**Nanovectorized Microalgal Extracts to Fight** *Candida albicans* **and** *Cutibacterium acnes* **Biofilms: Impact of Dual-Species Conditions**

**Virginie Lemoine 1, Clément Bernard 1, Charlotte Leman-Loubière 2, Barbara Clément-Larosière 3, Marion Girardot 1, Leslie Boudesocque-Delaye 2, Emilie Munnier <sup>4</sup> and Christine Imbert 1,\***


Received: 30 April 2020; Accepted: 23 May 2020; Published: 26 May 2020

**Abstract:** Biofilm-related infections are a matter of concern especially because of the poor susceptibility of microorganisms to conventional antimicrobial agents. Innovative approaches are needed. The antibiofilm activity of extracts of cyanobacteria *Arthrospira platensis*, rich in free fatty acids, as well as of extract-loaded copper alginate-based nanocarriers, were studied on single- and dual-species biofilms of *Candida albicans* and *Cutibacterium acnes*. Their ability to inhibit the biofilm formation and to eradicate 24 h old biofilms was investigated. Concentrations of each species were evaluated using flow cytometry. Extracts prevented the growth of *C. acnes* single-species biofilms (inhibition > 75% at 0.2 mg/mL) but failed to inhibit preformed biofilms. Nanovectorised extracts reduced the growth of single-species *C. albicans* biofilms (inhibition > 43% at 0.2 mg/mL) while free extracts were weakly or not active. Nanovectorised extracts also inhibited preformed *C. albicans* biofilms by 55% to 77%, whereas the corresponding free extracts were not active. In conclusion, even if the studied nanocarrier systems displayed promising activity, especially against *C. albicans*, their efficacy against dual-species biofilms was limited. This study highlighted that working in such polymicrobial conditions can give a more objective view of the relevance of antibiofilm strategies by taking into account interspecies interactions that can offer additional protection to microbes.

**Keywords:** antibiofilm; antimicrobial agent; bacteria; fungi; polymicrobial biofilm; microalga; free fatty acids; encapsulation

### **1. Introduction**

Biofilms are involved in numerous diseases, both superficial and systemic, for instance those affecting the oral cavity, skin or related to an implanted medical device. They can be single species, but most often they are polymicrobial and contain both fungi and bacteria. For example, dermal wounds are colonized by aerobic and anaerobic bacterial and fungal species, most of them belonging to resident microbiota of the surrounding skin, oral cavity and gut, or from the external environment [1]. It has been shown that 60% of chronic wounds exhibit a biofilm which is a major factor in delayed

wound healing [2–5]. Also, it is considered that *Candida* spp. are among the primary causes of delayed healing and infection in both acute and chronic wounds, especially those of a surgical nature [6,7]. Literature data suggest that *Candida* spp. rarely colonize human skin but can cause infection especially in specific conditions such as immune deficiency, diabetes or after antibiotic use [8,9].

Gram-positive bacteria *Cutibacterium acnes* (formerly *Propionibacterium acnes*) [10] are a main colonizer and inhabitant of the skin [11,12] and a good biofilm former, producing both single species and polymicrobial biofilms [13–16]. Its involvement in chronic skin disease such as acne vulgaris is very well-known and this species is also occasionally involved in non-skin-related infections such as prosthetic joint infections, some of them being related to the formation of a biofilm [17,18]. As for many other microbial species, sessile *C. acnes* cells as well as *Candida albicans* cells have been shown to be more tolerant to conventional antibiotics than their planktonic counterparts [19–21].

Our team recently showed that *C. acnes* and *C. albicans* can form dual-species biofilms, with *C. acnes* adhering to both hyphal and yeast forms of *C. albicans* [15]. The presence of metabolically active *C. albicans* cells enhanced the early growth of *C. acnes* under aerobic conditions, while no influence was observed in anaerobic conditions. We also recently demonstrated that the co-presence of these species in biofilms influenced their sensitivity to micafungin, a major conventional antifungal agent. Actually, *C. acnes* was shown to protect *C. albicans* cells from the effect of micafungin in dual-species biofilms [22]. Along the same lines, Montelongo-Jauregui et al. showed that the resistance of *C. albicans* to amphotericin B and caspofungin, as well as the resistance of *Streptococcus gordonii* to clindamycin were increased due to a dual-species biofilm produced by *C. albicans* and with *Streptococcus gordonii*, compared to single-species conditions [23]. Therefore, a double issue should be thus observed—biofilm lifestyle causes itself a decreased susceptibility to antimicrobial agents and the polymicrobial nature of the biofilm can make this lack of susceptibility even worse.

Biofilms are infectious reservoirs and the most effective way to prevent biofilm-related infections requires the eradication of these complex microbial structures, that is their detachment, their disorganization and the killing of all released microbial cells, with these three events needing to be concomitant. Unfortunately, the available antimicrobial conventional molecules fail to reach this challenging goal.

Free fatty acids (FFAs) are physiological antimicrobial agents occurring on skin, exhibiting a wide antimicrobial spectrum (antibacterial, antifungal, antibiofilm ... ) [24,25]. Microalgae have been well-described as abundant sources of lipids and especially FFAs [26,27]. Those FFAs, especially polyunsaturated (PUFAs), may also represent a potential source of topical drugs against polymicrobial biofilms. Indeed, a previous screening of 29 FFAs based on topical antibacterial activity highlighted that PUFAs were among the most active [25]. *Arthrospira platensis* (formerly *Spirulina platensis*) appeared as a good model among all microalgae as it was the most studied microalgae with a well-known FFA profile.

Due to their lipid nature, FFAs are not able to penetrate the biofilm made of highly hydrophilic exopolysaccharide. Recently, nanosized systems showed their ability to vectorize active molecules in biofilms. Core-shell nanosystems, with a hydrophilic shell and a lipophilic core, seem to be very appropriate vectors for low-polarity active molecules [28,29]. Alginate-based nanocarriers, nanosystems made of a triglyceride core and an alginate gel shell, were shown to be efficient to vectorize FFAs in *C. albicans* biofilm [30]. The reproducibility of the preparation, the stability of the systems and their FDA-approved ingredients constitute key advantages for their use in dermatology. Moreover, they were shown to be stable in dermatological preparations [31].

We previously developed a proof of concept of the potential of *A. platensis* extracts and alginate-based nanocarriers combination as a possible strategy to fight *C. albicans* single-species biofilms. The antibiofilm strategies are all the more innovative and promising in that they are able to act on taxonomically distant and diverse microbial species, because of the polymicrobial nature of most biofilms developing in humans.

Thus, this current study aimed to obtain lipid extracts from Cyanobacteria *A. platensis*, to develop extract-loaded copper alginate-based nanocarriers able to carry a lipid extract and to evaluate the antibiofilm activity of these lipid extracts nanovectorized or free, against both single-species fungal and bacterial biofilms and interkingdom dual-species biofilms.

### **2. Results**

### *2.1. A. platensis Extraction*

*A. platensis* biomass was extracted using two sustainable solvents—EtOAc and DMC. Resulting extracts were enriched in lipids (Table 1), with a closely related FFA profile. Both extracts contained mainly ω6 PUFA, i.e., linoleic and γ-linolenic acid (more than 60% of total FFAs) (Table 2). Also, lipophilic dyes (chlorophyll and carotenoids) were co-extracted (Table 2), but their content remained low, highlighting again the good selectivity of these solvents towards lipids.

**Table 1.** *A. platensis* extracts composition (total lipids, chlorophylls, carotenoids).


Data are shown as mean ± SD; *n* = 3.


**Table 2.** Free fatty acid (FFA) ratios in *A. platensis* extracts, in relative percentage of total FFA.

nd = nondetected; *n* = 1.

### *2.2. A. platensis Extracts Vectorization*

Extract-loaded ANCs were prepared with EtOAc extract and DMC extract. Physicochemical characteristics are shown in Table 3.

**Table 3.** Physicochemical characteristics of extract-loaded alginate-based nanocarriers.


Extract-loaded ANCs show similar size and surface potential as empty ANCs with a pure Labrafac®® WL 1349 core. The polydispersity index lower than 0.2 shows a monodispersity of the suspensions, guaranteeing the reproducibility of the dosage. The negative surface charge of the nanocarriers participates to the colloidal stability of the nanocarriers and should not limit their interaction with the biofilms. Indeed, even if biofilms are generally considered negatively charged and could thus bind more easily to cationic nanoparticles [32], several negatively charged systems displayed antibiofilm efficacy [28,33]. The native ANC suspension shows a concentration in *A. platensis* extract of ~1 mg/mL.

### *2.3. Ability of A. platensis Extracts to Prevent Biofilm Formation*

In single-species conditions, EtOAc extract used at 0.2 mg/mL displayed a significant (*p*=0.0001) but very limited antibiofilm formation effect against *C. albicans* (24.4% inhibition) (Figure 1A). This extract used at 0.1 mg/mL was not active against *C. albicans*, and DMC extracts (at both 0.1 and 0.2 mg/mL) as well. Both EtOAc and DMC extracts significantly reduced the growth of *C. acnes* biofilms, regardless of the tested concentrations—inhibition ranged between 66.0% and 78.4% (EtOAc extract, *p* ≤ 0.003) and between 67.6% and 86.2% (DMC extract, *p* ≤ 0.0008) (Figure 1C). However, no real conclusion can be made in the case of EtOAc (0.1 and 0.2 mg/mL) and DMC (0.1 mg/mL) as the error bars are very high.

**Figure 1.** Ability of *Arthrospira fusiformis* extracts to prevent biofilm formation. Single-species biofilms (*C. albicans*) (**A**); *C. albicans* concentration obtained in dual-species biofilms (*C. albicans* + *C. acnes*) (**B**); single-species biofilms (*C. acnes*) (**C**); *C. acnes* concentration obtained in dual-species biofilms (*C. albicans* + *C. acnes*) (**D**). Results are expressed as mean ± SD. \* *p* < 0.005: test condition vs. BHI-control (biofilms treated with BHI only).

In dual-species conditions, neither EtOAc nor DMC extract solutions were able to reduce the biofilm formation of *C. albicans* and no reduction was observed in the fungal and bacterial populations.

### *2.4. Ability of A. platensis Extracts to Eradicate Preformed Biofilms*

None of the extracts, whatever the tested concentration, had any effect on *C. albicans* or *C. acnes* preformed single-species or dual-species biofilms. No reduction was observed in the fungal and bacterial populations after a 24 h treatment (Figure 2).

**Figure 2.** Ability of *Arthrospira fusiformis* extracts to eradicate preformed biofilm. Single-species biofilms (*C. albicans*) (**A**); *C. albicans* concentration obtained in dual-species biofilms (*C. albicans* + *C. acnes*) (**B**); single-species biofilms (*C. acnes*) (**C**); *C. acnes* concentration obtained in dual-species biofilms (*C. albicans* + *C. acnes*) (**D**). Results are expressed as mean ± SD. \* *p* < 0.005: test condition vs. BHI-control (biofilms treated with BHI only).

### *2.5. Ability of A. platensis Extracts Encapsulated in Alginate-Based Nanocarriers to Prevent Biofilm Formation*

In single-species conditions, empty nanocarriers inhibited the growth of *C. albicans* biofilms by 51.55% (0.1\_emptyNC, *p* = 0.001) or 54.14% (0.2\_emptyNC, *p* = 0.0002), while they had no effect on *C. acnes* biofilms (Figure 3A,C). Nanocarriers loaded with extract solutions at 0.2 mg/mL (0.2 mg/mL\_EENC) inhibited the growth of *C. albicans* biofilms by 51.35% (EtOAc, *p* = 0.0031) or 43.77% (DMC, *p* = 0.0021) while those loaded with extract solutions at 0.1 mg/mL (0.1 mg/mL\_EENC) had no significant influence. Regarding the growth of *C. acnes* biofilms, only nanocarriers loaded with EtOAc

extract at 0.2 mg/mL and DMC extract at 0.1 mg/mL demonstrated a weak inhibitory activity of 22.48% (*p* = 0.0016) and 32.74% (*p* = 0.0004), respectively (Figure 3C).

**Figure 3.** Ability of *Arthrospira fusiformis* extracts encapsulated in alginate nanocarriers to prevent biofilm formation. Single-species biofilms (*C. albicans*) (**A**); *C. albicans* concentration obtained in dual-species biofilms (*C. albicans* + *C. acnes*) (**B**); single-species biofilms (*C. acnes*) (**C**); *C. acnes* concentration obtained in dual species biofilms (*C. albicans* + *C. acnes*) (**D**). Results are expressed as mean ± SD. \* *p* < 0.005: test condition vs. BHI-control (biofilms treated with BHI only).

In dual-species conditions, nanocarriers loaded with extract solutions did not limit the growth of either *C. albicans* or *C. acnes* in biofilms. Only empty nanocarriers (0.1\_emptyNC, *p* = 0.0046) displayed a weak activity but were not significant (*p* > 0.005) against *C. albicans* growth (21.0%) and no reduction was observed on *C. acnes* population (Figure 3B,D).

### *2.6. Ability of A. platensis Extracts Encapsulated in Alginate-Based Nanocarriers to Eradicate Preformed Biofilms*

In single-species conditions, empty nanocarriers inhibited preformed biofilms of *C. albicans* by 58.7% (0.1\_emptyNC, *p* < 0.0001) or 76.69% (0.2\_emptyNC, *p* < 0.0001), whereas they had no effect on *C. acnes* biofilms (Figure 4A,B). Whatever the conditions, all nanocarriers loaded with extract solutions inhibited preformed single species *C. albicans* biofilms (*p* < 0.0001) by at least 55%; nanocarriers loaded with EtOAc extracts inhibited biofilms by 76.9% (0.1 mg/mL\_EENC-EtOAc) and 62.35% (0.2 mg/mL\_EENC-EtOAct) whereas those loaded with DMC extracts induced a 55.69% (0.1 mg/mL\_EENC-DMC) and a 77.32% (0.2 mg/mL\_EENC-DMC) inhibition. On the contrary, whatever the conditions, both empty and loaded nanocarriers failed to significantly reduce an already formed single-species biofilm of *C. acnes* (*p* > 0.005) (Figure 4C).

