A cost-effective combination of Rose Bengal and off-the-shelf cationic polystyrene for the photodynamic inactivation of Pseudomonas aeruginosa

Two new photoactive materials have been prepared, characterized and tested against Pseudomonas aeruginosa bacteria (planktonic suspension). The synthesis of the polymeric photosensitizers can be made at a multigram scale, in few minutes, starting from inexpensive and readily available materials, such as Rose Bengal (photosensitizer) and ion exchange resins Amberlite® IRA 900 (macroporous) or IRA 400 (gel-type) as cationic polystyrene supports. The most notable feature of these systems is their notable bactericidal activity in the dark (4-5 log10 CFU/mL reduction of the population of P. aeruginosa) which becomes enhanced upon irradiation with visible light (to reach a total reduction of 8 log10 CFU/mL for the macroporous polymer at a fluence of 120 J/cm2 using green light of 515 nm).


Introduction
Nosocomial, or hospital-acquired, infections are a rampant problem for national health services, causing yearly hundreds of thousands of deaths worldwide. [1] Signs of alert have been increasing during the last decade and cannot be overlooked. In 2008 Rice coined the acronym ESKAPE to describe the group of bacteria causing the most frequent nosocomial infections (Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumannii, Pseudomonas aeruginosa and Enterobacter species), [2] and the World Health Organization (WHO) has recently warned about the need for new therapeutic approaches to treat antibiotic-resistant bacteria. [3] Bacteria identified by the WHO as serious threats for the human health in the whole world include 12 species. In this list, P. aeruginosa is classified within the most dangerous class "Priority 1: critical". The number of adverse clinical scenarios associated to this pathogen includes not only chronic respiratory infections, sepsis and lung damage in children with cystic fibrosis, but also infections related to surgical wounds / catheters, recurrent pneumonia, sepsis, urinary tract infections and a long list of other problems in immunocompromised patients.
Despite the concentrated efforts of the scientific community to develop alternatives to the existing antibiotics, [4,5] the current realistic therapeutic options are very scarce and the most sensible approach to tackle this problem is the prevention of contagion.
Since adherence of microorganisms to surfaces leads to the propagation of diseases by simple contact, much effort has been devoted recently to the development of materials with antimicrobial activity. [6] One advanced strategy to deal with the problem of microbial infections is the so-called antimicrobial photodynamic inactivation (aPDI). This preventive / therapeutic approach, related conceptually to the photodynamic therapy (PDT) of cancer, [7] relies on the absorption of light by a photosensitizing molecule which, in the presence of oxygen, is able to generate reactive oxygen species (ROS) capable of killing (or avoiding the growth of) bacteria, virus, fungi, protozoa, and a varied number of other infective agents. [8][9][10][11] One of the foreseen applications of aPDI is the prevention of the spreading of microorganism causing diseases by using materials with self-sterilizing surfaces activated by the light used for room illumination. [12,13] The list of elements in a healthcare facility that bacteria can colonize is almost unlimited, including catheters, stethoscopes, armpit thermometers, computer screens, cell phone cases, door handles, wall paints, fabrics, etc. The field of aPDI using immobilized photosensitizers has been reviewed by several groups, [14][15][16][17]. Most of the described photosensitizers and many of the polymeric materials use to support them are difficult to synthesize on a large scale, making their use in an extended surface a practical challenge. Hence, from a practical point of view, both photosensitizer and support must be readily accessible and they should be combined in a manner involving simple operational procedures. Here we would like to report two photoactive materials combining the well-known photosentitizer Rose Bengal (RB) with two cationic exchange resins differing in their nature, a macroporous (Amberlite® IRA-900) and a gel-type (Amberlite® IRA-400) cationic polystyrene. The photoactive polymers here presented display outstanding aPDI activity against P. aeruginosa when irradiated with visible light, leading to eradication of the population of this microorganism, starting from a concentration of 10 8 colony forming units (CFU) / mL. Many materials have been reported to date within the field of aPDI, but the one here described stands out for its very favorable cost-benefit ratio against one of the top threats identified by the WHO.

