Previous Article | Next Article ![]()
Antimicrobial Agents and Chemotherapy, May 2009, p. 2205-2208, Vol. 53, No. 5
0066-4804/09/$08.00+0 doi:10.1128/AAC.01166-08
Copyright © 2009, American Society for Microbiology. All Rights Reserved.

Department of Biological Sciences, University of Alberta, Edmonton, Alberta, Canada T6G 2E9
Received 29 August 2008/ Returned for modification 23 September 2008/ Accepted 3 February 2009
|
|
|---|
|
|
|---|
Chronic bacterial colonization of the major airways resulting in debilitating pulmonary infections and immune response exacerbations is the major cause of morbidity and mortality in cystic fibrosis (CF) patients (9). In the late 1970s and early 1980s, members of the Burkholderia cepacia complex (BCC) emerged as a group of opportunistic bacterial pathogens primarily affecting CF patients (for a review, see the work of Mahenthiralingam et al. [15]). In these patients, life-threatening infection with BCC bacteria is confounded by the fact that there are few therapeutic options available. Effective treatment of BCC infections is difficult due to the microorganisms' very high level of intrinsic resistance to many antibiotics and their ability to develop further resistance during therapy (26). Most BCC species are resistant to all classes of commonly used chemical antibiotics (1, 20). Triple-combination antibiotic therapy is believed to provide the most effective therapy against BCC infections (1); however, it is important to note that even when susceptibility to antibiotics is demonstrated in vitro, aggressive therapy often does not result in clinical improvement or even a reduction in bacterial numbers (9, 20). Although there are notable regional differences in the prevalences of species-specific BCC infections in CF patients (21), B. cenocepacia isolates are associated with higher prevalences and higher rates of morbidity and mortality (16, 27).
In this study, we examine B. cenocepacia strains K56-2 and C6433, both of which are BCC strains that have spread epidemically among patients with CF (14).
Multiple studies have demonstrated the effectiveness of phage therapy in animal models for the treatment of various bacterial pathogens, including Escherichia coli (24, 25), Pseudomonas aeruginosa (10, 18, 32), Klebsiella pneumoniae (31), Campylobacter jejuni (13), Enterococcus spp. (4, 30), and Staphylococcus aureus (5, 17). Although several BCC phages have recently been isolated and characterized (8, 22, 28, 29), to date there have been no published reports demonstrating the efficacy of phage therapy for BCC infections. Here, we demonstrate the efficacy of phages for treatment of fatal infection caused by clinical isolates of the BCC, using the Galleria mellonella infection model (23). Although G. mellonella has been used extensively for the study of human pathogens, including P. aeruginosa (12, 19), Bacillus cereus (7), and Francisella tularensis (2), there are few examples of antimicrobial agents being tested in this system.
BCC phage KS4-M is an uncharacterized phenotypic mutant of double-stranded DNA tailed Myoviridae phage KS4 (22), which, unlike its parent, is able to lyse liquid cultures of Burkholderia cenocepacia K56-2. High titers of phage KS4-M were prepared by adding KS4-M phage to K56-2 (pregrown for
3.5 h at 30°C in one-half-strength Luria-Bertani broth), and the mixture was shaken at 30°C until complete bacterial lysis was observed. Preparations were purified by centrifugation (1 x 104 x g for 10 min) and supernatant filtration (pore size, 0.45 µm). Phage KS4 was prepared as previously described (22). Following incubation of liquid B. cenocepacia K56-2 with KS4-M, the optical density at 600 nm (OD600) was 0.09, while the sample inoculated with KS4 had an OD600 of 0.44 and the uninfected control had an OD600 of 0.48. The concentrations of bacteria remaining in the samples were determined to be 6.9 x 106 CFU/ml for the KS4-M-infected culture, 1.6 x 108 CFU/ml for the KS4-infected culture, and 2.1 x 108 CFU/ml for the uninfected control. Concomitantly, the KS4-M-infected sample produced more phage, having a titer of 2 x 107 PFU/ml, than the KS4-infected sample (2 x 105 PFU/ml).
