Phage Abp1 Rescues Human Cells and Mice from Infection by Pan-Drug Resistant Acinetobacter Baumannii

Acinetobacter baumannii (A. baumannii) is a non-fermentative, non-motile, catalase-positive, Gram-negative bacterium [1, 2], widely found in soil, water, sewage, and many healthcare environments. In clinical settings, this pathogen causes severe pneumonia, as well as infections in the urinary tract, bloodstream, and wounds [2-4]. In the 1970s, A. baumannii was susceptible to most antibiotics. However, an increasing number of clinical A. baumannii strains now exhibit multi-drug and pan-drug resistance. The rapid development of antibiotic resistance is due to the acquisition of plasmids, transposons, and integrons that carry clusters of genes encoding resistance to several antibiotic families [2, 4]. Effective management of A. baumannii infections has become an increasingly urgent issue, requiring the development of new tools.

A growing body of literature has validated the use of phages for therapy and prophylaxis in the war against drug-resistant bacteria [5-9], and many lytic phages have been isolated and tested in animals for their efficacy against bacterial infections. In 2011, the U.S. Food and Drug Administration approved a mixture of bacteriophages as a food additive for protection against foodborne bacterial disease. In addition, promising results have been obtained using phages isolated on-demand to eradicate gram-negative bacteria [10].

In a previous report we described Abp1, a bacteriophage isolated from hospital sewage [11]. Abp1 is a lytic phage of the pan-drug resistant A. baumannii strain AB1. In this study, we analyzed the potential of Abp1 as a phage therapy candidate and tested its effects in both local and systemic mouse infection models.

A. baumannii AB1 was isolated from a burn patient at Southwest Hospital in Chongqing, China. It is resistant to multiple antibiotics, including aminoglycosides, carbapenems, cephalosporins, tetracyclines, and quinolones, but is sensitive to polymyxin with a MIC of 0.236 mg/L [11]. Abp1 is a lytic phage of AB1, isolated from sewage at the same location. The characteristics of this phage are described in our previous report [11]. Endotoxin was removed from Abp1 cultures using the ToxinEraser endotoxin removal kit (GenScript, China).

Serial dilutions of A. baumannii AB1, ranging from 10 to 10⁹ CFU/ml, were mixed with 5×10¹⁰ plaque-forming units (PFU) of Abp1. 10 µL from each infected culture was spotted onto an agar plate and incubated at 37°C for 24 h. Colonies formed by surviving bacteria were considered to be resistant to Abp1. Bacteriophage resistance was calculated as the ratio of CFU relative to the number of plated bacteria.

To test the thermal stability of Abp1, phage samples (7.5×10¹⁰ PFU in a volume of 1.5 ml) were incubated at 50, 60, 70, and 80°C. 100 µL aliquots were withdrawn after 5 min, 15 min, 40 min, and 1 h, and diluted 10-fold. 10 µl of diluted phage was combined with 200 µl host bacteria, incubated for 15 minutes, mixed with 3 ml of 0.75% LB agar, and poured onto plates containing solid medium to determine phage titer. For the pH stability assay, aliquots were collected one hour after incubation in liquid media ranging from pH 2 to 13.

LDH release from HeLa and THP-1 cells was determined using a CytoTox96 cytotoxicity assay (Promega G1780) [12]. HeLa cells were cultured in Dulbecco’s Modified Eagle’s Medium and THP-1 cells in Roswell Park Memorial Institute 1640 Medium, both supplemented with fetal bovine serum (10%) and penicillin- streptomycin (1%), and incubated in 5% CO₂ at 37°C. When cells reached 70-80% confluence, various amounts of Abp1 were added to the wells. 2% Triton X-100 was added as a positive control, and an equal volume of PBS was used as a negative control. After incubation for 24 hours, 50 µL of supernatant was added to the wells of a 96-well plate containing 50 µL of reaction mixture. The plate was incubated at room temperature for 30 minutes, and then 50 µL Stop Solution was added to halt the reaction. Absorbance was measured at 490 nm.

To conduct killing assays, an overnight culture of AB1 was diluted 1: 100 into 50 ml of medium and incubated for 2 hours at 37 °C to reach a cell density of approximately 5×10⁷ CFU/ml (OD₆₀₀=0.5). One milliliter of the culture was added to the wells of a 24-well plate, followed by addition of 100 µl diluted Abp1. To achieve multiplicity of infection (MOI) values of 100, 10, 1, and 0.1, a total of 5×10⁹ PFU, 5×10⁸ PFU, 5×10⁷ PFU and 5×10⁶ PFU Abp1 were added to each well. Cell death was analyzed using a microplate reader (SpectraMax M2e, Molecular Devices). The parameters used by SoftMax Pro were as follows: read time interval = 1 hour, with shaking for 50 minutes at 37 °C between reads, for a total of 8 hours. Absorbance at 600 nm was recorded.

