Evaluation of Changes Induced in the Probiotic Escherichia coli M17 Following Recurrent Exposure to Antimicrobials

Introduction: It is already well known that the exposure of certain bacteria, pathogenic or not, to antimicrobials is likely to increase their virulence and induce the development of direct or cross resistance to antimicrobials, but there is almost no information available regarding probiotics. Aim: To assess the changes induced in susceptibility to antibiotics, biofilm formation, growth rate and relative pathogenicity in the probiotic Escherichia coli M17 (EC-M17) after long exposure to antimicrobials namely ampicillin, kanamycin, cefazolin and silver nanoparticles (AgNPs). Methods: After determining the minimum inhibitory concentrations (MIC) and minimum bactericidal concentrations (MBC) of the 4 antimicrobials above-mentioned by the microdilution method, ECM17 was exposed to increasing subinhibitory doses ranging from MIC/8 to MIC for 8 days. The susceptibility to antibiotics of the mutants obtained was assessed by the Kirby Bauer disc diffusion method, biofilm formation by the Congo red agar method and with crystal violet bacterial attachment assay, and relative pathogenicity was assessed using a Galleria melonella waxworm model. Results: Exposure to antimicrobials induces noticeable changes in EC-M17. The highest adaptation to antimicrobials was observed on AgNPs with 8-fold increase in MIC and 16-fold Original Research Article Mbarga et al.; JPRI, 33(29B): 158-167, 2021; Article no.JPRI.69063 159 increase in MBC of AgNPs. EC-M17 exposed to ampicillin, kanamycin and silver nanoparticles became resistant to ampicillin, ceftazidime, ceftazidime/clavulanate and tetracycline while exposure to cefazolin induced a significant decrease in sensitivity to tetracycline and ampicillin and resistance to ceftazidime/clavulanate and ceftazidime. The strain exposed to ampicillin was the only one to produce more biofilm than the control strain and except the EC-M17 exposed to cefazolin, all other EC-M17 strains were more pathogenic on G. melonella model than the control. Conclusion: Data in this investigation suggest that repeated exposure of the probiotic EC-M17 to antimicrobials may induce changes in antimicrobials susceptibility, biofilm formation, growth rate, and relative pathogenicity. Therefore, as far as possible, the probiotic E. coli M17 should not be used in combination with antibiotics and further investigations are required to expand similar work on more probiotics in order to avoid resistance build-up which might be transmitted by horizontal transfer.


INTRODUCTION
The probiotic Escherichia coli M17 (EC-M17) was identified in Russia more than 90 years ago. EC-M17 is a non-pathogenic bacterium [1,2] of the family Enterobacteriaceae and a non-sporeforming gram-negative rod of the serotype O2 with flagellar antigen H type 41 [2]. EC-M17 should not be confused with E. coli Nissle 1917. Although no direct comparison has yet been made to establish the differences between these two strains, Kokesova et al. [3] and Fitzpatrick et al. [2] reported that Nissle strain differs from M17 in both serotype (O6 v. O2 for M17) and flagellar H antigen (H1 v. H41). The parent strain EC-M17 was deposited under the designation ATCC 202226 in the American Type Culture Collection [2].
EC-M17 is used in both humans and animals for its beneficial effects on the gastrointestinal tract [4]. This strain is particularly known for its effectiveness against a variety of gastrointestinal problems, including inflammatory bowel disease (ulcerative colitis or Crohn's disease) and diarrhea in infants and children [2,5]. EC-M17 has been reported to have had a positive result in the treatment of dysentery in thousands of individuals [6,7]. Moreover, the toxicity, immunogenic properties and protective activity of the live culture of EC-M17 and antigenic preparations obtained from cell suspensions of this strain have been studied by Levina et al. [3], and the results showed the capacity of EC-M17 to induce the production of 02 and H6 antibodies and to play a significant immunomodulatory role in rabbits. In addition, a study conducted by Fitzpatrick et al. [2] demonstrated the beneficial effect of EC-M17 combined with an antibiotic (metronidazole) on the immunomodulation and attenuation of murine colitis, probably due to an inhibitory effect on NF-kB signaling. However, this study did not attempt to assess the consequences that such a combination could have on EC-M17 itself. Indeed, several recent studies have reported that microorganisms, including E. coli, whether pathogenic or not, are likely to develop resistance to antibiotics and increase the expression of their virulence factors following exposure to biocides and antimicrobials more globally [8-13].
Since there is almost no information on the potential impacts of antimicrobials on probiotics, this investigation aimed to assess the changes induced in susceptibility to antibiotics, biofilm formation, growth rate and relative pathogenicity in the probiotic Escherichia coli M17 (EC-M17) after long exposure to antimicrobials namely ampicillin, kanamycin, cefazolin and silver nanoparticles (AgNPs).

