that macrophages produced ROS in response to intracellular MDRAB, however, could not eliminate intracellular MDRAB completely. As A. baumannii is a catalase-positive bacterium, we hypothesised that the catalase activity of MDRAB clinical isolates was upregulated. As shown in Fig. 3a, the expression of katE mRNA in MDRAB clinical isolates was significantly and substantially higher than that in ATCC 19606. Moreover, the expression of katG mRNA in strains R4 and R5 was significantly higher than that in ATCC 19606 (Fig. 3b). As shown in Fig. 3c, the catalase activity of ATCC 19606 was significantly higher than that of E. coli. Moreover, the catalase activity of the strains R1, R2, R4, and R5 was significantly higher than that of ATCC 19606, whereas catalase activity of strain R3 was almost equal to that of ATCC 19606. These results suggest that MDRAB clinical isolates exhibit upregulated catalase activity, which primarily depends on the expression level of the katE gene.
MDRAB clinical isolates resist toxicity caused by hydrogen peroxide. ROS plays a crucial role in the rapid killing of A. baumannii ingested by phagocytes 24 . However, our study indicated that macrophages produced ROS in response to MDRAB clinical isolates, albeit, could not eliminate them. Additionally, MDRAB clinical isolates had high catalase activity (Fig. 3c). Therefore, to clarify the resistance of MDRAB clinical isolates to ROS, we evaluated the growth of MDRAB clinical isolates in a medium containing H 2 O 2 in vitro. MDRAB clinical isolates showed very similar growth when cultured for 24 hours in the absence of H 2 O 2 (Fig. 4a). However, E. coli and ATCC 19606 showed no growth when cultured for 24 hours in the presence of 51 mM H 2 O 2 , whereas MDRAB strains R2 and R3 showed little growth when cultured for the same time period (Fig. 4b) (Fig. 4c). Moreover, strains R2 and R3 showed little growth in the presence of 51 mM H 2 O 2 (Fig. 4d), whereas strains R4 and R5 showed about 50% growth in the presence of 102 mM H 2 O 2 (Fig. 4e). These results suggest that some MDRAB clinical isolates have the ability to survive under oxidative stress within the phagolysosome of macrophages. Pearson correlation analysis revealed that the intracellular bacterial count in the macrophages was positively and significantly correlated with the growth rate of bacteria in the presence of 51 mM H 2 O 2 in vitro (r = 0.845, P = 0.010) (Fig. 4f). However, Pearson correlation analysis revealed that the catalase activity of these bacteria was not correlated with both, the intracellular bacterial count in macrophages as well as the growth rate of bacteria in the presence of 51 mM H 2 O 2 in vitro (data not shown). This was consistent with the observation that the MDRAB strain R2 exhibited high catalase activity, however, could not survive in a medium containing H 2 O 2 .

MDRAB clinical isolates alter the expression of proinflammatory cytokines in macrophages.
We have reported previously that MDRAB clinical isolates alter the expression of proinflammatory cytokines in human epithelial cells, suggesting that the clinical isolates had acquired renewed virulence characteristics 17 . Therefore, in this study, we evaluated the mRNA levels of proinflammatory cytokines in J774A.1 cells at 24 hours after phagocytosis of E. coli, ATCC 19606, and the MDRAB representative strains R1, R3, and R5. The expression of TNF-α, IL-1β, IL-6, and MIP-2 mRNA was induced in J774A.1 cells at 24 hours after phagocytosis of the bacteria ( Fig. 5a-d). The mRNA levels of TNF-α and IL-1β in J774A.1 cells at 24 hours after phagocytosis of ATCC www.nature.com/scientificreports www.nature.com/scientificreports/ 19606 were significantly lower than those after phagocytosis of E. coli, whereas those of the three representative MDRAB strains were significantly higher than those of ATCC 19606 (Fig. 5a,b). The mRNA level of IL-6 in J774A.1 cells at 24 hours after phagocytosis of ATCC 19606 was significantly lower than that after phagocytosis of E. coli, whereas that of strain R3 was significantly higher than that of ATCC 19606 (Fig. 5c). The mRNA levels of MIP-2 in J774A.1 cells at 24 hours after phagocytosis of ATCC 19606 were significantly lower than those after phagocytosis of E. coli, whereas those of strains R1 and R5 were significantly higher than those of ATCC 19606 (Fig. 5d). The expression of IL-10 mRNA was lowest in J774A.1 cells at 24 hours after phagocytosis of ATCC 19606, and this expression was significantly higher in that of the three MDRAB strains, and highest in that of E. coli (Fig. 5e). These results suggest that MDRAB clinical isolates alter the expression of proinflammatory cytokines in macrophages through their virulence factors. However, Pearson correlation analysis revealed that www.nature.com/scientificreports www.nature.com/scientificreports/ the mRNA level of each proinflammatory cytokine was not significantly correlated with the intracellular bacterial count of the macrophages (data not shown).
