An In vivo Wound Model Utilizing Bacteriophage Therapy of Pseudomonas aeruginosa Biofilms

Bacteriophages have been used as effective therapy against bacterial biofilms on devices such as catheters, in the lungs such as in cystic fibrosis, and even in infected food. Unlike antibiotics, they are bacteria-specific and produce the desired effect without systemic complications; they can develop bacterial resistance, although in ways different from antibiotics.The present study aimed to assess the effect of bacteriophages against multidrug-resistant Pseudomonas aeruginosa in a mouse wound model. P. aeruginosa obtained from laboratory culture of burn wounds were characterized, harvested, and titrated, and biofilms were generated on sterile catheter sections (10⁵ colony forming units/mL). Subcutaneous pockets were created on the backs of 24 male albino mice. Animals were randomized into 4 groups of 6 each. After evaluating a significant phage-bacteria interaction in vitro, 2 biofilm-laden catheter sections were implanted in subcutaneous pockets in mouse groups C and D. Sterile catheter sections only were implanted in group B. Group A had only a subcutaneous pocket without any catheter section. Phage cocktail solutions (10 µL of 10⁷ phage forming units/mL) were injected daily in group D pockets only. Groups B and C received 10 µL of normal saline. After 10 days, the catheter sections were explanted from groups B, C, and D and tissue biopsy was taken from group A pockets and cultured for bacterial and phage colony counts. A significant drop in bacterial counts from 3.87 x 10⁶ to 3.52 x 10⁴ was observed in group D when compared with group C  (3.87 x 10⁶ to 3.85 x 10⁵, P <0.05)  A significant rise in the phage counts from 1 x 10⁷ to 6.81 x 10⁸ (P <0.05)  also was observed in group D when compared with the baseline counts, indicating active phage proliferation and successful bacterial kill in group D. The present laboratory study could be indicative of a new treatment approach for multidrug-resistant bacterial infections, including wound infections.

Biofilm is a multicellular community composed of prokaryotic and/or eukaryotic cells embedded in a matrix composed, at least partially, of material synthesized by the sessile cells in the community.¹ It is essentially a survival strategy adopted by micro-organisms to resist environmental stress and antibiotics and disinfectants, which have been shown in vitro to be otherwise effective against the planktonic forms.^2,3 The biofilms thus formed have been shown in various in vitro, in vivo, and ex vivo studies to colonize indwelling medical devices such as catheters, intravenous cannula, stents, pacemakers, artificial heart valves, and orthopedic prosthesis.⁴ Once biofilms have formed, as shown in vitro, they are difficult to eradicate and are highly resistant to antibiotic therapy, often resulting in persistent and endemic populations.^5,6 A number of clinical studies have found such persisters are a source of subtle continuous and chronic infections within the host.⁷ Furthermore, colonization of indwelling medical devices results in deterioration, blockages, and loss of function, necessitating frequent replacement.⁷

In a cross-sectional study that involved 50 patients with chronic wounds and 16 with acute wounds recruited consecutively, James et al⁸ observed biofilms in 60% of the chronic wounds in comparison with 6% of acute wounds. Although the authors used light microscopy, biopsy culture, and polymerase chain reaction to detect bacterial types, they confirmed the presence of biofilms on detection of extracellular polymeric substance (EPS) and 3-dimensional confluent bacterial colonies using scanning electron microscope. Reviews of the literature^8,9 have speculated biofilm bacteria are associated with chronic wound infections and difficult to eradicate or kill because these organisms are encased in EPS and are able to resist phagocytic action and impede the action of the host immune system and antimicrobials. Moreover, lipopolysaccharide molecules in the bacterial cell wall have been shown to inhibit keratinocyte migration in laboratory study.¹⁰ All of these factors may contribute to the inability of wounds with biofilms to heal.

Pseudomonas aeruginosa is known for infection in chronic wounds¹¹ and one of the most common organisms responsible for hospital-acquired infections.¹² Given the degree of morbidity infections generate, it appears prevention is a more logical option than treatment. No technique currently is known to successfully prevent or control the formation of unwanted biofilms without causing adverse side effects, although some efforts using hydrogel-coated catheter pretreated with bacteriophage bred against Staphylococcus epidermidis¹³ and a phage cocktail against P. aeruginosa¹⁴ have shown promise. Because these are both in vitro studies, the outcome in clinical settings is not known.