Synthesis of the supported photosensitizers
The preparation of the RB loaded polymers was performed by mixing 50 mL of Rose Bengal sodium salt (RB, Sigma-Aldrich) solution in absolute EtOH (30 µM) with 1 g of Amberlite® IRA-900 or Amberlite® IRA-400 (both from Sigma-Aldrich, chloride form; ion exchange resins were previously washed with ethanol and vacuum dried overnight). The suspension was stirred smoothly at room temperature. Loading was followed by UV-Vis (JASCO V-630 UV-vis spectrophotometer), monitoring the decrease of the absorbance at 560 nm (quantitative RB loading after 70 min of stirring). The beads were filtered off on a sintered glass filter and washed with EtOH. Finally, the polymers were dried under vacuum at 55 ºC overnight.
Characterization of the supported photosensitizers was performed by fluorescence spectroscopy, porosimetry, and scanning electron microscopy (see below).

Fluorescence Spectroscopy.
The steady-state fluorescence of the samples was recorded with a Varian Cary-Eclipse spectrofluorometer, using quartz cell cuvettes. The polymeric samples were measured in the solid state whereas spectra of RB was recorded in solution. In the case of the polymers, five measurements were carried out for each sample in different areas of the solid and the results were averaged. Excitation wavelength was set at 555 nm. To record the excitation spectra, the emission wavelength was set at 575 nm (RB in EtOH) or 600 nm (RB@Pmp and RB@Pgel).

Porosimetry.
Measurements were carried out on a Micromeritics ASAP 2020 gas porosimeter equipped with a SmartVac degasification system. 1 g of RB@Pmp and RB@Pgel were washed with ethanol and dried for 24 h in a vacuum oven at 60 ºC before performing the analysis.

Scanning Electron Microscopy.
Images were acquired using a JEOL 7001F scanning electron microscope equipped with a digital camera. The samples were placed on top of a tin plate and sputtered with Pt in a Bal-tec SCD500 sputter coater. Vis spectrophotometer. The kinetic traces were fitted to a pseudo-first order model (ln C/C0 =kobs · t, where C is the concentration of DMA at a certain time t and C0 is the initial concentration of DMA; for low concentrations of substrate, the concentration ratio C/C0 is considered to be equivalent to the absorbance ratio A/A0).

Photooxygenation cycles.
Irradiations were performed as indicated above (20 minutes to obtain the rate constant),.
After complete reaction of the DMA substrate (1h irradiation), the polymer was easily separated by filtration and re-suspended in 50 mL of 1x10 -4 M DMA freshly prepared solution in the corresponding solvent, and irradiated again, up to seven cycles.