Tigecycline is an effective antibiotic for intracellular MDRAB. Colistin and tigecycline are considered effective antibiotics for the treatment of MDRAB 28 . We evaluated whether these antibiotics were effective in killing intracellular MDRAB clinical isolates in the macrophages. Treatment of J774A.1 cells with a high concentration of colistin (50 μg/mL), used for the killing of extracellular MDRAB in the phagocytosis assay, did not successfully kill intracellular MDRAB (Figs. 1a and 6). These results indicate that colistin is not transported into the host cells. However, in mouse and human macrophages, the survival rates of ATCC 19606 and an MDRAB representative strain R1, were decreased in the presence of tigecycline in a dose-dependent manner (Fig. 6). These results suggest that tigecycline is an effective antibiotic for the killing of intracellular MDRAB.

Discussion
A. baumannii has recently emerged as a major nosocomial pathogen 1,2 . Although, we have previously reported the virulence characteristics of A. baumannii clinical isolates in human epithelial cells 17 , the pathogenicity of A. baumannii clinical isolates within macrophages has remained elusive. Therefore, in the current study, we focused on the survival of intracellular MDRAB clinical isolates phagocytosed by macrophages, and the virulence factor responsible for the resistance to killing exhibited by this organism. Additionally, we evaluated the effect of antibiotics on intracellular MDRAB clinical isolates, commonly used for the treatment of MDRAB infections.
The role of macrophages in A. baumannii infection has been analysed in a previous study, and they are known to play a crucial role in early host defence, especially against respiratory A. baumannii infection 25   www.nature.com/scientificreports www.nature.com/scientificreports/ isolates induced rapid bacterial replication in the lungs, significant extrapulmonary dissemination, and severe bacteremia by 24 hours of postintranasal inoculation, eventually leading to death of the animals, whereas infecting the animals with ATCC strains did not produce the same results 29 . In our previous study, we have reported that MDRAB clinical isolates express different levels of virulence factor omps, and exhibit high adherence capacity for human epithelial cells, compared with A. baumannii ATCC 19606 17 . These results suggest that clinical isolates of A. baumannii exhibited renewed virulence characteristics. Considering the above findings, in the current study, we examined the intracellular survival of MDRAB clinical isolates in macrophages in vitro. In fact, 2 of 5 clinical isolates survived and increased obviously in mouse macrophages for 24 hours after phagocytosis. These results suggest that MDRAB clinical isolates have acquired increased virulence capacity.
Considering the intracellular survival of clinical isolates of MDRAB in macrophages, it is likely that A. baumannii might evade the bactericidal action of ROS in phagolysosomes of macrophages. We evaluated the growth of MDRAB clinical isolates in a medium containing H 2 26 . In addition, the expression of katE mRNA in A. baumannii was drastically increased during the stationary growth phase, as compared to during exponential growth 26 . In the present study, we clarified that MDRAB clinical isolates have expressed high katE mRNA expression levels in the exponential growth phase. These results suggest that the clinical isolates have the capacity to degrade H 2 O 2 in the exponential growth phase. Moreover, as the expression of katG mRNA in strains R4 and R5 was higher than that in other strains, we consider that, compared to the other strains, these two strains could grow in high concentrations of H 2 O 2 .