HeLa cells were cultured as described above in 4 wells of a 6-well plate. When cells reached 70-80% confluence, AB1 (10⁷ CFU) in 100 µL DMEM medium was added to 2 wells. For the negative controls, DMEM only was added. After 2 hours, Abp1 (10⁸ PFU) and an equal volume of PBS were added to all wells. The cells were incubated for additional 22 hours, and the number of living cells was determined using trypan blue.

The ability of Abp1 to confer protection against AB1 was evaluated using a mouse wound infection model as described previously [13]. Briefly, 36 six-week-old male BALB/c mice were randomly assigned to three groups (n=12 per group). Each mouse received cyclophosphamide (150 mg/kg body weight) via intraperitoneal injection on day -4 and 100 mg/kg on day -1. One day after the second injection, mice were anesthetized with amobarbital sodium and shaved. Two 5.0×5.0 mm full-thickness wounds were created on the dorsal side of each mouse using a skin biopsy punch. 25 µl of PBS containing 5.0×10⁴ AB1 cells were deposited on each wound and the region was air-dried for 5 min.

Four hours after infection, mice were inoculated with phage (5.0×10⁸ PFU) in PBS either via a subcutaneous injection or by directly pipetting into the wound. Inoculation with PBS alone was used as a negative control. The phage treatments were repeated once a day for 7 days after infection. On days 1, 3, and 7, wound sizes were measured and recorded.

In the mouse systemic infection model, 42 six-week-old male BALB/c mice were randomly assigned to three groups (n=14 per group). 200 µl of PBS containing 5.0×10⁷ AB1 cells was injected intraperitoneally into the mice. Immediately after infection, phage (5.0×10⁸ PFU) in PBS were injected intraperitoneally on the same side of the mouse used for the bacterial injection. Phage were injected daily for 6 days after infection. Because AB1 is sensitive to polymyxin B, this antibiotic was used as a positive control in the antibiotic treatment group, which received 10 mg/kg polymyxin B daily for 6 days. The same volume of PBS was injected as a negative control in the third group. The animals were monitored for seven days and the survival rate of each group was calculated. During the observation period, dead mice were dissected immediately to obtain the livers and kidneys. 7 days post-infection, survivors were also sacrificed and dissected. For all animals, livers and kidneys were weighed, homogenized in 1 ml saline, serially diluted, and bacterial counts and phage titers were determined.

All animal experiments complied with the International Guiding Principles for Biomedical Research involving Animals (1985) and were approved by the Laboratory Animal Welfare and Ethics Committee of Southwest Hospital, Third Military Medical University.

Data were analyzed by one-way analysis of variance (ANOVA) or log-rank test (Mantel-Cox) as appropriate. A value of P < 0.05 was considered significant.

Abp1 was isolated from sewage at Southwest Hospital. It has a genome size of 42 kb containing 51 genes, including an endolysin gene [14] as expected for a lytic phage. As shown in Fig. 1A, Abp1 can cause lysis in a bacterial culture after 3 hours, and has the ability to form large and transparent plaques (Fig. 1B). We attempted to isolate lysogens of Abp1 by recovery of colonies from the middle of plaques after incubation for one week. However, 10 of the tested colonies were still sensitive to Abp1, suggesting that they are not lysogens. Abp1 is a member of the Podoviridae family, it has a short tail and regular icosahedral head about 60-70 nm diameter. (Fig. 1C). A. baumannii AB1 resistance to Abp1 occurs at a frequency of 3.51±0.46×10^-8.

As shown in Fig. 2A, Abp1 is relatively thermostable, with high survival rates after incubation at 50 °C. However, only 0.01% of phage were viable after 1 h incubation at 60 °C, after incubation for 15 min at 70 or 80 °C, almost no phage survive. These results demonstrate that Abp1 retains activity well above 37 °C, which simplifies its storage and delivery. As shown in Fig. 2B, Abp1 also exhibits good pH stability. Abp1 survival was high from pH 5.0 to 10.0, although survival rates decreased significantly between pH 4.0 and 5.0, and from 10.0 to 11.0. No viable phage were detected below 3.0 or above 12.0. The broad range of thermal and pH stability facilitates the application of Abp1 in phagotherapy.

Potential cytotoxic effects of Abp1 were evaluated using HeLa and THP-1 cell assays (Fig. 3). No cytotoxicity was detected at concentrations of 10⁶, 10⁸, or even 10¹⁰ PFU/ml, in comparison to the negative PBS group. This suggests that Abp1 is safe for use in antimicrobial treatment in vivo.

To determine the effective dose range of Abp1, AB1 was infected with phage over a range of MOI, and cultures were monitored for lysis (OD at 600 nm). As expected, OD increased steadily during the 8-hour period in the negative control (Fig. 4), but decreased significantly in the cultures infected by Abp1. AB1 growth was inhibited at all MOIs, at the highest phage concentration (MOI=100), AB1 failed to grow, and cultures began to decline at an early stage. After slight increases, AB1 growth eventually declined at MOI of 10, 1, and 0.1, with higher phage concentrations resulting in more rapid lysis. Cultures at all four MOI declined to the same OD₆₀₀ level after 5 hrs incubation.