Bacteriophages are viruses that infect bacteria and may provide a natural, highly specific, nontoxic, feasible approach for controlling several micro-organisms involved in biofilm formation.¹⁵ These are better than traditional antibiotics in that they are bacteria-specific and kill selectively without generating any systemic complications to the host. This high degree of specificity along with potential for deeper penetration through water channels in biofilms makes them an attractive mode of therapy. The fact bacteriophages multiply in the presence of sensitive bacteria (active therapy) allows limited delivery and dose control and prevents recurrent application. Furthermore, unlike antibiotics dose titration is not required in the presence of comorbid conditions such as renal or hepatic compromise. However, bacteriophages have been shown in vitro¹⁶ to form bacterial resistance in ways different from that of antibiotics such as adsorption resistance, restrictions, superinfection immunity, or abortive infection. The technology for harnessing this benefit successfully has been shown in laboratory settings,^17-19 although its application in clinical settings is still evolving²⁰ but not as yet convincing. In view of concerns regarding emerging multidrug-resistant bacteria,²¹ finding an alternative antibacterial approach should be pursued. The present study aims to evaluate the potential for use of novel phages for P. aeruginosa to control biofilm infection in vitro and in a murine wound model.

Overview. The present study was conducted in several steps, including P. aeruginosa identification and culture; bacteriophage characterization, reharvesting, and titration; enumeration of phage morphology and host range; P. aeruginosa biofilm formation on catheter; phage-biofilm interaction in vitro and in vivo; and identification of colony forming unit (CFU) and phage forming unit (PFU) counts after phage therapy of biofilms. The study was approved by the authors’ Institute Ethics Committee and the Institute Animal Ethics Committee for research. Care of the laboratory animals was provided according to the Indian Council of Medical Research Guidelines for use of Laboratory Animals in Medical Colleges 2001.²²

P. aeruginosa identification and culture. Multidrug-resistant strains of P. aeruginosa were procured from stock culture collected from the Burn Unit of the authors’ University Hospital. One hundred (100) strains from stock were streaked on Muller Hilton agar plates (Hi media, India) and incubated overnight. Smears were prepared on a glass slide from 1 loopful of bacterial colony, Gram-stained and observed under microscope, and subjected to oxidase and catalase tests. For further experiments, the bacterial strains were grown in Luria Bertani (LB) broth at 37˚ C and harvested while in the exponential phase of the growth cycle. The harvested bacteria were washed with saline (0.85% NaCl in distilled water) and resuspended in saline to achieve 10⁸ CFU/mL.

Phage characterization and re-harvesting and titration. Eleven (11) phages against P. aeruginosa were obtained from Bacteriophage Research Lab (BRL) of the Microbiology department of the University. The phages had been isolated from samples of various bodies of water including rivers, ponds, and sewers. The processing and reharvesting of the phages were done using the double agar overlay plaque assay technique,²³ and PFU was measured. Small-scale concentration of  Banaras Hindu University (BHU) bacteriophages was performed by spreading phages on the top-agar layer containing the respective host bacterium. The  BHU bacteriophage titer was analyzed as described by Adams.²⁴ Phages were characterized on the basis of their lytic profile, and a dendogram was obtained (see Figure 1), which depicts the level of similarity between the different phages used. When the 11 BHU phages were screened against the 40 MDR strains of P. aeruginosa, 11 strains were found to be lysed by 9 or more phage strains (see Table 1). These 11 strains then were re-subjected for testing against the  BHU phages, leading to the conclusion P. aeruginosa strain 10958 is lysed by all the BHU phages; it was selected for further experimentation.

Enumeration of phage morphology and host range. Polyethylene glycol-precipitated BHU bacteriophage particles were subjected to transmission electron microscopy for viral morphology. The host range specificity and lysis efficiency in screening tests against different P. aeruginosa strains were measured. Eleven (11) strains, which showed lysis by 9 or more phages, were rechecked by repeating the process.