Bacterial Strains and Culture Conditions
In this investigation, we used the standard strain Escherichia coli M17 provided by the laboratory of microbiology and virology of the Peoples' Friendship University of Russian. All the cultures were made on BHIB (Brain Heart Infusion Broth) (HiMedia™ Laboratories Pvt. Ltd., India) and Muller Hinton Agar (MHA HiMedia™ Laboratories Pvt. Ltd., India) and incubated aerobically at 37°C for 18-24h.

Stocks Solution of Antibiotics and AgNPs
Stock solutions of each antimicrobial were prepared at a concentration of 1024 µg/ml and dilutions were made as needed. Ampicillin, cefazolin and kanamycin were prepared in physiological water (NaCl 0,9%) and 2 nm silver nanoparticles (Nanoserebro Argitos, OOO NPP Sintek Nano, Russia) were prepared in distilled water. All the solutions were sterilized by microfiltration (0.45 μm) prior to use.

Determination of Minimal Inhibitory Concentration (MIC) and Minimal Bactericidal Concentration (MBC)
The MIC and MBC were determined as previously reported [14]. In brief, two 5-ml overnight cultures of test bacteria were prepared in BHIB prior to overnight incubation (18 to 24h) at 37°C. Cultures were diluted to a turbidity equivalent to 0.5 McFarland, then the stock solutions of antimicrobials (100µL) abovementioned were submitted to serial twofold dilutions in sterile BHIB (100µL) on U-bottom 96well microplates prior to addition of the bacterial inoculum (10 µL). The plates were incubated overnight (18 to 24 h) at 37°C. The MIC was defined as the lowest concentration of antimicrobial at which growth was completely inhibited (viewed as turbidity relative to that for a sterile negative control). Further, MBCs were determined by subculturing the wells without visible growth (with concentrations ≥ MIC) on MHA plates. Inoculated agar plates were incubated at 37°C for 24 h. MBC was considered the lowest concentration that did not give any bacterial growth on agar.

Long-term Exposure of Bacteria to Antibiotics and Silver Nanoparticles
Bacteria were exposed to increasing concentrations of ampicillin, cefazolin, kanamycin and AgNPs using U-bottom 96-well microplates. 8 concentrations of each antimicrobial were prepared in sterile BHIB then the mixture was sterilized by microfiltration prior to use. The concentrations varied from MIC/2 to MIC with an increment of MIC/16. For each antimicrobial, 200 µL of the preparation of MIC/2 concentration was introduced into line 1 of the microplate, then MIC/2 + 2MIC/16 in line 2, MIC/2 + 3MIC/16 in line 3, etc. 15 µL of overnight culture of Escherichia coli M17 prepared at a concentration equivalent to 0.5 of McFarland was inoculated in the first line and after 24h of incubation at 37°C, the wells of line 1 were homogenized and 15 µL was transferred to line 2 of the corresponding columns and the same operation was repeated for the following lines until the 8th day. Bacteria passaged 8 times on antimicrobial-free medium were also included and were considered as the controls. During incubation, the microplates were placed in a container containing distilled water to limit water loss by evaporation. After successive passages, the bacteria were kept at −80°C in cryovials (Cryoinstant; Deltalab, Spain) for subsequent testing.