**Figure 4.** Ability of *Arthrospira fusiformis* extracts encapsulated in alginate nanocarriers to eradicate preformed biofilm. Single-species biofilms (*C. albicans*) (**A**); *C. albicans* concentration obtained in dual-species biofilms (*C. albicans* + *C. acnes*) (**B**); single-species biofilms (*C. acnes*) (**C**); *C. acnes* concentration obtained in dual-species biofilms (*C. albicans* + *C. acnes*) (**D**). Results are expressed as mean ± SD. \* *p* < 0.005: test condition vs BHI-control (biofilms treated with BHI only).

In dual-species conditions, empty nanocarriers as well as those loaded with extract solutions induced inhibition always less than 28% of already formed dual-species biofilms, whatever the target population (*C. albicans* or *C. acnes*) (Figure 4B,D). Significant *p*-values demonstrating an inhibition of the *C. albicans* population were only observed in the case of 0.2\_emptyNC (inhibition: 27.5%), 0.2 mg/mL\_EENC-EtOAc (inhibition: 25.3%) and 0.2 mg/mL\_EENC-DMC (inhibition: 26.3%) (Figure 4B).

### **3. Discussion**

The results are in accordance with those previously obtained when studying the activity of EtOAc extract at 0.2 mg/mL on *C. albicans* biofilms [30]. The ability of EtOAct extract at 0.2 mg/mL to inhibit *C. albicans* biofilms growth evidenced by the significant decrease in the number of cells forming treated biofilms (FCM approach) (Figure 1A) agrees with previous results showing that this extract was able to reduce the metabolic activity of *C. albicans* forming treated biofilms (XTT method). However, EtOAc extract at 0.1 mg/mL and DMC extract at 0.1 or 0.2 mg/mL did not manage to decrease yeast concentration in biofilms, although they previously showed antimetabolic activity. The XTT method is a classical method used to quantify fungal biofilms [34–36]. However, this method does not allow a differentiation between bacterial and fungal populations in dual-species biofilms. That is why the FCM approach used for the current study was recently developed [22,37]. A comparative study previously suggested that results provided by colony-forming unit (CFU) counts, XTT reduction or FCM counts were generally comparable and occasional differences could be explained by the specificity and targets of each method [37]. For example, metabolic activity can be reduced without any change in the cell number explaining some divergence in XTT versus CFU or FCM count results. Slight differences between previous and present results could also be at least partially explained by the fact that two different *A. platensis* biomasses were used in these studies, leading to different compositions of extracts. Growth conditions impact the FFA profile as large amounts of ω6-MUFAs and PUFAs were highlighted here, with decreased rates of saturated FFAs, the latter being known to exhibit higher antifungal activity.

By comparing results obtained from growing biofilms (prophylactic activity) and preformed ones (curative activity), we observed that EtOAc extract at 0.2 mg/mL loses its activity once the biofilm is formed (Figures 1A and 2A). Similarly, although all tested extracts significantly limited the growth of single species*C. acnes* biofilms, they were not active anymore once the biofilm was preformed (Figures 1C and 2C). The extracts, whether free or nanovectorized, were not active against dual-species biofilms, growing or already formed as well (Figure 1C,D and Figure 2B,D). Moreover, since single-species *C. albicans* biofilms were prepared aerobically and those involving *C. acnes* anaerobically, a role of the presence of oxygen could not be excluded to explain the different levels of antibiofilm activity that have been observed. In fact, the mechanism of action of the FFA is not completely elucidated. Some studies suggested that their antimicrobial activity would be partly explained by the formation of PUFA peroxidation products [24], which would be favored in an aerobic environment. These oxidized metabolites would act according to a mechanism different from that of native FFAs [24], explaining the residual activity observed on *C. acnes*. As we could expect, these results suggest that preventing the formation of a biofilm is easier than eradicating this biofilm once it is formed.

Different teams demonstrated that biofilms made of more than one species presented reduced susceptibility to antimicrobial treatment compared to single-species biofilms [38,39]. In addition to studying the activity of the extracts and nanocarriers loaded or not by extracts on single-species biofilms, our work assessed the impact of the dual-species nature of the biofilms. Indeed, our results showed that nanocarriers loaded or otherwise with *A. platensis* EtOAc extracts or loaded or not with *A. platensis* DMC extracts as well significantly reduced both already formed and formation of *C. albicans* single-species biofilms, but displayed no or poor activity against *C. albicans* in dual-species biofilms (Figure 3; Figure 4A,B). These results thus suggest that *C. albicans* growing with *C. acnes* in dual-species biofilms is more difficult to inhibit than in single-species ones, which agrees with previous studies on the efficacy of micafungin against *C. albicans* in these two conditions [22]. More generally, results published in recent

years suggest that bacteria and fungi from dual-species biofilms such as *C. albicans–Staphylococcus* spp. or *C. albicans–Streptococcus* spp. often exhibit reduced susceptibilities towards antibiotic or antifungal agents, which is at least partially caused by their synergistic interaction [23,38,40–43]. This study confirmed the activity of the empty nanocarriers against *C. albicans* biofilms which was already observed by Boutin et al. in 2019 [30], suggesting that copper ions could efficiently reach *C. albicans* cells through this single-species biofilm. Cheong et al., 2020 recently confirmed that copper displayed a high antifungal activity against *C. albicans* [44]. Unfortunately, we observed that empty nanocarriers lose their activity at least partially against *C. albicans* as soon as *C. acnes* is present in biofilms, whatever the age of the studied biofilm. Punniyakotti et al. 2020, recently reported the antibiofilm activity of copper nanoparticles studying *Pseudomonas* and *Staphylococcus* species [45]. They hypothesized that Cu2<sup>+</sup> ions liberated from the nanoparticles would be engrossed by the bacterial cell surface and cause cell damage, affecting biofilm development. These authors suggested that the surface binding capability of copper ions would play a key role in the biofilm inhibition. Although we can hypothesize a similar mechanism to explain the activity against fungi, there is no clear explanation as to why empty nanocarriers failed to inhibit biofilm in the presence of *C. acnes*. Nanocarriers loaded with *A. platensis* extracts failed to significantly prevent the formation of *C. acnes* biofilms whereas *A. platensis* extracts without nanocarriers did it in the range of 66.0% to 86.2%. As empty nanocarriers display no activity either, we can hypothesize that nanocarrier loading would counteract the action of extracts against these bacteria (Figure 1; Figure 3C). Conversely, the encapsulation of *A. platensis* extracts induced up to 51.35% of inhibition against the formation of *C. albicans* single-species biofilms (Figures 1A and 3A). As the empty nanocarriers inhibited *C. albicans* single-species biofilm formation and eradicated biofilms, the activity cannot be totally attributed to the extracts. Unfortunately, this encapsulation did not allow the growth inhibition of dual-species biofilms, whatever the studied species (Figure 1; Figure 3B,D).

Finally, *A. platensis* extracts alone or encapsulated in nanosystems displayed an absence of activity against *C. acnes* preformed biofilms (Figures 2C and 4C) whereas the encapsulation of *A. platensis* extracts gave a promising activity against *C. albicans* preformed single-species biofilms, inducing inhibition up to 77.32% (Figures 2A and 4A). Whatever the microorganism studied, the encapsulation does not lead to the obtention of an efficient and significant inhibition of preformed dual-species biofilms (Figure 2B,D and Figure 4B,D)

Very few authors compared the effect of nanosystems vectorizing antimicrobial agents on monoor multispecies biofilms [46,47], and even less on biofilms mixing Gram-positive bacteria and fungi. It is now established that the efficacy of nanosystems on biofilms is linked to their capacity for deeply penetrating the matrix [32]. However, the penetration of nanoparticles into biofilms is highly dependent on the surface characteristics of the nanoparticles [46,48]. Our results suggest that ANCs can diffuse through the extracellular polymeric substance (EPS) of *C. albicans* biofilm, but are not able to diffuse in the EPS of *C. acnes* and in that of polymicrobial biofilm matrix as well. Anjum et al. showed that PLGA nanoparticules loaded with xylitol successfully penetrated into the EPS matrix of single-species biofilms of *S. aureus* or *Pseudomonas aeruginosa*, and also of dual-species biofilms [46]. In the study of Anjum et al., penetration was made easier by adding a ligand onto the nanoparticle surface targeting the biofilm matrix. Tan et al. measured the antibiofilm activity of nanoparticles including enzymes targeting the matrix of biofilms composed of *S. aureus* and *C. albicans* [47]. The particles were able to disrupt in a similar manner single-species or dual-species biofilms, but it was observed that adhesion of bacteria to *Candida* hyphae made their surface less accessible to antimicrobial molecules. This obstacle was already described for free antimicrobial molecules in dual-species biofilms of *S. aureus* and *Fusarium falciforme* [49]. This interaction between the microbial species was also observed for *C. acnes* et *C. albicans* [22] and could participate in the loss of activity of ANCs on *C. albicans* in the dual-species biofilm.

In conclusion, our results highlight the interest of *A. platensis* extracts in preventing the formation of *C. acnes* single-species biofilms. They also suggest that even if the nanocarrier developed by our team offers interesting features, especially in the case of *C. albicans*, its activity against dual-species biofilms is much more limited at the concentrations tested. Even if in vitro models represent simplified models, far from real clinical conditions, developing polymicrobial conditions gives a more realistic representation of clinical biofilms that develop in the human body. This study clearly demonstrated the impact of polymicrobial conditions on the antibiofilm efficacy of nanovectorized antimicrobial systems and highlighted the importance of working in such polymicrobial conditions to have a more objective view of the tested molecules or systems.

### **4. Materials and Methods**

### *4.1. Chemicals*

Ethyl acetate (EtOAc), methanol (MeOH), toluene, hexane, formic acid, diethylether, glacial acetic acid, petroleum ether, sulfuric acid 96% (H2SO4) and dimethylsulfoxide (DMSO) were purchased from Carlo Erba (Val de Reuil, France). Dimethyl carbonate (DMC), sodium alginate, (±)-α-tocophérol 96% (vitamin E), oleic acid, linoleic acid, palmitic acid, myristic acid, stearic acid, palmitoleic acid, γ-linolenic acid, 2,3-bis-(2-methoxy-4-nitro-5-sulfophenyl)-2H-tetrazolium-5-carboxanilide (XTT), menadione and glucose were purchased from Sigma Aldrich (Saint-Quentin Fallavier, France). Phosphoric acid 85% was purchased from Merck pro analysis (Darmstad, Germany). Labrafac®® WL 1349 was purchased from Gattefossé (Saint-Priest, France). Montane 80®® and Montanox 80®® were purchased from Seppic (Castres, France). Copper nitrate Cu(NO3)2 was purchased from Fisher Scientific SAS (Illkirch, France). Vanillin, acetic acid trihydrate 99+% were purchased from Acros Organics (Geel, Belgium). Water was purified using a Milli-Q system (Millipore Corporation, Bedford, MA, USA).

### *4.2. Biomass*

*Arthrospira platensis* was cultivated and harvested by DENITRAL SA (Lamballe, France) and kindly provided by Dr Barbara Clément-Larosière.

### *4.3. Extraction Protocol and Extracts Analyses*

A total of 1 g of biomass was extracted with DMC or EtOAc according to the protocol described by Boutin et al. [30]. The calibration curve was built up using castor oil and results were expressed as mg of equivalent of castor oil in the extract.

Pigments and total lipid rates were obtained using protocol described in Boutin et al. (2019) [30]. FFA profiles were obtained using the LC-ESI-MS protocol adapted from Samburova et al. (2013) [50]. Briefly, LC-ESI-MS analyses were performed on an Acquity H-Class with an SQD detector (Waters, Saint Quentin en Yvelines, France). The system was fitted with a BEH C18 (50 × 2.1 mm; 1.7 μm particle size). The column oven was set at 40 ◦C. Mobile phases were A Water 0.1% NH3 aq; B acetonitrile 0.1% NH3 aq. Flow rate was 0.25 mL/min and the gradient was set as follows—initial solvent B content was 10%, raised to 40% in 2 min, 90% in 23 min and 100% in 1 min and maintained for 9 min. ESI in negative mode was performed with cone voltage set at 50 V and capillary voltage at 2.8 kV.

### *4.4. Alginate-Based Nanocarriers Preparation and Characterization*

Alginate-based nanocarriers (ANCs) were prepared using ultrasound oil-in-water emulsification followed by surface gelation with cupric ions inspired by Nguyen et al. [31] and adapted by Boutin et al. [30]. Briefly, an *A. platensis* lipid extract solution in Labrafac ® WL 1349 (6 mg/mL) was emulsified with a sodium alginate solution in presence of nonionic surfactant, using an ultrasonic probe (Vibra-cell ultrasonic processor, Sonics, Newtown, CT, USA, 20 kHz). The resulting nanoemulsion was mixed under ultrasounds stirring with a solution of copper ions, which complex alginates to form an insoluble copper-alginate gel at the surface of the nanodroplets.

The hydrodynamic diameter and polydispersity index (PdI) of the ANC aqueous suspensions were measured using a dynamic light scattering (DLS) instrument (NanoZS, Malvern Panalytical, Malvern, UK). Each sample was diluted 1:50 in ultrapure water before measurements. Zeta potential was determined on the same sample with the same instrument. Measurements were made in triplicate at 25 ◦C.

### *4.5. Bacterial and Fungal Organisms*

*C. albicans* ATCC® 28367™ and *C. acnes* ATCC® 6919 were used for this study.

Yeasts were cultured on Sabouraud Glucose with Chloramphenicol agar plates aerobically at 37 ◦C whereas *C. acnes* was cultured on Brain Heart Infusion (BHI) agar plates supplemented with 10% of defibrinated horse blood anaerobically at 37 ◦C. Before biofilm experiments, *C. albicans* and *C. acnes* were cultured overnight in BHI at 37 ◦C in aerobic and anaerobic conditions, respectively. Following incubation, cultures were washed with PBS (centrifugation at 2000× *g*, 10 min) and adjusted to 2 × 10<sup>7</sup> cells/mL and 2 × 108 cells/mL in fresh BHI for *C. albicans* and *C. acnes* respectively.