P. aeruginosa biofilm formation on catheter sections. P. aeruginosa biofilms were formed on 0.5 cm, presterilized polyurethane catheter sections (6 French, single lumen) according to the method described by Cerca et al²⁵ with some alterations. Longitudinally cut open catheter sections were placed in a conical flask containing 6 mL of yeast peptone dextrose medium. A culture (10 µl) of P. aeruginosa strain 10958 with an optic density of 2.0 at 600 nm (corresponding to approximately 1.79 × 10⁸ cells/mL) was added and incubated at 37˚ C on an orbital shaker for 48 hours at a constant speed of 120 rpm with the change of medium every 12 hours. The number of cells present on these catheter sections was assessed after 48 hours of biofilm formation in order to estimate the infective dose. CFU counts varied from 2 x 10⁶ to 6 x 10⁶. Presence of P. aeruginosa biofilm was confirmed in vitro after crystal violet staining of the catheter sections and viewed under oil immersion microscopy (100x) and biochemical analysis (oxidase and catalase positive).

Phage-bacterial interaction in vitro. Biofilm were infected with a cocktail of 11 BHU phages (100 µl each). The catheter sections with biofilm were immersed twice in phosphate buffer saline (PBS) and placed in microcentrifuge tubes with 0.5 mL of tris-magnesium sulphate-gelatin (TMG) buffer and 0.5 mL of phage cocktail solution in a concentration of 10⁷ PFU/mL. The microcentrifuge tubes were incubated at 37˚ C. Control experiments were performed at the same conditions with the catheter sections after immersion in PBS in new microcentrifuge tubes with 0.5 mL of TMG and 0.5 ml of saline.

Phage-bacterial interaction in vivo. Male albino mice (25 ± 6 g) procured from the central animal house of the authors’ Institute were acclimatized in laboratory conditions for 7 days as per Indian Council of Medical Research guidelines.²² They were housed at 25˚ C, 45%–55% relative humidity, and 10:14 hours of light-dark cycle and had free access to rat pellets and water ad libitum. The animals were anesthetized with intraperitoneal ketamine (75 mg/kg). The back was shaved and disinfected with 0.5% chlorhexidine in 70% alcohol. A 10-mm transverse skin incision was made and a subcutaneous pocket was created between the skin and panniculus. Two catheter sections (with or without biofilm) were implanted in the subcutaneous pocket (see Figure 2). The incision was closed with interrupted 3-0 nylon stitches and disinfected with 0.5% chlorhexidine in 70% alcohol. The number of catheter sections to be used was assessed from a pilot study in which increasing numbers of catheter sections were implanted in 5 mice (ie, the first mouse had 1 catheter section while the fifth mouse had 5 sections). All developed pus formation in the subcutaneous pocket, but the first mouse showed complete resolution of pus, while the fourth and fifth mice perished after 48 hours of infection. The second and third mice showed pus formation without resolution for a week. Therefore, it was decided 2 catheter sections were optimum for the experiment.

The animals were randomly allocated into 4 groups of 6 mice each. In group A, the skin incision was followed by the creation of a subcutaneous pocket and 10 µl of sterile normal saline (NS) was injected daily into the pocket for 10 days. Group B had an incision with pocket creation in which uninfected catheter sections were placed; this group also had NS injections every day. In group C, catheter sections containing biofilm were placed in the pocket and the animals received daily NS injections. In group D, everything was similar to group C except instead of NS, 10 µl of the phage cocktail solution containing 10⁷ PFU/mL phages was injected in the subcutaneous pocket.

After 10 days, the animals were anesthetized and after local disinfection the subcutaneous pockets were explored and wounds were examined macroscopically for signs of infection. Catheter sections were removed in an aseptic technique and washed twice with PBS and analyzed for bacterial and phage colony counts.

CFU and PFU counts after phage therapy. After washing twice with PBS, the catheter sections were put in microcentrifuge tubes containing 0.5 mL of NS solution. The tubes were thoroughly mixed (vortexed 4 × 30 seconds) and serial dilutions were immediately performed in NS solution for CFU counts and in TMG buffer for PFU counts. For CFU counts, the samples were immediately plated on Mueller Hinton agar plates, and for PFU counts samples were immediately plated using the method described above for phage titration.