The universal stress protein UspA plays a crucial role in protecting A. baumannii from H 2 O 2 , low pH, and 2,4-DNP 27 . Moreover, a recent study has reported that OxyR (defined as a transcriptional regulator of H 2 O 2 stress response) 30 regulates the major H 2 O 2 -degrading enzymes, encoded by katE and ahpF1, (which encodes alkyl hydroperoxide reductase), in A. baumannii 31 . In the present study, the expression of katE mRNA in MDRAB clinical isolates was significantly higher than that in ATCC 19606, whereas the expression of ahpF1 mRNA in MDRAB clinical isolates was slightly higher than that in ATCC 19606 (data not shown). Considering these results, novel transcription factors may regulate the expression of the katE gene. Additionally, the MDRAB strain www.nature.com/scientificreports www.nature.com/scientificreports/ R2 showed high catalase activity but could not survive in media containing H 2 O 2 . Further studies are required for in depth analysis of these clinical isolates.
Considering that A. baumannii has emerged as a major pathogenic species, this bacterium exhibits varying pathogenicity in different environments. Since A. baumannii clinical isolates may have been previously exposed to environmental stress conditions such as multiple antimicrobial agents, bactericidal substances in the serum, and host immune responses such as oxidative stress, they may have altered the expression of several genes in order to adapt to these stress conditions. Wright et al. found insertion sequence (IS) elements throughout the genome of A. baumannii, and these elements contributed to genome variation by interrupting genes or altering gene expression 32 . In addition, A. baumannii carrying an ISAba1 element upstream of the catalase-peroxidase gene katG, was selected by serial subculture in the presence of sub-inhibitory concentrations of H 2 O 2 32 , suggesting that A. baumannii has the ability of adapting to oxidative stress. In the present study, although the expression of katG mRNA in strains R4 and R5 was significantly higher than that in ATCC 19606, no insertion sequence in the upstream region of the katG gene in MDRAB clinical isolates was found (data not shown). The DNA sequence upstream of the katG gene in MDRAB clinical isolates was the same as that in A. baumannii BJAB0868, BJAB07104, AC29, and AC30 clinical isolates 33,34 . As MDRAB clinical isolates exhibit enhanced catalase activity, further studies are required to determine novel transcription factors of katG as well as katE in these clinical isolates.
Colistin and tigecycline are considered effective antibiotics for the treatment of MDRAB 28 . Tigecycline acted synergistically with colistin, to exert antibacterial effects on A. baumannii in vitro 35 . However, cohort studies have reported that a combination of colistin and tigecycline against MDRAB, showed disappointing results 7,36 and tigecycline-based therapy may not be the best option for treating MDRAB infection 37 . We clarified the antimicrobial effect of tigecycline on intracellular MDRAB that had survived in macrophages after phagocytosis in vitro. Our laboratory has previously reported a novel bacterial transport mechanism, where A. baumannii exploits human neutrophils by adhering to the cells and inducing IL-8 release for bacterial portage 38  In summary, we demonstrated that MDRAB clinical isolates acquired renewed virulence characteristics after phagocytosis by macrophages. High catalase production by MDRAB may impair the intracellular killing by macrophages and the consequent spread of A. baumannii infections. Further studies are required to understand the resistance mechanisms employed by MDRAB to evade killing by macrophages.

Materials and Methods
All methods were carried out in accordance with relevant guidelines and regulations.