### *4.6. Antibiofim Formation Assay*

Single-species *C. albicans*, single-species *C. acnes* and polymicrobial *C. albicans*-*C. acnes* biofilms were formed in 96-wells flat bottom nontreated polystyrene microplates. In the single-species condition, wells received 100 μL of microbial suspensions. In polymicrobial condition, wells received 50 μL of both microbial suspensions.

Antibiofilm formation activities of lipid extracts previously dissolved in DMSO were tested at two concentrations—0.1 and 0.2 mg/mL. Final DMSO concentrations did not exceed 2% of the overall volume in wells. For extracts included in nanocarriers (NCs), nanosystem tested concentrations were chosen to display extracts at 0.1 and 0.2 mg/mL in ultrapure water—"extract equivalent in nanocarrier" (mg/mL\_EENC). Finally, empty nanocarriers were tested as controls and the studied concentrations corresponded to those present in nanocarriers loaded with extracts at 0.1 and 0.2 mg/mL (0.1\_emptyNC and 0.2\_emptyNC). A total of 100 μL of extract or nanosystem solutions diluted in BHI were then added to the wells. Some wells without extract or nanosystem solution were reserved as a control and received 100 μl of fresh BHI (BHI control). Microplates containing only *C. albicans* were incubated 24 h at 37 ◦C in aerobic conditions while microplates containing *C. acnes* or both microorganisms were incubated in anaerobic conditions.

After incubation, cell concentrations were determined using a protocol adapted from the work of Kerstens et al., 2015 [51]. Planktonic cells were eliminated (2 rounds of washing with 200 μL of PBS) and sessile cells were scraped off from the microplate bottom using sterile tips. An extensive rinsing of the microplate bottom was performed to detach remaining microorganisms. The obtained suspensions were sonicated for 10 min to break down aggregates (Elmasonic S 30, Elma Electronic, Wetzikon, Switzerland, 37 Hz). This procedure has no effect on both *C. albicans* and *C. acnes* viability according to literature data [18,51].

In these microbial suspensions, cells concentrations were determined using flow cytometry (FCM). For dual-species conditions, FCM allowed us to distinguish the yeast population from that of bacteria according to their respective sizes and morphologies [22]. Measurements were performed on a CytoFLEX (Beckman Coulter, Indianapolis, IN USA) managed by CytExpert 2.0.0.153 software (Beckman Coulter, Indianapolis, IN, USA) and equipped with a blue laser (λex = 488 nm) and a 488/8 bandpass filter. The flow rate used was 30 μL·min<sup>−</sup>1.

### *4.7. Anti-Preformed Biofilm Assay*

Single-species *C. albicans*, single-species *C. acnes* and polymicrobial *C. albicans*-*C. acnes* biofilms were formed in 96-well flat-bottom nontreated polystyrene microplates. In single-species condition, wells received 100 μL of microbial suspensions. In polymicrobial condition, wells received 50 μL of both microbial suspensions. Final volume was adjusted to 200 μL using fresh BHI in all conditions. Microplates containing only *C. albicans* were incubated 24 h at 37 ◦C in aerobic conditions whereas microplates containing *C. acnes* or both microorganisms were incubated in anaerobic conditions.

After incubation, supernatants were removed, and biofilms were carefully rinsed twice with 200 μL of PBS. A total of 100 μL of fresh BHI was added to all wells. Then, wells received 100 μL of extract or nanosystem solutions diluted in BHI. Tested conditions were similar to those presented for antibiofilm formation assays. Some wells without extract or nanosystem solution were reserved as a control. A total of 100 μL of fresh BHI was also used in control wells. Microplates containing only *C. albicans* were incubated 24 h at 37 ◦C in aerobic conditions whereas microplates containing *C. acnes* or both microorganisms were incubated in anaerobic conditions.

After incubation, cell concentrations were determined as described previously for the antibiofilm formation assay.

### *4.8. Statistical Analysis*

Experiments were performed at least in duplicate with four replicates for each condition. Mann–Whitney *U* test was applied to determine statistical significance of the differences between the groups using GraphPad Prism® version 6.01 (GraphPad Software Inc, San Diego, CA USA). Differences were considered significant if *p* < 0.005.

**Author Contributions:** Conceptualization, C.I., E.M., M.G. and L.B.-D.; Methodology, C.B., M.G., L.B.-D., E.M. and C.I.; Resources, B.C.L., M.G., L.B.-D., E.M. and C.I.; Investigation, V.L., C.B. and C.L.-L.; Formal analysis, C.B., E.M., L.B. and C.I.; Writing–Original Draft Preparation, C.B., L.B.-D., E.M. and C.I. and S.M.; Writing–Review & Editing, C.B., B.C.-L., M.G., L.B.-D., E.M. and C.I.; Supervision, L.B.-D, E.M. and C.I.; Project administration, L.B.-D., E.M. and C.I. All authors have read and agree to the published version of the manuscript.

**Funding:** This work was supported by a grant (2018–2019) of the AAP Collaborative Research Action of the universities of Tours and Poitiers; This work was also supported by the 2015–2020 State-Region Planning Contracts (CPER), European Regional Development Fund (FEDER), and intramural funds from the Centre National de la Recherche Scientifique (CNRS) and the University of Poitiers. The authors thanks the ARD Cosmétosciences and the Centre Val de Loire Region for financial support (ARD 2017-00118114).

**Acknowledgments:** The authors also wish to thank Didier Debail and Garry Holding for revising the English text.

**Conflicts of Interest:** The authors have nothing to declare.

### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Carbapenem-Resistant** *Klebsiella pneumoniae* **Clinical Isolates: In Vivo Virulence Assessment in** *Galleria mellonella* **and Potential Therapeutics by Polycationic Oligoethyleneimine**

**Dalila Mil-Homens <sup>1</sup> , Maria Martins 1, José Barbosa 1, Gabriel Serafim 1, Maria J. Sarmento <sup>2</sup> , Rita F. Pires <sup>1</sup> , Vitória Rodrigues 3, Vasco D.B. Bonifácio 1,\* and Sandra N. Pinto 1,\***


**Abstract:** *Klebsiella pneumoniae*, one of the most common pathogens found in hospital-acquired infections, is often resistant to multiple antibiotics. In fact, multidrug-resistant (MDR) *K. pneumoniae* producing KPC or OXA-48-like carbapenemases are recognized as a serious global health threat. In this sense, we evaluated the virulence of *K. pneumoniae* KPC(+) or OXA-48(+) aiming at potential antimicrobial therapeutics. *K. pneumoniae* carbapenemase (KPC) and the expanded-spectrum oxacillinase OXA-48 isolates were obtained from patients treated in medical care units in Lisbon, Portugal. The virulence potential of the *K. pneumonia* clinical isolates was tested using the *Galleria mellonella* model. For that, *G. mellonella* larvae were inoculated using patients KPC(+) and OXA-48(+) isolates. Using this in vivo model, the KPC(+) *K. pneumoniae* isolates showed to be, on average, more virulent than OXA-48(+). Virulence was found attenuated when a low bacterial inoculum (one magnitude lower) was tested. In addition, we also report the use of a synthetic polycationic oligomer (L-OEI-h) as a potential antimicrobial agent to fight infectious diseases caused by MDR bacteria. L-OEI-h has a broad-spectrum antibacterial activity and exerts a significantly bactericidal activity within the first 5-30 min treatment, causing lysis of the cytoplasmic membrane. Importantly, the polycationic oligomer showed low toxicity against in vitro models and no visible cytotoxicity (measured by survival and health index) was noted on the in vivo model *(G. mellonella)*, thus L-OEI-h is foreseen as a promising polymer therapeutic for the treatment of MDR *K. pneumoniae* infections.

**Keywords:** *Klebsiella pneumoniae*; KPC and OXA-48-like carbapenemases; *Galleria mellonella* infection model; linear oligoethyleneimine hydrochloride

### **1. Introduction**

The widespread use of antibiotics in clinics caused an increased frequency of multidrugresistance bacteria mainly due to bacterial mutations [1]. In particular, the emergence of resistance to last resource antibiotic treatment options (including carbapenems) has contributed to the limitation of effective therapeutics. Recent reports revealed a weak pipeline for novel antibiotics. From all the compounds under development, very few target infections were caused by Gram-negative bacteria [2–4]. This is clinically relevant, as Gram-negative bacteria infections are significantly more lethal compared to those caused by the Gram-positive [3]. Several factors contribute to the scarcity of new antibiotics, market

**Citation:** Mil-Homens, D.; Martins, M.; Barbosa, J.; Serafim, G.; Sarmento, M.J.; Pires, R.F.; Rodrigues, V.; Bonifácio, V.D.; Pinto, S.N. Carbapenem-Resistant *Klebsiella pneumoniae* Clinical Isolates: In Vivo Virulence Assessment in *Galleria mellonella* and Potential Therapeutics by Polycationic Oligoethyleneimine. *Antibiotics* **2021**, *10*, 56. https:// doi.org/10.3390/antibiotics10010056

Received: 30 September 2020 Accepted: 6 January 2021 Published: 8 January 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

failure being the most relevant. As result of a low return investment, pharmaceutical companies lack incentives for novel antibiotics development. Antibiotics are fast-acting drugs (limiting patient requirements to a small-time window) and the use of novel antibiotics is often reserved since, ultimately, an unpredictable resistance may occur [1,5]. Therefore, the development of novel and efficient antimicrobial agents is of utmost priority.

*Klebsiella pneumonia*, a pathogen of the *Enterobacteriaceae* family, is resistant to last resource antibiotics and is the source of some of the most complicated hospital-acquired infections [6–8]. Resistance to carbapenem in *K. pneumonia* poses a significant threat to patients in hospitals as this organism can cause life-threatening infections such as pneumonia, bloodstream infections and sepsis [9]. Several factors are associated with the acquisition of *K. pneumonia* KPC(+) and OXA-48(+) bacteria, including prolonged hospitalization, infections caused by medical devices (including contamination of ventilators and catheters) and overuse of antibiotics (e.g., carbapenems). Carbapenems-resistant *K. pneumonia* bacteria are capable of inactivating carbapenems via the production of carbapenemase enzymes. Several carbapenemases have been identified and categorized into classes. The ambler classes A (KPC, plasmid-mediated clavulanic acid-inhibited β-lactamases) and D (OXA-48, expanded-spectrum oxacillinase) categories are considered relevant carbapenemases, being highly resistant to all β-lactam molecules, including carbapenems [10].

To investigate the in vivo relevance of MDR *K. pneumonia* infection, we obtained different KPC(+) and OXA-48(+) isolates and determined their virulence using *Galleria mellonella,* a caterpillar model of infection. The success of MDR *K. pneumonia* infections depends, "among other factors", on the ability of the pathogen to escape the host's defense mechanisms. The larvae of the greater wax moth *G. mellonella* have been successfully employed as a model host to study virulence of human pathogenic agents, including several human pathogens, and to investigate the efficacy of therapeutic drugs [11–14]. *G. mellonella* possess only an innate immune system (that includes melanization, hemolymph, and several antimicrobial peptides). However, this is enough to offer powerful resistance to microbial infections [15]. Additionally, their innate immune system shares a high degree of structural and functional homology with the innate immune systems of mammal's [16]. Thus, evaluation of *G. mellonella* responses to *K. pneumonia* isolates infection can provide indication of the mammalian response to these pathogens.

Very few treatment options are available for patients infected with *K. pneumoniae* producing KPC or OXA-48-like carbapenemases and are often limited to administration of multiple antibiotic therapies and to colistin [17,18]. In the light of this, in this study, we report the use of a polycationic synthetic oligomer, linear oligoethyleneimine hydrochloride (L-OEI-h), as an antimicrobial agent for the treatment of *K. pneumonia* KPC(+) and OXA-48(+) bacterial infections. We have previously reported the synthesis and biocidal activity of L-OEI-h against *Streptococcus aureus* and *Escherichia coli* [19]. Herein, we evaluate L-OEI-h antibacterial activity against MDR bacteria clinical isolates, namely *K. pneumoniae*, and investigate the underlying mechanism of action, which, as found for other polycationic antimicrobial agents, might involve disruption of the cell wall and/or the disintegration of the cytoplasmic membrane [10].

### **2. Results**

### *2.1. Evaluation of K. pneumoniae Virulance in G. mellonella Infection Model*

Ten *K. pneumoniae* isolates (Table S1) with reduced sensitivity to carbapenems were obtained from different clinical specimens. To investigate the virulence of *K. pneumoniae* isolates in vivo, we used the *G. mellonella* infection model (Figure 1). In this case, larvae survival rates were measured by injecting inoculums, incubating at 37 ◦C and recording the survival rate daily for up to three days. In all experiments, control groups with administration of phosphate-buffered saline (PBS) solution resulted in a 100% survival rate.

**Figure 1.** In vivo assays using the *Galleria mellonella* larva model. Inoculation by injection of different bacteria inoculum (**a**), healthy larva (**b**), and dead larva (**c**) as a result of *Klebsiella pneumoniae* infection.

Most larvae were found healthy following infection with 1 × 10<sup>4</sup> CFU (colony-forming unit) of each isolate per larva (Figure S1), which suggests that at this bacterial density the humoral immunity of the insects is enough to produce an adequate response to the infection. However, the larvae survival was significantly altered upon increase of infection ratio (1 × 105 CFU per larva). As shown in Figure 2, the virulence of isolates varied widely, with some KPC(+) isolates promoting total larvae mortality (e.g., SYN7 KPC clinical isolate). These differences in virulence are in accordance with data reported for patients suffering from *K. pneumoniae* KPC(+) infections [20], confirming the reliability of the *G. mellonella* infection model in reporting pathogenicity differences between all *K. pneumoniae* isolates. This infection model (*G. mellonella* infected with 1 × 105 CFU *K. pneumoniae* per larva) is now well established in our lab, and we believe that will be very helpful in future development of *K. pneumoniae* therapeutics.