Data collection and statistical analysis. The data generated from phage titration, bacterial colony counts, and phage colony counts both in vitro and in vivo were retrieved by a laboratory assistant blinded to the study. The statistical program SPSS version 16.0 (SPSS Inc, Chicago, IL) was used for data entry and analysis. An independent sample t-test was used to compare categorical variables between 2 groups. Analysis of variance (ANOVA) along with post-hoc Bonferroni test was done to compare variables in more than 2 groups. A P value <0.05 was considered statistically significant.

Phage-bacterial interaction in vitro. At the time of primary culture on the blood agar plate after the action of cocktail of  BHU phages, no colony appeared after 24 hours of incubation. The CFU count on catheter sections was in the range of 10⁶ CFU/mL (see Table 2). Following treatment of these catheter sections with phage solution (10⁷ PFU/mL) for 24 hours, no growth was observed after 24 hours when these sections were plated on blood agar. Similarly, no plaque was observed when these catheter sections were processed for PFU counts.

Phage-bacterial interaction in vivo. When the mean baseline bacterial colony count was compared with that of group C (3.87 x 10⁶ versus 3.85 x 10⁵) and group D (3.87 x 10⁶ versus 3.52 x 10⁴), respectively, both were found to be significantly higher than the baseline value (P <0.001) (see Table 3). The bacterial counts also were found to be significantly higher in group C than that in group D (3.85 x 10⁵ versus 3.52 x10⁴; P <0.045), respectively. No phage was detected from any explanted catheter in group B and group C. However, in group D, phages were detected in range of 10⁷ to 10⁹, which was significantly higher than the baseline value (10⁷ versus 6.81 x 10⁸; P <0.05), indicating active multiplication of the phages in vivo (see Table 4). 

Although planktonic forms of bacteria usually are controlled by conventional doses of antibiotics, bacteria in biofilms are notably resistant to antibiotic therapy.⁵ Several in vitro studies^16,17,19 demonstrate the successful action of bacteriophages on biofilms. A number of in vivo studies^26-28 also have shown the potential of using phages to treat infectious diseases in animals and in humans infected with antibiotic-resistant bacteria.²⁹ In the present study, P. aeruginosa biofilms were challenged with the lytic  BHU phages both in vitro and in vivo. The data presented here show the potential of  BHU phages for controlling and reducing P. aeruginosa. Similar observations of reduction in bacterial cell count (between 10³ and 10⁵) also have been reported in a number of experimental models.^4,30-33 Current findings indicate promising potential of  BHU phages as controlling agents for P. aeruginosa.

Although the formation of biofilms on catheters immersed in the growth media are unlike conditions observed in a variety of biological environments, it is a simple, rapid, and reproducible method and makes it easy to assess the influence of different parameters required for phage therapy. However, potential concerns for the use of this therapy must be carefully considered, such as narrow host range of phages, the development of resistance of host bacteria to phages during therapy, potential of inactivation of phages by patient’s immune system, and the safety of phage application in humans.³⁴

In the present study, a significant bacterial load reduction was observed in the test group. The reduction in the control group is probably due to the intrinsic immune response activated following implantation of the infected catheter sections. Therefore, the significant reduction in the test group is likely due to a combined effect of phage and host immune system-mediated bacterial death. Catheter sections were completely sterile in 1 mouse in the test group. This may be due to a widespread phage-induced bacterial kill supplemented with strong immune response. Furthermore, it also may be argued that a weak immune response toward the phage particles allowed abundant phage proliferation and subsequent eradication of the bacteria. Interestingly, in the same mouse, no phage was detectable on the catheter sections. A possible explanation of this finding may be that because all of the bacteria on this specimen were eradicated by the phages, in the absence of a host to act upon, the phage particles became inactive and were subsequently inactivated by the host immune system. This hypothesis is supported by the observation from the in vitro aspect of the study, which also demonstrates absence of phages after complete eradication of bacteria.

Although the phage cocktail was given in the dose of 10⁷ PFU/mL daily for 10 days (passive therapy), it is interesting to note phages were found in abundance (10⁸–10⁹ PFU/mL) on the catheter sections when removed after 10 days. This indicates the phages were actively multiplying in the presence of bacteria (active therapy), which is possible only when the bacteria remain sensitive to the phages. Thus, it may be extrapolated that continuing phage therapy for a longer duration could have eradicated bacterial population on the catheter sections. However, further research is needed to confirm.