Results and discussion
Amberlite ® IRA-900 is a popular commercial ion exchange resin used in catalysis, chromatography and environmental remediation. [18] It consists basically of macroporous crosslinked polystyrene containing trimethylammonium groups (with chloride as counterion). On the other hand, Amberlite ® IRA-400 presents a lower degree of cross-linking which does not allow a permanent porosity, but can be swollen in appropriate organic solvents, leading to gel-like materials. In the past we have employed synthetic macroporous resins with RB attached covalently and studied their photochemical and photobiological properties, including their PDT activity against melanoma cells. [19] In order to facilitate the synthetic methodology, recently we turned our attention to commercial Amberlite ® resins IRA 900 and 400 (called from now on Pmp and Pgel, respectively). The possibility of carrying out a simple anion exchange (replacing chloride anion by any anionic photosensitizer) makes them very appropriate for the production of large amounts (multigram scale) of photoactive materials in a very straightforward manner. In the field of photocatalysis this is a common method to produce, for instance, photocatalytic materials for oxygenations. [20] In previous reports we used this ion exchange strategy to explore the possibilities of molybdenum-based photosensitizers bound to cationic polymers. [21,22] We found that both supports, Pmp and Pgel, are appropriate for the development of photobactericidal materials, against planktonic cultures of S. aureus and P. aeruginosa, with a better performance of the Pmp resin over the Pgel one. In order to expand the range of useful photosensitizers employing this strategy, we decided to use RB, providing its dianionic nature, and considering that it is one of the most studied photo-antimicrobials so far, that it is safe for human use and also that its cost is very low, which could be a practical advantage for future application in real medical contexts.
The above mentioned exchange process is depicted in Figure 1 for the case of RB. From the experimental point of view, the procedure is very straightforward: the addition of 1 g of polymer to a 50 mL solution of RB (30 µM) in ethanol leads to complete discoloration of the stirred solution after 70 min. The exchange kinetics are very similar for both resins and leads to a red solid that is isolated by filtration. In Figure 1, pictures of resins before and after loading with RB are shown. It must be noted that both resins were loaded with 1.5 mg of RB per gram of polymer, but this loading is not optimized, and could be improved in the future.  The structural properties of the materials were checked by nitrogen porosimetry (see Figure   S1 in Supplementary Material). This technique showed much higher values of specific surface for RB@Pmp (21.96 m 2 g -1 ) than for RB@Pgel (0.0065 m 2 g -1 ), as expected. The different nanostructure of the matrices were better visualized using Scanning Electron Microscopy (SEM, Figure S2 in Supplementary Material). Images of both materials magnified 50 times show spherical beads very similar in appearance, but a closer look at the nanoscale level (50,000 times magnification), affords a much different view: RB@Pmp is comprised of a structured surface with large and small pores whereas RB@Pgel has a more dense appearance.
The ability of RB to induce bacterial killing is based on the production of singlet oxygen ( 1 O2) upon irradiation with visible light. [25] In order to demonstrate the production of 1 O2 species and to compare both photoactive polymers, a benchmark reaction like the oxygenation of 9,10dimethylanthracene (DMA) was used (Figure 3). In this reaction, the absorption of the DMA probe at 350-400 nm is progressively bleached upon reaction with 1 O2 to form the corresponding nonabsorbing endoperoxide (DMA·O2). Monitoring the absorption of DMA vs the irradiation time allows the determination of the pseudo-first order kinetic constant (k), which can be used as a relative guide to estimate the photoactivity of the material (Figure 3b). [26] In our case, the two polymers under study were tested using DMA as 1 O2 trap, in different solvents. As it can be seen in Figure 3c, the macroporous resin performs the oxygenation reaction in a more effective way than the gel-type material, especially in acetonitrile (ACN). The effect of the solvent on the kinetics of the reaction is attributable to the longer lifetime of 1 O2 in acetonitrile as compared to ethanol and water, as reported in the literature. [27] Control irradiations using polymers without photosensitizer afforded minimal values of k (2 -8 x 10 -3 min -1 ) as a result of selfphotosensitization (DMA absorption tail around 400 nm could be the cause of a slight excitation of the probe leading to residual 1 O2 production) In order to test the resistance of the materials to photobleaching, cycles of irradiations, using the reactivity of DMA as a test, were conducted. As it can be seen in Figure 4, both RB@Pmp and RB@Pgel are effective for seven cycles of oxygenation reactions, with low loss of activity. The polymeric matrices containing RB were irradiated in the presence of cultures (10 8 CFU / mL initial concentration) of Pseudomonas