Three novel BCC phages were isolated from soil obtained from the Muttart Conservatory (Edmonton, Alberta, Canada), using a procedure similar to that described previously (22), involving KS12 from soil planted to Dietes grandiflora (wild iris) and KS14 and DC1 from soil planted to a Dracaena sp. (dragon tree). We determined the host ranges of newly isolated KS12, KS14, and DC1 phages against a previously utilized BCC panel comprising 24 strains, representing nine different species of the BCC (6, 14, 46). In addition to this strain panel, three additional BCC strains were tested for host range in this study: B. cenocepacia PC184 and Cep511 (LMG 18830) and Burkholderia vietnamiensis DBO1 (ATCC 29424). KS12 was isolated with B. cenocepacia strain K56-2 and, of the 27 strains in the BCC panel tested, is able to form plaques only on K56-2 and Burkholderia multivorans C5274. KS14 was isolated using B. multivorans strain C5393 and has a considerably broader host range, capable of producing plaques on 8 of the 27 strains tested. These eight strains are B. multivorans strains C5393 and C5274; B. cenocepacia strains 715J, C6433, C5424, and PC184; Burkholderia dolosa strain 21443; and Burkholderia ambifaria strain 17828. DC1 was isolated using B. cepacia strain LMG18821, and this phage is able to form plaques on B. cenocepacia strains K56-2, C6433, PC184, and Cep511 and Burkholderia stabilis strain LMG18870. KS14 and DC1 are double-stranded-DNA-tailed myoviridae (data not shown).
Phages were tested for their ability to persist in the hemolymph of uninfected G. mellonella larvae over time (Fig. 1). All phage preparations were applied to a Detoxi-Gel endotoxin-removing column (Pierce Biotechnology, Rockford, IL) and serially diluted in 10 mM MgSO4 plus 1.2 mg/ml ampicillin prior to injection. Ampicillin was present in the inoculum to prevent infection with bacteria naturally present on the surfaces of the larvae. A 250-µl Hamilton syringe fitted with a reproducibility adapter was used to inject 5-µl aliquots into G. mellonella. Following injection, larvae were placed in a static incubator in the dark at 30°C. A 20-gauge needle was used to withdraw hemolymph samples from larvae for isolation of phage and bacteria on BCSA medium (11). At various times, 10 µl of hemolymph was collected from five larvae and combined; this sample was then serially diluted in suspension medium and plated with the appropriate BCC strain for detection and quantification of plaques. Larvae injected with phage KS4-M alone showed no decrease in number of phage collected from the hemolymph over a 48-hour period (Fig. 1). These results are also representative for phages DC1 and KS4, whose phage titers did not decline in the hemolymph over time. However, larvae injected with KS12 or KS14 alone showed a decline in number of phage isolated over a 48-hour period.
![]() View larger version (17K): [in a new window] |
FIG. 1. Bacteriophage persistence in G. mellonella larvae. Uninfected larvae were injected with individual phages (KS4 [ ], KS4-M [ ], DC1 [line without symbol], KS12 [ ], or KS14 [x]). Equal volumes of hemolymph were collected from five worms at each time point, combined, serially diluted, and plated with B. cenocepacia K56-2 or B. cenocepacia C6433 for quantification. Each point represents the average for three trials, and the standard deviations are indicated.
|
2.5 x 103 CFU of B. cenocepacia strain K56-2 had a mortality rate of nearly 100% at 48 h p.i. As shown in Table 1, larvae treated with KS4-M immediately following infection showed increased survival rates, and this rescue was dependent upon the number of phage used in the therapeutic dose. The optimal phage dose in this experiment was approximately 2.5 x 103 PFU, resulting in a multiplicity of infection (MOI) of 1. At this dose, 60% of the larvae were alive at 48 h p.i., whereas none of the untreated larvae were viable at 48 h p.i. Despite in vitro growth differences in liquid culture, phage KS4 showed results similar to those for KS4-M in its ability to treat K56-2-infected larvae (data not shown). |
View this table: [in a new window] |
TABLE 1. Mortality rates of phage-treated BCC-infected G. mellonella
|
2.5 x 103 CFU) was administered, and at various intervals thereafter, ranging from 6 to 24 h, the larvae received a single injection of phage KS4-M (MOIs from 1 to 1,000). Treatment at either 6 h or 12 h p.i. with an MOI of 1,000 was the most effective. In this experiment, more larvae were viable at 48 h p.i. than when the phage dose was administered immediately following infection (20% mortality versus 40% mortality). B. cenocepacia K56-2 isolated from hemolymph samples of treated larvae subsequently showed resistance to KS4-M infection in vitro. Bacterial DNA was isolated from individual colonies by using standard procedures (3). Oligonucleotide primers F8.3 (5'-TGTTCAGAGATGCGTTCGAC-3') and R9.2 (5'-ATGGCGCTTGACAGGTAATC-3') (28) were used in a PCR assay to determine the presence of KS4-M within B. cenocepacia K56-2 isolates from hemolymph samples of larvae undergoing KS4-M phage therapy. PCR was performed using Taq PCRx DNA polymerase under standard conditions (Invitrogen, Burlington, Ontario, Canada). All of the resistant K56-2 isolates were PCR positive for KS4-M, indicating that lysogeny had occurred in vivo (data not shown). Phage KS12 was also investigated as a potential candidate for treating B. cenocepacia K56-2-infected larvae. As previously observed with KS4-M phage therapy, immediate therapy of infected larvae with KS12 can increase the survival rate of larvae and this effect is dependent on the number of phage in the therapeutic dose (Table 1). At the highest MOI tested, KS12 was able to rescue over 90% of infected larvae. Delayed therapy with KS12 (up to 12 h p.i.) also resulted in an increase in survival rate, and this increase was proportional to the number of phage in the therapeutic dose. As with all phages tested, therapy delayed to 24 h p.i. was not effective at increasing the survival rate of infected larvae. As anticipated, hemolymph samples collected from G. mellonella larvae infected with K56-2 that received treatment with KS12 were often completely devoid of bacteria, producing sterile hemolymph. This effect was observed only with phage KS12.