It is interesting to observe the huge difference in efficacy of phage action in the in vitro and in vivo model. A number of factors may be responsible for this discrepancy. It has been proposed the P. aeruginosa biofilm EPS may impede penetration by trapping the phage particles or may produce proteolytic enzymes that cause bacteriophage inactivation.³⁵ This may be overcome by the phage elaboration of EPS degrading depolymerase in P. aeruginosa biofilms.³¹ However, this is unlikely to be the cause in the present study because the biofilm was effectively penetrated in vitro. Alternatively, active local inflammatory response could have acted directly against the viruses by producing antibodies against the viral proteins, leading to their neutralization. Issues such as temperature and culture medium composition may be responsible for the difference between in vitro and in vivo results.³⁶  In in vitro study, Sillankorva et al¹⁸ demonstrated a temperature of 26˚ C is favorable for the best effects of phage, which is unlikely to be achieved in the animal body. Similarly, a favorable medium is also beneficial for bacterial kill,¹⁸ which is probably responsible for the better in vitro outcome. The different metabolic status of the biofilm bacteria may affect phage-induced lysis; it has been observed that Escherichia coli in exponential phases of growth are lysed faster than those in the later growth phase.³⁷ Only a marginal reduction of susceptibility of old P. aeruginosa biofilm to phage-induced killing has been observed.³² The detection of the actual factor(s) responsible for the reduced phage efficacy in vivo will help researchers and clinicians refine this natural form of therapy to a newer level in healing of wounds with biofilms.

The phage therapy may be active or passive depending on the mode of use. In active therapy, a single small dose of phages can be useful in circumstances when bacteria are multiplying fast, which leads to active multiplication of the phages as well. But when bacteria are slow multipliers, endogenous multiplication of the viral particles is inadequate for bacterial kill. In these cases, phages may be required to be given repeatedly in large doses (passive therapy). Both slow-multiplying and fast-multiplying bacteria are inhabitants of a biofilm. Accordingly, a single small dose of phages may be inadequate to completely eradicate the slow multipliers. Therefore, phage therapy used repeatedly over a prolonged period of time is likely to act against both types of bacterial population in the biofilm.

The present study is not without limitations. The murine model does not closely mimic human wound healing, wound infection, and wound-biofilm interaction. Also, it is not known how the immune system of rodents reacts to the biofilm; these animals demonstrate strong immunogenic reaction and are mostly resistant to wound infection. Therefore, the behavior of the Pseudomonas biofilm in the presence of bacteriophages may differ in human wounds. However, in vivo models are better than in vitro models because the phage-biofilm interaction can be studied in the midst of host immune response.

In addition, the phage-bacteria interaction in the present study is limited to the effect of the phages on the bacteria only. It is not known to what extent this interaction influences the wound healing process. The passive mode of phage delivery is possible in well-controlled laboratory conditions that may not always be possible in the clinical setting. Although phage multiplication was observed after passive delivery, it is not known to what extent the response will be clinically significant.

In this study, P. aeruginosa organisms in the biofilm were sensitive to the respective phages in vivo and in vitro. In the presence of the biofilm, phages were capable of multiplication and bacterial kill and in combination with the host immune system were found to lead to significant reduction in bacterial population locally. The difference between in vitro and in vivo results indicates interplay of factors at the tissue level, which may slow bacterial kill. Repeated doses of phage application remain a promising technique to overcome these factors in wound biofilms. In the future, devising a technology for delivering phages to a chronic wound with biofilm would be challenging and may usher a new era of biological dressings for biofilms. 

Dr. Basu is Associate Professor of Surgery; Dr. Agarwal is a surgical resident cum research scholar; Dr. Bhartiya is Assistant Professor of Surgery; Dr. Nath is Professor of Microbiology; and Dr. Shukla is Professor of Surgery, The Departments of General Surgery and Microbiology, Institute of Medical Sciences, Banaras Hindu University, Varanasi, India. Please address correspondence to: Dr. Vijay Kumar Shukla, Department of General Surgery, Institute of Medical Sci-ences, Banaras Hindu University, Varanasi 221005, India; email: vkshuklabhu@gmail.com.