aeruginosa ( Figure 5 and Figure S3 in Supplementary Material). A high killing efficiency was observed with both materials. This result is very remarkable since it is well known that this Gram-negative species is particularly resistant to photodynamic treatments, due to their double protective membrane structure. This is even more striking when considering that it is a well-stablished fact that anionic photosensitizers like RB are not particularly effective against Gram-negative species. [28] As a matter of fact, it has been reported that RB alone is very inefficient against P. aeruginosa. [29] Survival data at a fluence of 120 J/cm 2 can be seen in Figure 5. The effect of RB@Pmp is very remarkable since no CFU can be detected after the irradiation period (8 log10 CFU/mL reduction). In the case of RB@Pgel the result is also very notable, with an estimated 7 log10 CFU/mL reduction (a longer irradiation dose of 140 J/cm 2 led also to the eradication of this bacteria, see Figure S3 in Supplementary Material).
But the action of 1 O2 is not the only factor to be considered in order to explain these results since a remarkable reduction of viability is also observed with resins Pmp and Pgel (with no RB encapsulated inside, third and fourth bars in Figure 5  In this graph C is control, only microbial suspension and RB means polymer with Rose Bengal. The red arrow indicates 99.999999% reduction of bacterial population. The error bars represent the standard deviation calculated for three measurements. One-way ANOVA statistical analysis: * p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001. Importantly, the results here reported are consistent with the previously reported effect of similar materials developed in our laboratories. As indicated previously, Pmp and Pgel loaded with a photoactive hexanuclear molybdenum cluster (Mo6) were also efficient against P. aeruginosa although with a more limited effect. Hence, at 120 J/cm 2 , Mo6@Pmp was able to reduce the population of this bacteria ca. 3 log10 CFU/mL whereas Mo6@Pgel did it only by ca. 1 log10 CFU/mL. [22] Probably, a combination of factors would explain the performance of RB@Pmp towards P. aeruginosa here reported: the appropriate distribution of ammonium groups on the polymer, the right nanostructure of the surface, the ability of the surface to bind (and sequester) the bacteria, the high yield of 1 O2 generated by RB at the employed concentration and maybe the manufacturing process (and even commercial batch) of the resin. Since those factors are not optimized yet, it is expected that a deeper knowledge of these materials will lead to an improved balance between them and hence it will allow obtaining polymeric materials with even a better performance.
Besides, it must be noted that ion-exchange resins must be operated within a certain temperature range in order to maintain their properties [35], so this factor must also be taken into account.
In order to appreciate the value of the reported 8 log10 CFU /mL reduction, a comparison to recently described photosensitizing systems active against P. aeruginosa is presented in Table   1. We are aware that different experimental setups make the comparison very difficult (concentrations, bacterial types, excitation wavelengths, light fluences, etc). The presented data do not constitute a systematic revision of the current literature, but an overview of recent results in order to illustrate the value of the cost-effective proposal here presented. We have included only studies in planktonic suspension, not in biofilms, [36] which use a different methodological approach. As can be seen in Table 1, the reductions of the bacterial populations of P. aeruginosa caused by both soluble and supported photosensitizers range typically from 1 to 7 logs10 CFU / mL. Values of 8 are obtained for methylene blue encapsulated in porous silica nanoparticles [49] and for chitosan used as carrier of Toluidine blue O. [43] However, according to the authors, [43,49] the action of these supports in the dark against P. aeruginosa is negligible. By contrast, Pmp and Pgel polymers already display a dark antibacterial activity (4-5 log10) which would make those polymers still bactericidal against P. aeruginosa even when the material is not illuminated.

Conclusions
In summary, the aPDI capacity of two new materials, RB@Pmp and RB@Pgel, made from easily accessible sources, against P. aeruginosa bacteria, has been explored. The excellent results obtained make the materials here reported a good tool for the photo-inactivation of one of the most important threats identified by the WHO. We would like to emphasize on the fact that despite anionic photosensitizers have been relegated in favor of cationic ones for the inactivation of Gramnegative bacteria, if such anionic molecules are combined with simple cationic materials like exchange resins Amberlite ® IRA-900 or IRA-400, they could be actually highly effective against resistant microorganism. We expect that the proper combinations of support and photosensitizer, tailored for each pathogen, will be discovered in the coming years, which will help to stop the spreading of pathogenic microorganisms.

Declaration of competing interest
The authors declare no competing financial interest.