Sensitivity of Bacteria to Antibiotics
The modified Kirby-Bauer's disc method described by Mbarga et al. [15] was used to assess the sensitivity to antibiotics of the original Escherichia coli M17 and the mutant strains obtained. Briefly, after bringing the bacteria to room temperature, they were cultured at 37°C for 24 hours in sterile BHIB. 1.5 ml of each overnight culture was centrifuged (Eppendorf Centrifuge 5415 R) for 10 minutes at 3000 RCF and the pellet was collected, washed 3 times with Phosphate buffer saline (PBS) and resuspended in 5 ml of physiological water to obtain a concentration equivalent to 0.5 McFarland. 100 µL of the culture was plated on Muller Hinton Agar (MHA) (HIMEDIA®, Ref 173-500G) and the antibiotic discs were placed aseptically using a dispenser. After 18-24 hours of incubation, the inhibition diameters were measured and interpreted referred to the Clinical & Laboratory Standards Institute [16]. The petri dishes were again incubated for 48 hours at 37°C and the bacteria of the second growth in the inhibition zones were isolated and subjected to a second antibiogram as described above. The 11 antibiotics used were: tetracyclin (TE), 30 μg/disc, ceftazidime/clavulanic acid (CAC), 30/10 per disc; ceftazidime (CAZ), 30 µg/disc; ceftriaxone (CTR), 30 μg/disc; ciprofloxacin (CIP), 30 μg/disc; imipenem (IMP), 10 μg/disc; nitrofurantoin (NIT), 200 μg/disc; trimethoprim (TR), 30 μg/disc; Fosfomycin (Fo), 200 μg/disc; Amoxicillin (AMC), 30 μg/disc and ampicillin (AMP), 25 μg/disc.

Evaluation of Biofilm Formation by Crystal Violet Bacterial Attachment Assay
The biofilm formation of original strains and exposed strains was assessed in sterile 96-well microtiter plate. 200µL of sterile BHIB was introduced in each well and was inoculated by the corresponding overnight culture (18 to 24 h at 37°C and 100 rpm) centrifugated and resuspended in physiological water to obtain a turbidity equivalent to 0.5 of McFarland as described above. Sterile controls were also included. The plates were incubated statically for 48 h at 37°C. 100µL of the medium was transferred in the corresponding well in another microtiter plate for planktonic measurement. The remaining medium was removed from the wells and replaced with 200µl of 1% (w/v) crystal violet solution during 90s. The wells were rinsed three times with distilled water prior to drying at 37°C. The biofilm-bound crystal violet was solubilized in 200 µl of 100% ethanol and the A450 was determined and compared with the negative controls. Each test was repeated 6 times and each repeat was read 3 times.

Planktonic Growth Rate Measurement
Overnight suspensions of original and each adapted strain were diluted 1:10 in BHIB (200ml) and incubated at 37°C with shaking at 150 rpm (Heidolph Inkubator 1000 coupled with Heidolph Unimax 1010, Germany). Samples were regularly removed and diluted as appropriate. Optical density was measured at 490nm.

Galleria melonella Pathogenesis Assay
Relative pathogenicity was assessed as described by Henly et al.
[8] with slight modifications. Briefly, final larval-stage G. melonella (ECO BAITS, Moscow, Russia) was stored in the dark at 4°C for less than 7 days, before randomly assigning 20 to each treatment group and incubating at 37°C for 30 min. After this acclimatation phase, aliquots (15µL) of each suspension (eq 0.5 of McFarland) of overnight Escherichia coli M17 strains were injected into the hemocele of each larva via the last left proleg. Larvae were incubated at 37°C in sterile petri dishes and the number of surviving individuals was recorded after 12, 24, 48 72 and 96 hours. The group injected with sterile PBS were used as negative controls. The larvae were considered dead when were unresponsive to touch and appeared black

Statistical Analysis
All experiments were carried out at least in triplicate. The statistical significance was set at p≤0,05 and all the statistical analysis were carried out using the statistical software XLSTAT 2020 (Addinsof Inc., New York, USA). All the graphs were plotted by Excel software or SigmaPlot 12.5 (Systat Software, San Jose, CA, USA).