**Figure 2.** Evaluation of virulence of *Klebsiella pneumoniae* KPC(+) and OXA-48(+) isolates in *Galleria mellonella*. Survival of *G. mellonella* was followed for three days after infection with *K. pneumoniae* KPC(+) and OXA-48(+) with 1 × 10<sup>5</sup> CFU per larva. Ten larvae were analyzed in each condition and larvae survival was monitored daily. In all cases, no larvae death was observed upon administration of PBS (control). KPC(+) isolates: SYN1, SYN6, SYN7, SYN8, SYN9 SYN19 and SYN22; OXA-48(+) isolates: SYN3, SYN4 and SYN17.

### *2.2. Antimicrobial Activity of L-OEI-h*

The oligomer L-OEI-h was synthesized following our reported protocol [19]. We evaluated the antimicrobial activity of L-OEI-h against *K. pneumoniae* isolates by determination of the minimum inhibitory concentration (MIC) and the minimum bactericidal concentration (MBC). MIC is defined as the lowest drug concentration that prevents visible growth of bacteria. The MBC is the lowest concentration of an antimicrobial agent required to kill ≥99.9% bacteria over an extended period (18–24 h). The antimicrobial assays were

conducted according with CLSI guidelines in Mueller Hinton broth (MHB), a nutrient rich bacterial growth medium. The obtained results for the different isolates are shown in Table 1. We included other Gram-negative bacteria (*Pseudomonas aeruginosa* PAO and *E. coli AB1157*) and a Gram-positive strain (*S. aureus* MRSA JE2) as control strains. Except for SYN7 KPC, which is also the most virulent isolate, L-OEI-h displayed good antibacterial activity against Gram-negative strains with particular relevance for *P. aeruginosa* PAO and *E. coli* AB1157. The MIC and MBC values were almost identical, which is indicative that the oligomer exerts a bactericidal activity.

**Table 1.** Antimicrobial activity of linear oligoethyleneimine hydrochloride (L-OEI-h) against *Klebsiella pneumoniae* isolates, control bacterial strains (*P. aeruginosa* PAO, *E. coli* AB1157) and the methicillinresistant *Staphylococcus* strain *S. aureus* JE2.


\* Data acquired in a previous study [13].

We previously demonstrated that L-OEI-h is also able to target some Gram-positive bacteria with low efficiency [19], as demonstrated here for methicillin-resistant *S. aureus* JE2 (MIC > 915 μg/mL).

### *2.3. Biocompatibility Studies*

Some membrane-lytic agents are known to display selectivity towards bacterial membranes, allowing for elevated antibiotic activity and low toxicity to mammalian cells [21]. The differences in composition and lipid arrangement in bacterial and mammalian cell membranes can support the selectivity observed in these antimicrobial agents [21,22]. Here, we evaluated the cytotoxicity of L-OEI-h using in vitro (L929 mouse fibroblasts) and in vivo models (*G. mellonella* larvae) (Figure 3).

The polycationic oligomer compound has very little cytotoxicity against mammalian cell lines, even at the highest dose tested against clinical isolates (915 μg/mL, Figure 3a,b). The larvae were injected with 5 μL of different concentrations of L-OEI-h, and then incubated in Petri dishes at 37 ◦C and daily scored for survival. Up to a concentration of 915 μg/mL, all larvae were found healthy for a three-day period (Figure 3c). To obtain more differences in larvae health, we also determined the health index scores (Figure 3d), which scores four main parameters: larvae activity, cocoon formation, melanization and survival. The injection of the larvae with L-OEI-h even after 72 h resulted in high health index scores. The higher activity and more cocoon formation are regularly associated to a healthier wax worm [23].

All experiments included a control group injected only with a PBS solution. Overall, our results corroborate the biocompatibility of L-OEI-h.

**Figure 3.** Evaluation of linear oligoethyleneimine hydrochloride (L-OEI-h) toxicity and biocompatibility using in vitro mammalian cells (L929 mouse and A549 human epithelial cells) (**a**,**b**) and in vivo assays (*Galleria mellonella* larvae) (**c**). The health index scores of wax worms injected with L-OEI-h was also evaluated (**d**). Asterisks (\*) represent statistical significance in *t*-student tests (*p* < 0.05) compared to the untreated samples.

### *2.4. Exploring the L-OEI-h Mechanism of Action*

Cationic polymers are expected to target the microbial cell surface, via binding to negatively charged components, and to disrupt the cytoplasmic membrane [24]. A fastkilling kinetics is associated with surface membranolytic processes, while slow kinetics is usually associated with activation of intracellular processes [25–28]. In this sense a time-kill assay based on traditional colony count was carried out to discriminate between fast and slow kinetics of antibacterial activity.

Time-kill curves of a selected OXA-48(+) (SYN4 OXA-48) and KPC(+) isolates (SYN8 KPC) were obtained in the presence of two different oligomer concentrations. As shown in Figure 4 and Figure S2, after the addition of L-OEI-h, the number of *K. pneumoniae* viable and culturable colonies decreases significantly in the first 30 min of treatment, with almost complete bacterial removal achieved at 2 h treatment. This observation illustrates a possible fast L-OEI-h bactericidal activity. For other Gram-negative bacteria (e.g., *E. coli* AB1157 [19]), L-OEI-h had a very fast killing effects (within 5 min). In both cases, the verified fast activity is supportive of a surface membranolytic mechanism of action for L-OEI-h.

**Figure 4.** Killing kinetics of *Klebsiella pneumonia* SYN4 OXA-48 (**a**) and SYN8 KPC (**b**) clinical isolates induced by L-OEI-h. The killing kinetics was evaluated with a colony count assay using two different oligomer concentrations, 2 × MIC and 3 × MIC.

The higher antimicrobial activity of L-OEI-h against Gram-negative bacteria (e.g., *E. coli* AB1157 vs. *S. aureus* NCTC8325-4 [19]) may indicate that the presence of lipopolysaccharides (LPS) in the outer leaflet of the outer membrane of Gram-negative may facilitate the initial binding of the compound. Antibacterial cationic polymers are expected to permeabilize bacterial membrane by one of two possible mechanisms: (i) Perpendicular insertion in the membrane followed by pore formation and membrane permeabilization/depolarization, or (ii) accumulation at the membrane surface until a certain threshold concentration is achieved, which leads to membrane disruption and cell lysis.

To further characterize the mechanism of action of L-OEI-h, studies with membrane mimetics were carried out. In these simplified membrane models, size, geometry, and composition can be tailored with great precision.

The effect of L-OEI-h on the membrane was thus, studied using large unilamellar vesicles (LUVs) of controlled lipid composition, serving as mimetics of bacterial and mammalian cell membranes. The bacterial membrane model, with an overall negative charge, was composed of different ratios of phosphatidylcholine (POPC) and phosphatidylglycerol (POPG), while the healthy mammalian plasma membrane model contained only POPC [20]. POPC zwitterionic liposomes are not affected by the addition of L-OEI-h, but higher amount of negatively charged lipids (>20%) led to the formation of large clusters within a few seconds (Figure 5). We hypothesize that the formation of large clusters (>100 nm) is consistent with the binding/accumulation of the polymer in membranes with high negatively charged lipids.

**Figure 5.** Effect of L-OE-h on model liposomes composed of phosphatidylcholine (POPC) (**a**) and POPC with varying phosphatidylglycerol (POPG) content (**b**). Vesicle sizes were measured by DLS at 25 ◦C. Dashed lines are the respective controls without polymer.

Upon electrostatic binding of L-OEI-h to anionic lipid vesicles, and after a certain threshold concentration, the oligomer induces vesicle aggregation and/or vesicle fusion, which explains the appearance of a LUV population larger in size (Figure 5b). Vesicle aggregation and fusion are not unusual occurrences, being also observed in the case of cationic peptides addition to anionic vesicles [29]. Ultimately, accumulation of L-OEI-h in the membrane may induce micelle formation or the formation of transient pores, as illustrated for the mechanism of action of some antimicrobial peptides (e.g., [29]). In both cases membrane disintegration is a consequence of these events (Figure 6).

**Figure 6.** Hypothetical mechanism of action of L-OEI-h (model II) towards bacterial cell membranes. The accumulation of the polycationic oligomer at the membrane surface, due to electrostatic interactions, results in an oligomer threshold concentration capable of cell disruption and lysis.

### **3. Discussion**

Bacterial infections are becoming a major human health problem. Resistance towards antibiotics is becoming increasingly common, including to the ones only reserved for the treatment of severe infections. The World Health Organization has recently included *Klebsiella* in the critical list of microorganisms for which new therapeutics are urgently needed [30]. In the light of this demand, in this study we examine the sensitivity of *G. mellonella* to *K. pneumoniae* isolates that are resistant to last resources antibiotics and focused our attention towards a polycationic oligomer [19], a synthetic mimic of host defense peptides (HDPs), as a novel treatment for MDR *K. pneumonia* infections.

HDPs are a class of innate immunity components expressed by all multicellular organisms [31]. It is believed that their function is, in part, to kill invasive cells without prejudice to the host and without presenting itself as a stress agent for the development of resistance traits [31]. The discovery of HDPs was accompanied by the development of disinfectant polymers, which in the late 1990s led to HDP-mimicking polymers [32]. Although antimicrobial peptides (AMPs)/HDPs are an excellent alternative to conventional antibiotics, cheap and scalable bioprocessing is not yet available [33]. Additionally, AMPs are poorly stable in vivo due to protease liability.

In this way, HDP-mimicking polymers are considered a more cost-effective and stable alternative to HDP/AMP [31]. The relationship between structure and activity of HDPmimetic polymers relies on two main design principles: (i) Hydrophobic/hydrophilic component ratio and (ii) presence of a structured cationic group. It was demonstrated that a poly(methacrylate) random copolymer with 40% methyl side chains (hydrophobic segment) and 60% aminoethyl side chains (hydrophilic segment) show potent antibacterial activity and low hemolytic activity [34].

The effect of primary amine groups, instead of traditional quaternary ammonium salt (QAS), was already investigated [35]. For the same type of polymer, primary amines outperformed in comparison with tertiary or quaternary amines in terms of antimicrobial activity and toxicity. Following this study, it was found that primary ammonium groups can form a stronger complex with phospholipid headgroups when compared to QAS analogues [36]. Additionally, the effect of amine groups density is quite relevant since increasing the amine density by monomer unit enhances the polymers efficacy and decreases hemolysis in great extent [37].

The synthetic polymer polyethyleneimine (PEI), due to its intrinsic features, is regarded as a good alternative to fight antibiotic resistant organisms. The abundance of reactive amine groups in the backbone allows post-modifications to display both hydrophobicity and a positive charge density, primary requirements for a good antimicrobial activity [38]. In previous studies, we found that L-OEI-h, a PEI analogue, is a very effective biocidal oligomer [19], whose activity may be attributed to high positive charge density (hydrochloride salt, quaternary nitrogen atoms), if compared with commercial L-PEI (nonquaternary nitrogen atoms) [39] (see Figure 7). As we and others demonstrated, a higher positive charge density leads to higher antimicrobial efficacy [40].

**Figure 7.** Comparison of positive charge density between linear oligoethyleneimine hydrochloride (L-OEI-h) and linear polyethyleneimine (L-PEI) at physiological pH.

In this work, we evaluated the antimicrobial activity of L-OEI-h against a variety of *K. pneumonia* strains. The antimicrobial efficiency is attributed to favorable electrostatic interactions between the polycationic oligomer (having *ca.* +11 formal net charge, one positive charge per monomer unit) and the anionic bacterial membrane. This interaction can induce a very fast-bactericidal activity, as demonstrated by us [19]. Such a fast mode of action strongly suggests that membranolytic processes are responsible for antimicrobial activity. Through direct permeabilization of the lipidic membrane, instead of action on a specific cellular target, the probability of bacteria to develop resistance to this treatment is extremely low.

For *K. pneumoniae* clinical isolates, the bactericidal activity of the oligomer does not occur earlier than 30 min, in contrast with what we and others verified for other membrane disruption agents, whereby membrane permeabilization and/or lysis happened within 5 min [19,41,42]. It is likely that, for *K. pneumoniae* clinical isolates, L-OEI-h cannot diffuse so efficiently (or diffuses slowly) through the bacterial cell wall to reach the plasma membrane. This could be associated with the fact that the capsule (composed of extracellular polysaccharides) of *K. pneumoniae* is more "robust" than what was verified for other gram-negative bacteria, and, in this sense, constitutes an efficient barrier against several antimicrobial agents [43,44] including HDPs. Despite this, L-OEI-h was able to efficiently kill several *K. pneumoniae* clinical isolates (Table 1, Figure 4 and Figure S2).

Importantly, the L-OEI-h antimicrobial action against some *K. pneumoniae* clinical isolates such as SYN4 OXA-48 and SYN9 KPC (a highly virulent clinical isolate) is verified under a possible therapeutic window, since L-OEI-h induces low cytotoxicity against in vitro mammalian cell lines (<80% cell viability [45] within statistical error) and, more importantly, no visible cytotoxicity was detected against the in vivo model (*G. mellonella*). As shown here, low toxicity in these models is likely associated with reduced interaction with the plasma membrane of eukaryotic cells. Liposomes with a lipid composition mimicking the outer leaflet of eukaryotic plasma membranes were not affected by the presence of L-OEI-h (Figure 5a). On the other hand, liposomes rich in phosphatidylglycerol,

mimicking bacterial membrane lipid composition, showed dramatic aggregation upon interaction with this compound. Large unilamellar vesicles (LUVs) were used in these studies as their size is more amenable to DLS resolution. Other membrane models, such as multilamellar vesicles (MLVs), due to their multi-layered character, prevent polymer interactions with the internal bilayers (that could mask possible effects occurring only in the outer membrane), while small unilamellar vesicles (SUV)'s small size results in membranes with excessive curvature that do not mimic bacterial membranes. On the other hand, very large vesicles such as giant unilamellar vesicles (GUVs) have a less controlled lipid concentration/size and thus frequently show significant vesicle heterogeneity within the same sample.