After co-incubation with Abp1 for 24 hours, viable HeLa cell counts were identical to those in the negative control (Fig. 5), consistent with the HeLa cell toxicity assay described above. In contrast, the viability of HeLa cells infected with AB1 only for 24 hours decreased significantly. When HeLa cells were treated with phage 2 hours after infection by AB1, their survival was almost identical to the uninfected negative control. This indicates that Abp1 can rescue HeLa cells from AB1 infection, even 2 hours after infection.

To investigate the effect of phage therapy on local infection, we devised an A. baumannii infection mouse model. As shown in Fig. 6, 3 and 7 days post-infection, wound sizes in animals receiving locally applied phage were significantly smaller than in mice receiving either systemically administered phage or no treatment (negative control). Although systemic phage therapy was associated with reduced wound size relative to the negative control, the difference was not significant. Interestingly, the wounds in mice receiving locally administered phage had a drier and cleaner appearance than wounds in the other two groups (data not shown). These results suggest that local Abp1 therapy has the potential to be used clinically in the control of local infections by multi-drug resistant A. baumannii, and demonstrate that the treatment accelerates wound healing.

In addition to testing the efficacy of Abp1 treatment in local infections, we also tested Abp1 in a systemic AB1 infection model. As shown in Fig. 7A, infected mice receiving no treatment succumbed rapidly. Six mice died 8 hours after infection, with an additional 6 dying during the first day. Only 1 mouse in this group survived the entire seven-day observation period. In contrast, all mice treated with phage or polymyxin B survived the entire 7 days of the observation period. In untreated mice, livers and kidneys contained high bacterial loads, while mice treated with phage or polymyxin B contained almost no bacteria in these organs (Fig. 7B). Finally, in phage-treated animals, phage were detected in both livers and kidneys (approximately 10⁴ to 10⁵ PFU/g), suggesting that phage replication occurred in vivo (Fig. 7C). Together, these data suggest that Abp1 is highly effective against systemic A. baumannii infection.

Because of increasing drug resistance, A. baumannii has become a serious threat to human health [15]. Phages, as “our enemy’s enemy,” are potential weapons against drug-resistant pathogens [16, 17]. Some phage products have already been approved by the FDA as food biopreservatives, such as ListShield and ListexP100 [18]. The Phagoburn clinical study, supported by several European countries, was conducted to assess the safety, effectiveness, and pharmacodynamics of two therapeutic phage cocktails to treat E. coli and P. aeruginosa burn wound infections [19]. In our previous survey, we found that A. baumannii was highly prevalent in our hospital, and a majority of isolates exhibited multi-drug resistance or pan-drug resistance [20, 21]. To explore the possibility that phage therapy might be effective against this pathogen, we isolated 21 A. baumannii phage strains from the hospital environment. Although nearly all had narrow host ranges, Abp1 exhibited a relatively wide host range. About 5 percent of over 200 clinical A. baumannii isolates are sensitive to Abp1 (data not shown). We also tested several phage cocktails, but the cocktail host range was the simple sum of the phage component host ranges. In our future work, we want to expand the host range by two ways. Firstly, isolate more phages against A. baumannii strains belonging to different ST types. Secondly, explore CRISPR-Cas9 technology to manipulate the phage tails which play an important role in recognizing host receptors.

The results presented here show that Abp1 has several qualities that are desirable for a potential antimicrobial agent. Abp1 is stable under a wide range of temperatures and pH environments, and has almost no detectable toxicity against HeLa or THP-1 cells. In addition, the A. baumannii isolate we used in our work exhibits phage resistance at very low levels, and lysogeny is rare or perhaps nonexistent. Abp1 can lyse nearly all cells even at low MOI, suggesting that a small number of phage may be sufficient for effective phage therapy.

In the wound infection model, topically applied Abp1 accelerated the healing of AB1-infected wounds. However, when phage are introduced systemically via subcutaneous injection, the effect is indistinguishable from the negative control. This result indicates that topical administration is a better choice for the treatment of local infections. In the systemic infection model, phage-treated animals exhibited a 100% survival rate when phage are introduced immediately after bacterial infection. Abp1 efficacy is equivalent to that of polymyxin B (10 mg/kg), and phage therapy is therefore a potential treatment for pan-drug resistant A. baumannii infections.

We have evaluated the safety, effectiveness, and methods of administration for Abp1 in the treatment of A. baumannii local and systemic infections. Our results provide evidence that phage therapy can be effective against multi-drug resistant A. baumannii infections, and suggest that additional clinical studies are warranted to confirm and expand these findings.

No conflict of interests exists.

We would like to thank Dr. Shouguang Jin (University of Florida, USA) for providing assistance with English grammar.

This work was supported by National Natural Science Foundation of China (No. 81571896 and No. 81772073) and Technological Innovation Plan in Major Fields of Southwest Hospital, Key Projects (No. SWH2016ZDCX2001).