In order to determine whether the effects that we observed with phage therapy were associated with a nonspecific immune activation response, as a control, we tested heat-inactivated phage for their ability to rescue BCC-infected larvae. Heat inactivation was achieved by incubating KS12 (4 x 109 PFU/ml) at 80°C for 20 min. This sample, in which no viable phage was detected by in vitro plating, was used to treat larvae infected with a lethal dose of B. cenocepacia K56-2. Ten larvae were injected with a lethal dose of K56-2 (
2.5 x 103 CFU) and immediately treated with heat-inactivated KS12 (MOI > 5,000). Larvae that were treated with heat-inactivated KS12 and untreated larvae had 100% mortality rates at 48 h p.i. This suggests that larval survival is entirely due to phage antibacterial activity rather than to host immune stimulation.
In order to demonstrate phage therapy effects against BCC strains other than K56-2, G. mellonella larvae were injected with a lethal dose of B. cenocepacia C6433 and immediately treated with KS14. Because the 50% lethal dose for B. cenocepacia C6433 is 3.0 x 104 CFU in this infection model (23), an injected lethal dose of
1.0 x 105 CFU was required. We were therefore limited to testing the efficacy of low KS14 phage MOIs because even the highest attainable concentrations of injected phage resulted in an MOI of only 0.1. However, even at these low numbers of phage, a therapeutic effect was observed. G. mellonella larvae treated with the highest dose of phage (MOI = 0.1) had mortality rates of approximately 50%, while untreated, infected larvae had a 100% mortality rate at 48 h p.i. (Table 1). In vitro evaluation of B. cenocepacia C6433 isolates collected after KS14 treatment revealed the presence of both phage-sensitive and phage-resistant cells. In contrast, G. mellonella larvae challenged with a lethal dose of B. cenocepacia C6433 (
1.0 x 105 CFU) and immediately treated with BCC phage DC1 at various MOIs showed no protective effects. Although C6433 is highly sensitive to DC1 in vitro and we were able to achieve higher MOIs using DC1 than with KS14, at all DC1 MOIs tested (10 to 0.0001), no therapeutic effect of phage DC1 was observed (data not shown). Uninfected G. mellonella larvae that received a mock injection in place of bacteria and that were subsequently injected with phage DC1 exhibited no mortality, indicating that the larval mortality was not due to the injection of the DC1 phage.
The BCC is well known for its high-level multidrug resistance, which renders patients defenseless, with virtually no traditional antibiotics available to treat these infections. Bacteriophage therapy may provide a reasonable alternative to ineffective, aggressive antibiotic therapy for patients suffering from BCC infections. In this study, we used an alternative caterpillar infection model to evaluate the potential of phage therapy to eradicate experimental infections caused by clinically significant BCC strains. The in vivo phage therapy experiments with BCC infections in G. mellonella larvae conducted in this study are the first to demonstrate the efficacy of BCC phages as antibacterial agents. We were able to abolish the lethal effects of a B. cenocepacia K56-2 infection in more than 90% of infected larvae with a singe injection of phage KS12. Our results with heat-inactivated KS12 demonstrate that the ability of phage to rescue infected larvae is a direct result of phage activity and not a result of a host immune response mounted against the injected phage. This finding is further supported by the observation that treatment of BCC-infected larvae with KS12 often results in larvae with hemolymph that is sterile compared to that of untreated larvae. Our results suggest that variables such as phage persistence, in vivo antibacterial activity, and lysogenic potential all play a role in the effectiveness that a phage will have against a particular bacterial strain in a specific host organism. Although clear differences in treatment efficacy between BCC phages were observed, our results demonstrate that BCC phage therapy has the potential to be effective against the BCC in vivo.
We thank Danielle Carpentier for isolating the DC1 phage used in this study as well as Mandi Goudie for help in characterizing the DC1 phage host range.
Published ahead of print on 17 February 2009. ![]()
|
|
|---|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»