Although we obtained promising data for some *K. pneumoniae* isolates, we verified that in some cases MIC and MBC values were still high (if compared with the results obtained for the Gram-negative control strains). The polymer design and the polymermembrane interactions are crucial issues for bacterial infection eradication. In a recent study, linear and branched PEI polymers were found to have very similar MIC values, while εpolycaprolactone showed superior broad-spectrum antimicrobial properties over L-PEI [46]. Hence, in future work we will consider the use of a coarse-grain (CG) molecular dynamics model of L-OEI-h to better understand its interactions with bacterial and mammalian membranes. This study will allow the identification of key oligomer-membrane interactions which could lead to enhanced polymer antimicrobials.

### **4. Materials and Methods**

### *4.1. Synthesis of Linear Oligoethyleneimine Hydrochloride (L-OEI-h)*

The preparation of L-OEI-h was made in two steps following our previous protocol. First, linear 2-ethyl(2-oligooxazoline) (OEtOx) was synthesized by cationic ring-opening polymerization (CROP) in supercritical carbon dioxide, a green polymerization methodology [47]. The living OEtOx polymer was terminated with water and isolated as a brownish sticky oil. Next, OEtOx was hydrolyzed overnight using a HCl 5M solution. After this period, the precipitated solid was filtered and washed with acetone to obtain L-OEI-h as an off-white solid in quantitative yield. After vacuum drying, L-OEI-h is ready to use [19].

### *4.2. Clinical Isolates Collection and Identification*

The KPC and OXA-48-positive carbapenems-resistant *K. pneumoniae* clinical isolates were collected from patients treated in medical care units (Lisbon, Portugal). All strains were identified with matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) spectrometry [48,49] using the VITEK MS system (bioMérieux). Briefly, inoculation loops were used to select and smear the isolates onto the sample spots/target slide. Then 1 μL VITEK mass spectrometry α-cyano-4-hydroxycinnamic acid (MS-CHCA) matrix was applied over the sample and air dried (1–2 min). The target slide was loaded into the VITEK MS system to acquire the mass spectra of whole bacterial cell protein (which is mainly composed of ribosomal proteins). Then, the mass spectra acquired for each sample were compared to the mass spectra contained in the database.

In addition, the following antimicrobials were included in the microorganism's characterization: β-lactams (ceftazidime, cefepime, cefuroxime/axetil, amoxicillin/clavulanate, ticarcillin/clavulanate and piperacillin/tazobactam), carbapenems (meropenem, ertapenem), aminoglycosides (gentamicin, amikacin), nitrofurantoin fluoroquinolones (ciprofloxacin), polymyxin (colistin), fosfomycin, trimethoprim/sulfamethoxazole. The production of carbapenemases in these strains is evidenced through the antibiotic resistance profile and the type of carbapenemase (OXA-48-like, KPC, NDM or VIM) was identified through an immunochromatographic method (RESIST-4 O.K.N.V., Coris). CASFM-EUCAST 2016 defined breakpoints for *Enterobacteriaceae* were used to interpret susceptibility data for *K. pneumoniae* (http://www.sfm-microbiologie.org).

From pure culture on MacConkey agar plates, all identified *K. pneumoniae* isolates were transferred to 1.5 mL Eppendorf tubes contain Luria–Bertani (LB) broth with 20% (*v*/*v*) glycerol and were maintained at −80 ◦C for long-term storage. For the antimicrobial activity studies, bacterial cultures were inoculated in Mueller-Hinton broth (MHB) (Difco), at 37 ◦C. Tryptic soy agar (TSA) agar plates (bioMérieux, Marcy l'Etoile, France) were used to subculture *K. pneumoniae* isolates, *P. aeruginosa* PAO and *E. coli* AB1157. *S. aureus* MRSA JE2 was streaked on Columbia agar +5% sheep blood plates (COS, bioMérieux, Marcy l'Etoile, Auvergne-Rhône-Alpes, France) and grown overnight at 37 ◦C. *P. aeruginosa* PAO, *E. coli* AB1157 and *S. aureus* JE2 were used as reference strains. The antimicrobial activity of L-OEI-h against *E. coli* AB1157 was previously investigated by us [19].

### *4.3. Galleria mellonella Infection Model*

*G. mellonella* wax moth larvae were reared in our lab at 25 ◦C in the dark, from egg to last-instar larvae, on a natural diet (beeswax and pollen grains). Worms of the final-instar larval stage, weighing 250 ± 25 mg, were selected for the experiments. The *G. mellonella* survival experiment was adapted from previous studies with small changes [11,14]. Briefly, all *K. pneumoniae* isolates were grown overnight in TSA plates. Then, the assay was carried out by preparing two distinct inoculums of 2 × 106 and 2 × 107 CFU/mL in PBS. Using a hypodermic microsyringe, the larvae were injected with 5 μL of each bacterial suspension via the hindmost left proleg, previously surface sanitized with 70% (*v*/*v*) alcohol. Different groups were used (*n* = 10 each)—larvae injected with PBS to monitor the killing due to injection trauma (control) and larvae injected with *K. pneumoniae* isolates. After inoculation, larvae were kept in Petri dishes and maintained in the dark at 37 ◦C for 72 h. The larval survival was assessed daily during that period, and caterpillars were considered dead based on the lack of mobility in response to touch. Each larva was also scored daily to the *G. mellonella* health index, which scores four main parameters: Larvae activity, cocoon formation, melanization and survival, as described in [23].

All experiments were performed using a minimum of two independent experiments.

### *4.4. Antimicrobial Activity*

### 4.4.1. Minimum Inhibitory Concentration (MIC) Determination

To determine the MIC values, the bacterial suspension was initially adjusted to a concentration of 1 × 10<sup>6</sup> CFU/mL in MHB (according to CLSI guidelines) [50]. On a 96-well plate (Orange Scientific, Braine-l'Alleud, Belgium), two-fold serial dilution (in MHB) of L-OEI-h was then added to each well that contained the bacterial inoculums (dilution 1:1). Final bacteria inoculum in each well were diluted to 5 × 10<sup>5</sup> CFU/mL. The 96-well tissue culture plates were incubated for 18–20 h at 37 ◦C. All experiments were performed using a minimum of three independent experiments performed with three technical replicates each.

### 4.4.2. Minimum Bactericidal Concentration (MBC) Determination

The MBC values were determined by the traditional colony count assay [51]. At the end of the MIC assay, 20 μL samples from each well (corresponding to <sup>1</sup> <sup>2</sup> MIC, MIC, 2 × MIC, 3 × MIC of L-OEI-h) were transferred to a new 96-well plate and successively diluted (10-fold) in MHB. Then, each dilution was sub-cultured in TSA plates. After incubation at 37 ◦C for 24 h, the resultant viable colonies were counted. All experiments were performed using a minimum of three independent experiments performed with three technical replicates each.

### 4.4.3. In Vitro Time-Kill Curves

Time-kill curves of L-OEI-h were determined according to literature [52,53]. Briefly, a final inoculum of 1 × 106 to 1 × 107 CFU/mL was exposed to distinct doses of L-OEI-h and incubated at 37 ◦C. Aliquots at specified time points were taken; 10-fold dilutions of each well were prepared and plated onto a TSA plate for CFU enumeration. TSA plates were incubated for 24 h at 37 ◦C and bacterial colonies were counted. Viable cells (CFU/mL) are reported here as percentage of the control (bacterial suspension without

L-OEI-h exposition). From the number of bacterial colonies obtained, viable bacteria (in CFU/mL) are reported as percentage of the control. All experiments were performed using two independent experiments with three technical replicates each.

### *4.5. Biocompatibility Assays*

### 4.5.1. MTT Viability Assay

L-929 and A549 cell lines were cultured in T-75 cell culture flasks (Filter caps) using Dulbecco's modified Eagle's medium (DMEM, Catalog number 41966-029) supplemented with 10% fetal bovine serum (Thermo Fisher Scientific, Catalog number 10500-064, heat inactivated) and 1% penicillin-streptomycin (Thermo Fisher Scientific, catalogue number 15140-122) and maintained in a humidified atmosphere with 5% CO2 at 37 ◦C. All cell culture lines were maintained with routinely subcultures (using TrypLE Express without phenol red, GIBCO™, for chemical detaching). Mammalian cell lines were counted with a hemocytometer.

The MTT assay was used to detect changes on metabolic activity of mammalian cells [54]. Briefly, the L-929 and A549 cell lines were seeded in 96-well flat-bottomed polystyrene plates with a density of 1 × 104 cells/well and left to adhere overnight in a CO2 incubator (5%) at 37 ◦C. After 24 h, the cell medium was discarded and replaced with fresh medium containing different concentrations of L-OEI-h. Cells were then incubated for a period of 24 h at 37 ◦C in a humidified 5% CO2 incubator. After this incubation period, the medium was discarded and 20 μL of MTT (5 mg/mL) were added to each well together with 100 μL of fresh DMEM and incubated at 37 ◦C for 3.5 h. The formazan crystals formed in the wells were dissolved using 150 μL of MTT solvent (4 mM of HCl, 0.1% of Nondet P-40 in isopropanol). The formation of formazan was monitored by measuring the absorbance at 590 nm in a microplate reader (BMG Labtech, Polar Star Optima). Cell viability was determined relatively to the untreated sample after correcting the data with the negative control.

All experiments were performed using a minimum of two independent experiments with three technical replicates each.

### 4.5.2. *Galleria mellonella* Toxicity Assay

The L-OEI-h toxicity was also evaluated in the larvae infection in vivo model. The *G. mellonella* killing assays were based on the above descriptions with small modifications. L-OEI-h doses were prepared and injected into the larvae hindmost left proleg. The larvae survival was assessed daily during a period of 72 h. A control group was also included in the assay. Two independent experiments were performed.

### *4.6. Exploring the L-OEI-h Mechanism of Action*

### Liposome Preparation

Large unilamellar vesicles (LUVs), with 100 nm of diameter, were prepared by extrusion of multilamellar vesicles [54]. The liposomes were prepared according to methods previously described [55]. Briefly, lipid mixtures composed of adequate amounts of lipids (POPC and POPG) were prepared in chloroform to a final lipid concentration of 2 mM. The solvent was slowly vaporized under a nitrogen flux and the resulting lipid film was left in vacuum for 3 h to ensure the complete removal of chloroform. Afterwards, the lipid was resuspended in 2 mL of DPBS (Thermo Fisher Scientific) and freeze–thaw cycles (liquid nitrogen/water bath at 60 ◦C) were performed to re-equilibrate and homogenize the samples. LUVs were finally obtained by extrusion of the solutions at 50 ◦C with an Avanti Mini-Extruder (Merck, Darmstadt, Germany) using 100 nm pore size polycarbonate membranes. All lipid stock solutions were prepared in chloroform and the respective concentrations were determined by the colorimetric quantification of inorganic phosphate.

Liposome size was determined by dynamic light scattering (DLS) using a Nanosizer ZS (Malvern Instruments). The POPC/POPG vesicles were incubated with L-OEI-h for 5 min. Data was collected at 25 ◦C and a backscattering angle of 173◦. Two independent experiments were performed.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/2079-6 382/10/1/56/s1, Figure S1: Evaluation of distinct virulence response of *K. pneumoniae* KPC(+) and OXA-48 (+) isolates in *G. mellonella*. Figure S2: Killing kinetics of *K. pneumonia* SYN4 OXA-48 (a) and SYN8 KPC (b) clinical isolates induced by L-OEI-h. Table S1: *K. pneumonia* clinical isolates origin and the minimum inhibitory concentration (MIC) values of trimethoprim (TM)/sulfamethoxazole (SM) antibiotics.

**Author Contributions:** Conceptualization, S.N.P., D.M.-H., V.D.B.B.; methodology, S.N.P., D.M.-H., V.D.B.B, R.F.P., M.M., J.B., M.J.S.; formal analysis, S.N.P., D.M.-H., V.D.B.B, M.J.S.; investigation, S.N.P., D.M.-H., R.F.P., M.M., J.B., M.J.S.; V.R.; writing—original draft preparation, S.N.P.; writing—review and editing, S.N.P., V.D.B.B, D.M.-H., M.J.S., G.S.; supervision, S.N.P., V.D.B.B. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by Fundação para a Ciência e a Tecnologia (FCT, Portugal), Ministério da Ciência, Tecnologia e Ensino Superior (MCTES) through projects PTDC/MEC-ONC/29327/2017, UIDB/04565/2020 and SAICTPAC/0019/2015.