Bacterial strains and growth condition. R1, R2, R3, R4, and R5 strains of A. baumannii were isolated from the Teikyo University hospital during an outbreak that occurred around 2010. The bacteria were isolated on CHROMagar ™ Acinetobacter and incubated for 24 hours at 37 °C. As shown in Table 1, the R1 strain was isolated from the sputum of a patient with interstitial pneumonia. The R2 strain was isolated from a urine sample of a patient with malignant lymphoma and pneumonia. The R3 strain was isolated from the blood of a sepsis patient with myelodysplastic syndrome. The R4 strain was isolated from a stool sample of a patient with multiple myeloma and bacterial colonization. The R5 strain was isolated from the sputum of a patient with cardiovascular disease and bacterial colonization. The isolates were streaked onto blood agar plates and cultivated for 24 hours to obtain monoclonal colonies and identified as A. baumannii by DNA sequencing of a partial RNA polymerase β-subunit (rpoB) gene (La Scola et al., 2006). Additionally, the isolates were confirmed as non-clonal by pulsed-field gel electrophoresis (data not shown). After identification, these isolates were stored in glycerol stocks at −80 °C at the Department of Microbiology & Immunology, Teikyo University School of Medicine. Antimicrobial susceptibility testing was performed using 5 strains of A. baumannii based on the minimum inhibitory concentrations (MICs) of imipenem, amikacin, and ciprofloxacin. Against these 5 strains, the MICs of imipenem, amikacin, and ciprofloxacin were >8, 16, and 2 mg/L, respectively. These strains were thus identified as MDRAB strains. The A. baumannii ATCC19606 strain (AB) and the Escherichia coli ATCC 25922 strain (EC) were used as standard strains. The MIC of colistin against both ATCC 19606 and MDRAB strains, was determined as 2 mg/L. The MIC of tigecycline for ATCC 19606, and strains R1, R2, R4, and R5 was found to be 0.5 mg/L. The MIC of tigecycline for the R3 strain was 1 mg/L (Table 1). These bacteria were cultured on Luria-Bertani (LB) agar plates (Becton, Dickinson and Company, MD, USA) for 16 hours at 37 °C. Thereafter, the bacteria were suspended in RPMI-1640 Medium with L-glutamine and sodium (RPMI-1640) (Sigma-Aldrich, Tokyo, Japan), supplemented with 10% heat-inactivated fetal bovine serum (FBS) (Gibco, NY, USA) at a concentration of 2 × 10 7 CFU/mL, with the concentration being adjusted via optical density (OD) measurements at 595 nm. The bacterial suspensions thus obtained were used for the phagocytosis assay. cell culture and phagocytosis assay. J774A.1 cells were purchased from the JCRB cell bank (Osaka, Japan), and maintained at 37 °C under 5% CO 2 in RPMI-1640 medium supplemented with 10% FBS. J774A.1 cells were seeded at a concentration of 2 × 10 5 cells/well in 24-well plates and cultured overnight. Prior to co-culture with A. baumannii, the J774A.1 cells were washed twice with PBS. The cells were then co-cultured at a multiplicity of infection of 100 bacteria per cell at 37 °C under 5% CO 2 for 2 hours. In order to determine the intracellular viable bacteria at 4 and 24 hours, after 2 hours of co-culture, cells were washed 3 times with PBS, following which they were maintained in RPMI-1640 medium, supplemented with 10% FBS and 50 mg/mL colistin to kill the extracellular bacteria. To inhibit the phagocytosis of J774A.1 cells, they were pre-incubated for 1 hour with cytochalasin D (CytD) (FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan) (5 μg/mL) before co-culture with bacteria and were kept with CytD during co-culture with bacteria. To analyse the number of intracellular bacteria, J774A.1 cells were washed thoroughly 3 times with PBS after co-culture. The bacteria were harvested after lysing the J774A.1 cells by adding sterile distilled water (1 mL) to each well; bacterial count was confirmed by the growth of serial dilutions of the bacterial suspension on LB agar in terms of CFUs after 24 hours of incubation at 37 °C.
Differentiation of human macrophages. Human peripheral blood mononuclear cells (PBMC) were isolated from the peripheral venous blood of healthy volunteers. Briefly, whole blood (20 mL) was mixed with 7 mL of a 6% dextran solution and 15 mL of HBSS and allowed to stand for 30 minutes at 25 °C until stratification occurred. The upper leukocyte-rich plasma layer was transferred to a new tube containing endotoxin-free Ficoll-Paque PLUS gradient (GE Healthcare Japan, Tokyo, Japan) and was centrifuged (500 g, 30 min, 25 °C). The cell layer was harvested and subsequently, the cells were resuspended at 100 million PBMC per mL in Monocyte