**Data Availability Statement:** The data presented in this study are available in the article and in the supplementary material.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**


*Article*

### **Synthesis of Electrospun TiO2 Nanofibers and Characterization of Their Antibacterial and Antibiofilm Potential against Gram-Positive and Gram-Negative Bacteria**

**Mohammad Azam Ansari <sup>1</sup> , Hani Manssor Albetran 2, Muidh Hamed Alheshibri <sup>3</sup> , Abdelmajid Timoumi <sup>4</sup> , Norah Abdullah Algarou 5,6, Sultan Akhtar <sup>5</sup> , Yassine Slimani <sup>5</sup> , Munirah Abdullah Almessiere <sup>5</sup> , Fatimah Saad Alahmari 7, Abdulhadi Baykal <sup>7</sup> and It-Meng Low 8,\***


Received: 23 July 2020; Accepted: 1 September 2020; Published: 3 September 2020

**Abstract:** Recently, titanium dioxide (TiO2) nanomaterials have gained increased attention because of their cost-effective, safe, stable, non-toxic, non-carcinogenic, photocatalytic, bactericidal, biomedical, industrial and waste-water treatment applications. The aim of the present work is the synthesis of electrospun TiO2 nanofibers (NFs) in the presence of different amounts of air–argon mixtures using sol-gel and electrospinning approaches. The physicochemical properties of the synthesized NFs were examined by scanning and transmission electron microscopies (SEM and TEM) coupled with energy-dispersive X-ray spectroscopy (EDX), ultraviolet-visible spectroscopy and thermogravimetric analyzer (TGA). The antibacterial and antibiofilm activity of synthesized NFs against Gram-negative *Pseudomonas aeruginosa* and Gram-positive methicillin-resistant *Staphylococcus aureus* (MRSA) was investigated by determining their minimum bacteriostatic and bactericidal values. The topological and morphological alteration caused by TiO2 NFs in bacterial cells was further analyzed by SEM. TiO2 NFs that were calcined in a 25% air-75% argon mixture showed maximum antibacterial and antibiofilm activities. The minimum inhibitory concentration (MIC)/minimum bactericidal concentration (MBC) value of TiO2 NFs against *P. aeruginosa* was 3 and 6 mg/mL and that for MRSA was 6 and 12 mg/mL, respectively. The MIC/MBC and SEM results show that TiO2 NFs were more active against Gram-negative *P. aeruginosa* cells than Gram-positive *S. aureus*. The inhibition of biofilm formation by TiO2 NFs was investigated quantitatively by tissue culture plate method using crystal

violet assay and it was found that TiO2 NFs inhibited biofilm formation by MRSA and *P. aeruginosa* in a dose-dependent manner. TiO2 NFs calcined in a 25% air-75% argon mixture exhibited maximum biofilm formation inhibition of 75.2% for MRSA and 72.3% for *P. aeruginosa* at 2 mg/mL, respectively. The antibacterial and antibiofilm results suggest that TiO2 NFs can be used to coat various inanimate objects, in food packaging and in waste-water treatment and purification to prevent bacterial growth and biofilm formation.

**Keywords:** TiO2 nanofibers; electrospinning; biofilm prevention and control; multidrug-resistant bacteria; biomedical application

### **1. Introduction**

Titanium dioxide (TiO2) is among the investigated photocatalytic nanomaterials and is used extensively in diverse applications and for diverse purposes [1]. TiO2 nanomaterials are widely used in waste-water treatment and purification, air-pollutant decomposition, implantable devices, air-conditioning filters, hydrophilic coatings, self-cleaning and self-disinfecting devices, pesticide degradation (e.g., herbicides, insecticides and fungicides) and in the production of hydrogen fuel [2,3]. TiO2 is usually non-toxic, highly durable with a high refractive index, high absorption of light and a lower-cost production with antibacterial activity [4,5]. Because of its strong stability, TiO2 materials can be applied easily on inanimate items, e.g., metal, glass and biomedical implants [5]. Recently, TiO2 nanoparticles (NPs) have attracted increased interest in the scientific and industrial community because of their extensive applications in biological and pharmaceutical areas, purification of environmental sources, electronic system, solar energy cells, photocatalysts, photo-electrodes and gas sensors. TiO2 NPs are proven to be employed in food technology, drugs, cosmetics, paint pigment, ointments and toothpaste [6,7]. Because of their cost-effective, safe, stable, non-toxic, non-carcinogenic, photo-induced super-hydrophobicity and antifogging properties, TiO2 NPs have been used to kill bacteria, remove toxic and harmful organic elements from water and air and for self-sterilize glass surfaces [8–11].

However, it is difficult to separate TiO2 NPs after a photochemical reaction, which limits their practical applications [12]. TiO2 NPs aggregate easily in solution, which reduces their photocatalytic efficacy because of the decreased surface area. These limitations can be overcome by preparing TiO2 nanofibers (NFs) using simple, rapid and cost-effective electrospinning (ES) methods [13–18]. TiO2 NFs have gained increased attention because of their mesoporous structure [19], stability in solution, little or no aggregation, high surface to volume ratio that enhances photocatalytic reactions and their ease in separation and collection from solution after photochemical reactions [20,21]. However, the photocatalytic efficacy of TiO2 NFs is comparatively low and is effective only under ultraviolet (UV) light because of their relatively large band-gap energy and low-ordered crystalline structure [22]. An exceptional feature of TiO2 nanoparticles (NPs) is their photocatalytic activity that enhances the bacterial killing when exposed to UV light [7,23]. TiO2 NPs tend to exist in three principal forms, namely brookite, rutile and anatase, and it has been reported that the anatase form has a high photocatalytic and antibacterial activity [23–26]. A major biomedical application of TiO2 NPs is to prevent biofilm formation on medical devices that is related to infections and sepsis [3,27,28]. Several researchers have focused on the antibacterial and antibiofilm activities of TiO2 NPs under UV light against standard bacterial strains, e.g., ATCC, MTCC and NCIM. However, limited work has been published on the antibacterial and antibiofilm activities of TiO2 NFs without application of UV light against drug resistant isolates. The objective of present investigation is to explore the antibacterial and antibiofilm efficacies of TiO2 NFs in dark against two major human pathogenic drug resistant bacteria i.e., Gram-positive *methicillin-resistant Staphylococcus aureus* (MRSA) and Gram-negative *Pseudomonas aeruginosa* by using different methods.

### **2. Experimental Methodology**

### *2.1. Electrospinning and Heating Protocol*

Both the sol-gel and electrospinning approaches were used to synthesize electrospun TiO2 NFs. Briefly, Titanium isopropoxide (IV), acetic acid and ethanol were mixed and stirred with respect to volume ratio of 3:1:3. After that, 12% by weight of polyvinylpyrrolidone (PVP) was dissolved in the obtained TiO2 solution. This mixed TiO2/PVP sol-gel was then placed within a plastic syringe for electrospinning experiment. Additional details are provided in a preliminary study [15]. Thermal gravimetric analysis (TGA) and differential scanning calorimetry (DSC) for the non-isothermal heating of electrospun TiO2 NFs were performed on a Mettler Teledo thermal gravimetric analyzer TGA/DSC. The samples were heated from ambient temperature to 900 ◦C at a rate of 10 ◦C/min with an argon protective gas of 20 mL/min in various mixtures of air and argon. The thermal experiments were carried out by utilizing alumina crucibles that were charged with 25 mg of sample and in mixtures of 50% air-50% argon, 25% air-75% argon and 100% argon. It is worth noting that the argon shielding gas is included in the relative percentage of air to argon gas. For safety reasons, samples that were contacted with 100% air were heated in an oven under the same conditions [29].

### *2.2. Characterization of Electrospun TiO2 NFs*

The morphological and structural properties of as-prepared NFs were characterized by SEM (FEI Inspect S50) and TEM (FEI Morgagni 268). The elemental composition was determined by energy-dispersive spectroscopy (EDX). A strong correlation can be established from the initial microstructure images. The TiO2 grains were structured as microspheres and a complete description of the microstructure is provided. The microstructure relates to monitoring by three-dimensional imaging of the evolution of internal porosity as a function of annealing temperature. A Jasco V-670 UV–visible diffuse reflectance spectrophotometer (DRS) under a wavelength ranging between 200 and 750 nm was used to estimate the band gap energy (*Eg*) of various TiO2 NFs.

The values of band gap energy (*Eg*) were calculated from the absorption spectra versus wavelength using the following expression:

$$E\_{\mathcal{S}} = \frac{h\mathcal{C}}{\lambda\_0} \tag{1}$$

In this expression, *h* is Planck's constant (6.626 × 10−<sup>34</sup> J.s) and *C* is the speed of light (3 × 10<sup>8</sup> m/s). λ<sup>0</sup> (expressed in nm) is the cut off wavelength obtained from the absorption spectra [30]. Accordingly, λ<sup>0</sup> denotes the absorption edge wavelength, obtained from the offset wavelength derived and extrapolated from the low energy absorption band.

### *2.3. Evaluation of Antibacterial Activity of Electrospun TiO2 NFs*

### 2.3.1. Bacterial Culture

The laboratory strain of Gram-negative *Pseudomonas aeruginosa* PAO1 and Gram-positive methicillin resistant *Staphylococcus aureus* (MRSA) ATCC 33591 used in this study was obtained from Molecular Microbiology Laboratory, Institute for Research and Medical Consultations, Imam Abdulrahman Bin Faisal University, Dammam, Saudi Arabia. The bacterial strains preserved in glycerol cultures (−80 ◦C) were cultivated on Tryptic soy broth (TSB) at 37 ◦C in a shaker incubator before being used for microbial studies.

2.3.2. Investigation of Minimum Inhibitory and Minimum Bactericidal Concentration (MIC/MBC) Values of Electrospun TiO2 NFs

The MIC values of TiO2 NFs against *P. aeruginosa* and MRSA was estimated by serial two-fold dilutions of TiO2 NFs from 32 to 1 mg/mL as described previously [31,32]. The determination of MBC values was also investigated as method described in previous studies [32,33].

### *2.4. E*ff*ect of TiO2 NFs on Biofilm Formation*

The antibiofilm potential of TiO2 NFs against *P. aeruginosa* and MRSA biofilm was examined quantitatively in a sterilized 96-well polystyrene (flat bottom) microtiter tissue culture plate using crystal violet assay as described in our previous study [31,33].

### *2.5. E*ff*ect of TiO2 NFs on the Morphology of P. aeruginosa and MRSA: SEM Analysis*

Further, the effects of TiO2 NFs on the morphological features of *P. aeruginosa* and *S. aureus* cells were analyzed by SEM. In Brief, ~106 CFU/mL of *P. aeruginosa* and *S. aureus* cells treated with 1 mg/mL of TiO2 NFs for 18 h were incubated at 37 ◦C [33,34]. After incubation, the treated and untreated samples were centrifuged at 10,000 rpm for 15 min. The obtained pellets were washed with PBS (1×) three times and fixed with primary fixative (i.e., 2.5% glutaraldehyde) for 6 h at 4 ◦C and then further fixed with secondary fixative (i.e., 1% osmium tetroxide) for 1 h. After fixation, the samples were dehydrated by a series of ethanol [34,35]. The cells were then fixed on the aluminum stubs, dried in a desecrator and coated with gold. Finally, the treated and untreated samples were examined by SEM.

### **3. Results and Discussion**

### *3.1. E*ff*ects of the Calcining Atmosphere on TiO2 Colour*

Figure 1 presents a gradual color change from white to dark grey after heat treatment in 100% air and in different mediums of air–argon compositions up to 100% argon medium. This change is likely because of oxygen vacancy defects. The change and the intensification of the color are mainly a result of defects associated with oxygen vacancies that rise from an increase in argon content [36].

**Figure 1.** Color changes in electrospun titanium oxide (TiO2) nanofibers (NFs) in argon-air mixtures.

### *3.2. Microstructure Analysis of the Prepared NFs*

Figures 2 and 3 show the typical SEM and TEM micrographs of the as-spun TiO2 and calcined NFs. The electrospinning process could produce good quality TiO2 NFs, possibly without nodes and defects. The diameter of the as-spun fibers varied between 80 and 600 nm, whereas the estimated average thickness was ~400 nm (Figure 2a). Upon annealing in different mediums of air/argon (100%-0%, 50%-50%, 25%-75%, and 0%-100%), the fibers shrank, and their morphology changed slightly from smooth to rough. This figure also shows the presence of a heterogeneous matrix made up of agglomerated grains for the initial microstructure and leads to faster granular growth. The fibers size was between 50 and 300 nm (Figure 2b–e). Several thin-fibers of about 50 nm were perceived in specimens annealed under 25-75% air-argon. The quality and shape of fiber mats were preserved after calcination as clarified by the TEM images (Figure 3b–e) unlike the electrospun fibers that are often

composed of oxide nanoparticles (Figure 3a) [37]. The as-spun fibers showed organic species, whereas the annealed fibers exhibited a solid morphology with high-quality individual particles in the range of 100 nm. The annealing of TiO2 NFs at 900 ◦C in 50%-50% air-argon led to pure TiO2 fibers formation, which was proven by EDX and TGA characterization techniques. In Figure 4, the EDX spectrum illustrates high-intensity O and Ti peaks and a small Pt peak from the platinum coating on the TiO2 NFs heated in 50-50% air-argon, which is mainly similar to those observed in specimens annealed in 100% air, 25% air-75% argon, and 100% argon. Figure 5 shows the TGA result for samples heated under 50-50% air-argon medium. The PVP polymer and organic material are completely removed from the electrospun TiO2 NFs at ~450 ◦C, and ~100 ◦C, respectively.

**Figure 2.** Scanning electron microscopy (SEM) of TiO2 NFs calcined in different air and argon mixture. (**a**) As-spun TiO2, (**b**) 100% Air, (**c**) 50% Air and 50% Argon, (**d**) 25% Air and 75% Argon and (**e**) 100% Argon.

**Figure 3.** Transmission electron microscopy (TEM) of TiO2 NFs calcined in different air and argon mixture. (**a**) As-spun TiO2, (**b**) 100% Air, (**c**) 50% Air and 50% Argon, (**d**) 25% Air and 75% Argon and (**e**) 100% Argon.

**Figure 4.** An energy-dispersive X-ray spectroscopy (EDX) spectrum of electrospun TiO2 NFs prepared in 50% air-50% argon mixture.

**Figure 5.** Thermogravimetric analysis (TGA) performed for electrospun TiO2 NFs prepared under 50% air-50% argon mixture.

### *3.3. Wide-Band Gap Analysis of Calcined Electrospun TiO2 NFs*

Figure 6 shows the UV-vis DRS spectra of as-electrospun TiO2 NFs calcinated in air-argon media at 900 ◦C and cooling to ambient temperature. Table 1 shows the values of band-gaps at room temperature for various TiO2 NFs. The band-gap value reduced from 3.33 eV for as-spun and non-calcinated samples to about 3.09 eV for the ones calcinated in 100% air. Under various air-argon environments, the value of *Eg* decreased from about 3.09 to 2.18 eV with an increase in argon content. A previous study on similar specimens revealed that the growth of vacancies was minimal and the reduction of *Eg* value was ascribed to the increase in crystallinity [38]. The measured difference agrees with that weighted according to the concentration of pure anatase and rutile phases [38–40]. Alterations in levels and phase mixing and gradual development of oxygen vacancies are two factors that can reduce the band-gap energy with argon introduction. The measured energy gap was 2.18 eV for sample heated in 100% argon and for the phase composition for which the difference according to the concentration would be 3.05 eV. The difference of 0.87 eV is assigned to the development of oxygen vacancies and allows a greater density of charge carriers. The development of oxygen vacancies leads to the creation of Ti3<sup>+</sup> centers or unpaired electrons that generate vacant states under the conduction band [41,42]. The development of oxygen vacancies for different argon concentrations has been previously discussed [38]. When the specimen is annealed in argon, oxygen disappears and the non-stoichiometric anatase (TiO2−x) forms [43]. The formation of oxygen vacancy defects in titanium oxide is induced from the occurrence of new localized states of oxygen vacancies between the conduction and valence bands. The excitation of electrons from the valence band to the vacant oxygen states can be done in visible light. With rising argon amount, the effective *Eg* moves thoroughly to the red region, the specimen is being active under visible light and thus the *Eg* is reduced. So, the mutual effects of the formation of oxygen vacancies and crystallinity treatment have prolonged the excitation of light of electrospun TiO2 NFs from ultraviolet to visible light range without the need of chemical doping.

**Table 1.** Band gap energies for as-electrospun TiO2 nanofibers (non-calcinated), and TiO2 NFs obtained after calcination at 900 ◦C in various air-argon media.


**Figure 6.** UV-vis diffuse reflectance spectrophotometer (DRS) spectra of electrospun TiO2 NFs obtained before calcination and those obtained after calcination in various air-argon media.

### *3.4. Antibacterial and Antibiofilm Activity of TiO2 NFs*

### 3.4.1. MIC and MBC

The microbiocidal activities of TiO2 photocatalysis were reported for the first time by Matsunaga and co-workers in 1985 [44]. They investigated the killing of bacteria and yeast cells in water by employing TiO2-Pt photocatalysts in near-ultraviolet radiation. They reported that the inhibition of respiratory activity was the mechanism for cell death.

In this research work, the antibacterial property (MIC/MBC) of TiO2 NFs calcined with different ratios of air–argon mixtures (i.e., 100% air, 50% air-50% argon, 25% air-75% argon, and 100% argon) has been investigated against *P. aeruginosa* and MRSA (Supplementary Figure S1). The MIC/MBC values of TiO2 NFs heated with different ratios of air-argon mixtures against *P. aeruginosa* and MRSA are presented in Table 2. TiO2 NFs heated in the presence of 25% air-75% argon showed a maximum antibacterial activity and MIC/MBC values against *P. aeruginosa* were 3 and 6 mg/mL and for MRSA it was 6 and 12 mg/mL, respectively (Table 2). Based on the MIC and MBC results, it was observed that Gram-negative *P. aeruginosa* was more susceptible to TiO2 NFs than Gram-positive MRSA. These results agree with results from previous studies [45,46], and may occur owing to differences in their cell wall structures and to bacterial strain growth rate [45–47]. Pigeot-Rémy and co-workers [48] investigated the effects of TiO2 particles against *E. coli K-12* in the dark and reported that the attachment of NPs to bacterial surfaces causes membrane damage and perturbation, which may increase the permeability of the outer cell membrane and the resultant damage to the envelope of bacterial cells leads to bacterial cells death.

**Table 2.** Minimum inhibitory concentration (MIC) and minimum bactericidal concentration (MBC) (mg/mL) values of tested electrospun TiO2 nanofibers against Methicillin resistant *S. aureus* and *P. aeruginosa*.


### 3.4.2. Effects of Electrospun TiO2 NFs on the Morphology of Bacterial Cells

Morphological alterations in Gram-negative P. aeruginosa (Figure 7) and Gram-positive MRSA (Figure 8) after exposure to TiO2 NFs were further examined by SEM. The untreated P. aeruginosa had a normal, rod-shaped structure and regular, smooth and intact cell surface (Figure 7). However, the morphology of *P. aeruginosa* cells was altered considerably, and cells were damaged to different extents after treatment with TiO2 NFs. After 18 h of treatment, the cell envelope and cell wall were rough, irregular, abnormal in form and main damage was categorized by the creation of "pits" and depressions that probably lead to a loss of bacterial cell membrane integrity (Figure 7). Similarly, the untreated Gram-positive MRSA was normal with smooth and regular cell surfaces (Figure 8). However, MRSA cells treated with TiO2 NFs exhibited noticeable alterations and damage and the clusters of NFs were linked and anchored on the surface of bacterial cells (Figure 8). Irregularities, shallows and depressions on the cell envelopes and cell walls of certain MRSA cells suggest that bacterial damage occurred (Figure 8). SEM analysis showed that TiO2 NFs were more effective against P. aeruginosa bacterial cells in comparison with MRSA and were severely injured compared with Gram-positive MRSA. The obtained results may be due to morphological dissimilarities in the cell walls of bacteria. Gram-negative bacterial cells display thin layers of peptidoglycan that facilitate the mobility of metal-ion NPs within cells and facilitate the interaction among NPs and walls of bacterial cells. Gram-negative bacteria exhibit a negative charge due to their high content of lipopolysaccharides. This negative charge attracts and interacts with positive metal ions, which may lead to the NP penetration, intracellular damages and protein and DNA destruction [46]. It was suggested that the interaction of TiO2 NPs with bacterial cells in the dark caused bacterial membrane integrity destruction, especially of lipopolysaccharides [48]. TiO2 NPs form pores in bacterial cell walls and membranes, which increases the permeability and leads to cell death [10]. However, other published work has shown that the contact among metal oxides and bacterial cells provokes oxidation and formation of reactive oxygen groups including O2 •–, •OH, and H2O2. These free radicals attack bacteria cell walls and alter the membrane integrity and permeability, which leads to bacterial cell death [48–51]. It has been reported that the destruction of cell envelope by incorporation of TiO2 NPs inside the cells damages bacterial DNA and RNA, which could provoke cell death [48]. The antimicrobial activity of TiO2 in the absence of photoactivation has been also reported. Nakano and co-worker [51] stated that TiO2 deactivates bacterial DNA and enzymes via coordination of electron-donor groups, like hydroxyls, indoles, carbohydrates, amides, and thiols in the absence of light. Pit formation in bacterial cell walls and envelopes that enhanced the permeability lead to bacterial cell death [51,52]. It has been reported that there is proportional relationship between the light and the antimicrobial activity of TiO2. Senarathna et al [53] and Lee et al [54] reported that the presence of sunlight enhanced the antimicrobial activity of TiO2 against S. aureus might be due to generation of free radicals [53,54].

**Figure 7.** Effect of electrospun TiO2 NFs on the morphological aspects of *P. aeruginosa* as examined by scanning electron microscopy: (**A**) control without any treatment and treated with TiO2 calcined in (**B**) 100% Air, (**C**) 50% Air and 50% Argon, (**D**) 25% Air and 75% Argon; and (**E**) 100% Argon.

3.4.3. Inhibition of Biofilm Formation by TiO2 NFs

The antibiofilm potential of TiO2 NFs heated under different air-argon environments was evaluated at various amounts of 0.25, 0.5, 1.0 and 2.0 mg/mL against MRSA and *P. aeruginosa* biofilms using crystal violet microtiter assays in a 96-well flat-bottom polystyrene plate at OD595 nm. Plots in Figure 9A,B show that TiO2 NFs inhibit the biofilms formation by MRSA and *P. aeruginosa* in a dose-dependent manner. It was reported that a rise in TiO2 concentration provoked a reduction in the cultivability of bacteria [48]. As shown in Figure 9A,B, TiO2 NFs heated in a 25% air-75% argon mixture exhibited the highest biofilm inhibition of about 75.2% for MRSA and 72.3% for *P. aeruginosa,* respectively at 2 mg/mL of TiO2 NFs. These results agree with those reported in previous studies [55,56]. In a previous study, epoxy/Ag-TiO2 nanocomposites were found to inhibit biofilm creation of *S. aureus* ATCC 6538 and *E. coli K-12* by 67% and 77%, respectively [56].

**Figure 8.** Effect of electrospun TiO2 NFs on the morphological aspect of *S. aureus* as examined by scanning electron microscopy: (**A**) control without any treatment and treated with TiO2 NFs calcined in (**B**) 100% Air; (**C**) 50% Air and 50% Argon; (**D**) 25% Air and 75% Argon, and (**E**) 100% Argon.

**Figure 9.** Effect of TiO2 NFs calcined in various air–argon environments (a) 100% Air, (b) 50% Air and 50% Argon, (c) 25% Air and 75% Argon, and (d) 100% Argon on biofilm formation abilities of (**A**) *P. aeruginosa* and (**B**) methicillin-resistant *Staphylococcus aureus* (MRSA).

### **4. Conclusions**

This study focuses on the heat treatment of TiO2 NFs to develop photoactive titanium photocatalysis in the visible spectrum and to evaluate their antibacterial and antibiofilm potential against Gram-negative bacteria *P. aeruginosa* and Gram-positive MRSA. The *Eg* value was 3.09 eV for specimens heated in 100% air and 2.18 eV for the ones heated in 100% argon. The value of *Eg* decreased systematically with rising argon amount in the various air-argon mixtures. The increase in the amount of argon brings the state under the TiO2 conduction band. TiO2 NFs calcined in a 25% air-75% argon environment showed maximum antibacterial and antibiofilm activities. The MIC/MBC and SEM results show that TiO2 NFs were more operative against Gram-negative *P. aeruginosa* than Gram-positive *S. aureus*. The inhibition of biofilm formation by TiO2 NFs shows that TiO2 NFs inhibit the biofilms formation by MRSA and *P. aeruginosa* in a dose-dependent manner. From the obtained data on antibacterial antibiofilm analysis, it has been concluded and suggested that TiO2 NFs can be used in hydrophilic coatings, coating of various inanimate object surfaces, such as metals, glass, medical devices and equipment to prevent biofilm formation on medical devices or medical device-related infections and sepsis, and also can be applied in food packaging, wastewater treatment and purification, self-cleaning and self-disinfecting, killing of bacteria and the removal of toxic and damaging organic compounds from water and air.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2079-6382/9/9/572/s1, Figure S1: represents MHA plates showing MBC values of tested electrospun TiO2 NFs in various air–argon environments.

**Author Contributions:** Conceptualization, H.M.A. and I.-M.L.; investigation, methodology, writing—original draft preparation H.M.A.; M.A.A. (Mohammad Azam Ansari) and S.A.; formal analysis, visualization, software, writing—review and editing, S.A.; M.H.A.; A.T.; N.A.A.; S.A.; Y.S.; M.A.A. (Munirah Abdullah Almessiere); F.S.A.; A.B.; I.-M.L.; data curation, S.A. and Y.S.; Supervision, I.-M.L. All authors have read and agreed to the published version of the manuscript.

**Funding:** The authors acknowledge financial support from Imam Abdulrahman Bin Faisal University (IAU-Saudi Arabia) through the project application No. 2020-086-CED. The authors thank all participants for their effort as well as for their insightful discussions.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Communication* **Biofilm Control Strategies: Engaging with the Public**

### **Joanna Verran 1,\* , Sarah Jackson 1, Antony Scimone <sup>1</sup> , Peter Kelly <sup>2</sup> and James Redfern <sup>3</sup>**


Received: 17 June 2020; Accepted: 29 July 2020; Published: 30 July 2020

**Abstract:** There are few peer-reviewed publications about public engagement with science that are written by microbiologists; those that exist tend to be a narrative of an event rather than a hypothesis-driven investigation. However, it is relatively easy for experienced scientists to use a scientific method in their approach to public engagement. This short communication describes three public engagement activities hosted by the authors, focused on biofilm control: hand hygiene, plaque control and an externally applied antimicrobial coating. In each case, audience engagement was assessed using quantitative and/or qualitative methods. A critical evaluation of the findings enabled the construction of a public engagement 'tick list' for future events that would enable a hypothesis-driven approach with more effective communication activities and more robust evaluation.

**Keywords:** Biofilm; Public Engagement; Outreach; Control Strategies; Oral Biofilm

### **1. Introduction**

It is increasingly being recognised by 'experts' that science literacy is of key importance for the public [1]. At a time where antimicrobial resistance (AMR) continues to pose significant public health threats (or indeed, at a time of a global pandemic), an understanding of statistics, epidemiology and microbiology is even more desirable. As a subject, microbiology offers many topics with which we can engage non-experts, such as microbial diversity (including fungi, algae, protozoa and viruses as well as bacteria), beneficial microbes (for example, probiotics, fermented foods, the human microbiome), and messages that can influence behaviour in a positive manner (including vaccination, hand hygiene, antimicrobial stewardship) [2–4].

Biofilms (an assemblage of microbial cells that are irreversibly associated with a surface—not removed by gentle rinsing—and enclosed in a matrix of primarily polysaccharide material [5]) are of great importance to microbiologists, but also to many other professionals (such as engineers, biocide manufacturers, architects), and are found in a variety of environments (water distribution systems, industrial processing, hospitals). Biofilm research is multi-disciplinary, extensive and significant, with many applications. There are several research centres which focus on biofilm, such as the US-based Centre for Biofilm Engineering (http://www.biofilm.montana.edu/) and the UK-centred National Biofilm Innovation Centre (https://www.biofilms.ac.uk/), and conferences about biofilm are regular and not uncommon. Some individual researchers, research groups and research centres are keen to engage with external public audiences through outreach activities, although evidence of such activities (websites, articles, learning materials and other peer-reviewed outputs) is not easy to find. But why do we want the public to know about biofilms? And what does the 'public' need to know about biofilms? How will we know if our activity has been effective? How can you identify good practice? How can you share success?

Science communication/public engagement can be seen as an emerging discipline, particularly for those scientists who have begun to question the effectiveness of their public engagement work. Evaluation of effectiveness using both quantitative and qualitative methods ('mixed methods') is strongly supported by education researchers [6,7], enabling the assessment of both reach (i.e., numbers) and impact (change in attitudes, perception). There are few peer-reviewed publications on the topic that are written by microbiologists: those that exist tend to be a narrative of an event rather than a hypothesis-driven investigation with appropriate evaluation. However, it is relatively easy for experienced scientists to use a scientific method in their approach to public engagement. This short communication describes three different biofilm-related public engagement activities hosted by the authors, who used lessons learned to develop a tick list for future events to enable more effective communication activities with more robust evaluation.

### *1.1. Activity One: 'Now Wash Your Hands'*

'Now wash your hands' was developed as part of a University faculty family fun day during National Science and Engineering Week/Healthcare Science Week in the UK. The aim was to raise awareness of effective handwashing, whilst also engaging the participants in a discussion about the skin microbiome/biofilm. This event guarantees an audience of predominantly families who are likely to have an existing interest in science. Hand hygiene activities are well established as interactive learning activities with demonstrable public health impact (for example, as an intervention in reducing the spread of coronavirus [6]). In this activity, demonstrators (academic staff and student volunteers) engaged audiences to demonstrate surface contamination and effective handwashing (Figure 1). Thus, visitors at this activity (in a walkway area) had their hands 'contaminated' with a UV hand gel (www.hand-washing.com). This kit uses a fluorescent dye and ultraviolet light to illustrate the transmission of 'germs' from hands to other surfaces (and vice versa) and the importance of handwashing. In addition, the participants were invited to press their hands onto large agar plates for subsequent incubation to reveal the culturable microorganisms present on their skin. Of course, they were unable to see the results of this work until after incubation, thus images of plates pre-inoculated with microorganisms present on hands and mobile phones [7] were available to view, and post-incubation images of their own plates were uploaded to Flickr, a social media site that hosts images (http://tinyurl.com/howcleanareyourhands, Figure 2). Within a week from results going online, almost 100 downloads were recorded (the participants were provided with a card/web address), equivalent to the number of plates inoculated. From this, we deduced that visitors demonstrated interest and engagement with the activity. Throughout the activity, conversations were ongoing. It was unfortunate that these interactions were not noted in some form: informal observations revealed points of interest from the participants such as their inability to clean hands effectively (especially the adults!) and amazement at the mobile phone contamination. The handprint technique has been used as an engagement tool for other events, such as an art installation called 'Hands across the cultures' for registrants to a qualitative research conference and as part of the 'bioselfies' project (https://blogs.bl.uk/science/2020/02/introducing-bio-selfies-11-february-2020.html) initiated by the University of Salford. Flickr has been used for other events that require incubation of plates [8,9], and download numbers have on occasion exceeded the number of images posted, showing that the participants may have been sharing the findings with others. The fluorescent hand technique was used to illustrate person-to-person transmission by handshaking prior to a screening of the movie *Contagion* (directed by Soderbergh, 2011). One person 'contaminated' his/her hands, shook the hand of their neighbour, who shook her/his neighbour's hand and so on. Thus, the passing-on of fluorescence was used to illustrate the transmission of infection through poor hand hygiene, reinforcing the message as to how the movie pandemic was initiated (hand contact).

**Figure 1.** Activity one: 'Now wash your hands': the audience engaged in hands-on activities focusing on the topic of hand hygiene. Here, a participant's hands can be seen during the use of the UV glow gel.

**Figure 2.** Example of the images uploaded to the Flickr page following the 'Now Wash Your Hands' event. Each image represents the handprint of one participant, revealing the range of microorganisms present on the hand.

Hand hygiene activities are common in microbiology engagement, the aim of the activity being primarily to inform, and hopefully to change, participants' behaviour so that effective handwashing techniques are employed. Explanation regarding the presence or importance of the

skin microbiome/biofilm are likely rare (especially if the results are not available until a later date): the activity is inevitably more focused on the removal of temporary contaminants and on the importance of good handwashing. Some discussion could take place regarding the hygiene-versus-cleanliness hypothesis [10,11]. The Flickr method used for posting images and monitoring downloads at least gives an indication of interest, but much more could be made of this activity. It would also be interesting to know if the 'good handwashing' messages are retained and employed in the future. However, longitudinal studies are rare in this type of public engagement, probably because of the significant advanced planning required in terms of gaining approval for personal data access (e.g., emails) and also because only short-term awareness raising tends to be the primary aim of the activity.

### *1.2. Activity Two: Plaque Attack!*

The plaque biofilm is one of the best-known medical biofilms [12,13], and oral hygiene advertising frequently provides cartoons of plaque being removed to demonstrate the effectiveness of a paste, mouthwash or brush. It is known that good toothbrushing helps to remove plaque [14] and should be carried out regularly. Different dentifrices claim varying activities, but virtually all formulations include fluoride (to 'strengthen the teeth') [15], and many contain antimicrobial agents (to reduce the number of microorganisms, with claims around gum health) [16].

'Plaque attack!' was a laboratory-based activity designed for children and their parents, taking place during Manchester Science Festival's family fun day at Manchester Metropolitan University. The aim of the event was to encourage good oral hygiene but also to captivate visitors with the components of the plaque biofilm as well as the laboratory and its equipment. Being time-consuming and space-limited, the participants had to register for the event, were limited to 3 groups of 20 participants, be escorted to the laboratory, provided with appropriate clothing and instruction and supervised at all times. Oral microbiology is a key research area in our laboratories, and the delivery team thought it would be valuable for visitors to encounter activity in a working (teaching) laboratory. The delivery team comprised PhD students, technical staff and an academic. Several activities were conducted as part of a 'round-robin' activity: sampling plaque (microscopy demonstration and take-home photo [ZIP Mobile Printer, Polaroid]); disclosing plaque (using commercially available disclosing tablets), with photographs taken before and after cleaning teeth (in a wash area adjacent to the laboratory); looking at cultures of oral bacteria on agar plates; investigating biofilm structure/building a biofilm (using 'Model Magic' [Crayola Bedford UK], a white air-drying modelling clay) (Figure 3a); and destroying a biofilm (using a water pistol to remove plaque (whose microorganisms were pre-constructed from Fimo, a multi-coloured clay which can be hardened in the oven [www.staedtler. com]) hampered by plaque matrix (a translucent hair gel) [17] (Figure 3b). The participants were provided with a basic information sheet on plaque and oral hygiene, onto which they could attach their Polaroid images. They were also given a bag containing complimentary toothbrush and toothpaste (courtesy of Unilever [www.unilever.co.uk]). At the end of the activity, they were asked for free text feedback on what they thought of the event, and the information was coded into categories to allow for comparison [18,19] (Figure 4). The participants were particularly engrossed in the microscopy demonstration, being able to see their own plaque at high magnification. They also clearly had fun 'destroying' the biofilm but were less interested in the more passive/less exciting activity (agar plates demonstration, building a biofilm). The free text provided by the participants (allowing more thorough insight compared to multiple-choice or leading questions such as 'give three things you have learned', or 'smiley face/sad face' evaluations [18,20]) gave valuable qualitative information that was used to inform subsequent activities.

ǻǼȱ

ǻǼȱ

**Figure 3.** (**a**/**top**) Participants at the 'Plaque attack!' event were encouraged to create their own oral bacteria flora from modelling clay, which was assembled into the oral biofilm representation here shown. (**b**/**bottom**) Participants were encouraged to 'destroy a biofilm' by removing bacteria (coloured plastic pieces) encased in biofilm extracellular matrix (hair gel) with a spray bottle filled with water.

**Figure 4.** Themes identified from 'Plaque attack!' feedback. There was a total of 19 comments that were coded based on their focus—with each comment possibly being coded into more than one category.

### *1.3. Activity Three: A Photocatalytic Wall*

Our research into titanium dioxide coatings included a range of laboratory-based studies that compared different titanium dioxide concentrations in paint formulations [21]. The work described in this paper was to see whether the effect of a photocatalyst in paint could be detected by the human eye. Thus, as part of a PhD project investigating the activity of photocatalytic surfaces, one of the external walls of the University was used to illustrate the effectiveness of titanium dioxide paints in terms of self-cleaning and reduction of the formation of biofilm on the wall material. Photocatalytic material such as titanium dioxide can exhibit self-cleaning, anti-fouling and antimicrobial properties in the presence of light, which makes these materials excellent candidates for incorporation into urban buildings and infrastructure [22–24]. The self-cleaning properties stem from their superhydrophilic nature—as, for instance, that of a liquid (e.g., rain) rolling off the surface of a continuous body. This sheeting carries away dirt and debris, cleaning the surface in the process—as seen in the Sydney Opera House [25]. Thus, biofilm formation on the surface is delayed or prevented.

In our study, the wall, comprising concrete panels (smaller panels 190 cm × 76 cm, larger panels 406 cm × 76 cm) on a 1970s University building, was west-facing (location on Chester Street, Manchester, UK M1 5GD). Six of the panels were painted with a siloxane external paint formulation that contained or lacked the photoactive pigment (kindly provided by Tronox, www.tronox.com). Our aim was to inform the passing public about our research (an interpretation panel was affixed to the wall), and on occasion, we encouraged passers-by to participate in a longitudinal subjective assessment of the impact of titanium dioxide-containing paint on the perceived cleanliness of the panel. This engagement activity was done directly by interview and indirectly using photographs at specific times over a 44-month period.

Initially there was no apparent difference in the brightness of the painted panels (Figure 5a). Members of the public attending a Manchester Science Festival event (October 2014) were asked to rank the painted panels in order of cleanliness/whiteness, with 1 being most clean, and 6 being least clean (n = 18). The experiment was also conducted via a social media platform (Facebook), with participants asked to assess whiteness using photographs (n = 48). The direct assessment was repeated after three years (n = 21). In all cases, the participants ranked two or three of the photocatalytic panels as the 'whitest'. In 2014, around 60% of the participants selected the three photocatalytic panels correctly. In 2017, this figure rose to 78%. After six years, the test-paint panels appeared whiter than the control panels (Figure 5b, May 2020).

The presence of the wall with its accompanying information panel at the side of the University Science and Engineering building provided a useful pointer to introduce visitors to some of the research ongoing in the faculty. The use of the public to assess the cleanliness of the wall proved unnecessary within a few months, when the impact of the test paint was apparent. The fact that almost all participants could discriminate between the panels after less than 12 months was also of interest. This approach might therefore be useful in the future for the assessment of test formulations.

ǻǼȱ

ǻǼȱ

**Figure 5.** Images of the wall at Manchester Metropolitan University used in the study of photocatalytic paint (panels labelled 1–6). Panels 1, 3 and 6 were painted with photocatalytic paint, whilst panels 2, 4 and 5 were painted with paint that did not contain the photocatalytic agent. The image on the top (**a**) was taken in 2014, eight months following the application of the paint: whiteness/brightness difference between the two paint types is hard to distinguish. The lower image (**b**) was taken six years later (2020); panels painted with photocatalytic paint are visibly brighter compared to control paint panels.

### **2. Discussion**

Much was learned from each event (as noted above), particularly through observation, in terms of what components participants like and engage with when discussing biofilm. In addition, quantitative evidence of engagement was derived from the 'Now wash your hands' event; qualitative evidence of enjoyment and engagement was obtained from 'Plaque attack', and the potential for acquisition of research data was indicated by the photocatalytic wall activity. These various outcomes informed how subsequent events for the public would take place, with more focus on design, delivery and evaluation.

More recently, there has been increasing effort to ensure that these criteria for effective public engagement are met. Microbiology has a particularly dynamic approach to public engagement, and many teams are now publishing the outcomes of their public engagement research in peer-reviewed journals, magazines or online. Yet, in a review of public engagement activity around AMR, a rich bedrock of activity was found only through personal contacts and communication rather than through a literature search [4]. It is even more important when talking to audiences about biofilms that intended messages are clear. Thus, we describe in Table 1 the planning of a hypothetical public engagement event designed to inform a large number of adults about biofilm and AMR. Our focus was on the combination of the two phenomena, which occurs, for example, when biofilms on medical devices present increased resistance to antibiotics [26]. In order to address this combined effect, it was first necessary to define the two phenomena separately. We particularly wished to avoid intrusive aspects of evaluation, relying instead on observation and other (subjective and objective) indicators from participants. We hope that this checklist may be useful for others who might wish to engage audiences with their biofilm/antibiotic research.

The National Biofilm Information Centre has recognised the importance of public engagement and is providing a hub for the dissemination of biofilm-focused outreach and engagement activities, which will enable, over time, ideas, expertise and outcomes to be shared and developed, in order to improve the effectiveness of engagement encounters for scientists and their audiences alike. We hope that our experiences in the area are of interest in this context.


**Table 1.** Checklist for public engagement events, with accompanying information detailing planning for a proposed event focusing on antimicrobial resistance (AMR) and biofilms.


**Table 1.** *Cont*.

### **3. Conclusions**

Public engagement activities can be designed with clear aims that enable effective evaluation using both quantitative and qualitative methods. This is particularly important for complex phenomena such as biofilms and AMR.

**Author Contributions:** Conceptualization, J.V., P.K. and J.R.; Methodology, J.V., P.K., J.R.; Formal Analysis, J.V., J.R., A.S.; Investigation, J.V., J.R., A.S., S.J.; Resources, J.V. and P.K.; Writing – Original Draft Preparation, J.V. and J.R.; Writing – Review & Editing, J.V., A.S. and J.R.; Visualization, J.V., J.R. and A.S.; Supervision, J.V. and P.K.; Project Administration, J.V. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding

**Conflicts of Interest:** The authors declare no conflict of interest

### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel. +41 61 683 77 34 Fax +41 61 302 89 18 www.mdpi.com

*Antibiotics* Editorial Office E-mail: antibiotics@mdpi.com www.mdpi.com/journal/antibiotics

MDPI St. Alban-Anlage 66 4052 Basel Switzerland

Tel: +41 61 683 77 34 Fax: +41 61 302 89 18