Changes in MIC and MBC after Long Exposure to Ampicillin, Cefazolin, Kanamycin and Silver Nanoparticles
Minimum Inhibitory Concentrations (MICs) and Minimum Bactericidal Concentrations (MBCs) were determined for Escherichia coli M17 before and after exposure to kanamycin, cefazolin, ampicillin and Silver nanoparticles (AgNPs) ( Table 1). The results were interpreted as the fold change relative to the original strain. No variation was observed in the MICs and MBCs between the initial E. coli M17 and it analogue which underwent 8 repeated passages in the BHIB antimicrobials-free. In contrast, all E. coli M17 exposed to antimicrobials developed direct and cross resistance. With regard to the MIC, we observed ≥2-fold increase in 2/4 strains for ampicillin and cefazolin in 3/4 strains for kanamycin and AgNPs. Regarding the MBCs, for ampicillin there was 2-fold increase in bacteria exposed to kanamycin and ampicillin while for Cefazolin there was 4-fold increase for bacteria exposed to cefazolin and 2-fold increase for bacteria exposed to ampicillin and kanamycin. Interestingly, direct adaptation appears to be more significant on E. coli M17 exposed to AgNPs with 8-fold increase in MIC and 16-fold increase in MBC of AgNPs.

Changes in Susceptibility to Other Antibiotics
The original strain E. coli M17 and the control (passed in BHIB-antimicrobial free) were both sensitive to all antibiotics tested. However, exposure to antimicrobials has led to significant changes in susceptibility to some antibiotics (Fig.  1). Indeed, E. coli M17 strains exposed to ampicillin, kanamycin and silver nanoparticles became resistant to ampicillin, ceftazidime, ceftazidime/clavulanate and tetracycline. In addition, exposure to cefazolin induced a significant decrease in sensitivity to tetracycline and ampicillin and resistance to ceftazidime and ceftazidime /clavulanate. No significant change in susceptibility was observed on Fosfomycin and trimethoprim while changes occurred in susceptibility to ciprofloxacin, Ceftriaxone, nitrofurantoin and imipenem without causing resistance.

Biofilm Formation, Growth Rate and Relative
Pathogenicity on

G. melonella Model
As shown in Figs. 2 and 3, the original bacteria and the control were both biofilm producers. However, exposure to kanamycin, cefazolin and AgNPs resulted in decreased biofilm formation while the strain exposed to ampicillin became more biofilm-producing than the original bacteria.
Interestingly, as reported by others [17,18] there was a good correlation between the evaluation of biofilm formation on Congo red agar (Fig. 2) and the assessment by crystal violet attachment assay (Fig. 3). In addition, compared to the original strain, E. coli M17 exposed to ampicillin grew faster, followed by the strain exposed to kanamycin while all others grew relatively similar to the control (Fig. 4). Finally, the pathogenicity assay performed with the Galleria melonella waxworm model showed that exposure to Cefazolin significantly attenuates the potential pathogenicity of E. coli M17 while exposure to kanamycin, ampicillin and AgNPs tend to accentuate this pathogenicity (Fig. 5).

DISCUSSION
It is very common to determine the susceptibility to antibiotics of probiotics when studying their probiotic suitability [19,20] and for their potential combination with antibiotics [20,21]. However, to date no study in the literature has been focused on the consequences that exposure to antimicrobials could have on probiotics. In this study we evaluated the impact of antimicrobial exposure in the probiotic Escherichia coli M17. Data of this investigation suggest that longexposure of the probiotic E. coli M17 to antimicrobials may influence it antimicrobials susceptibility, biofilm formation, growth rate, and even relative pathogenicity. These results are similar to those reported on common bacteria [8-13].
The potential causes of the changes in susceptibility observed in this study may be attributed to the stress response [8], structural changes on the membrane which changes the interaction with antimicrobial agents [8], the activation of efflux pumps that make bacteria able to expel antimicrobials from the cell [8,11].
Adaptation to AgNPs was particularly high with 8fold increase in MIC and 16-fold increase in MBC of AgNPs. Previous investigations have also documented the selection of silver resistance in Gram-negative pathogens [22], including E. coli and other Enterobacteriaceae [23]. This resistance has been correlated with increased efflux activity [24] or a loss of outer membrane porins [25], thereby decreasing cell permeability which may explain the decrease induced in AgNPs susceptibility observed in the E. coli M17 exposed to AgNPs.
Regarding biofilm formation, only the E. coli M17 exposed to ampicillin produced more biofilm than the original strain. Machado et al. [26] and Henly et al. [8] reported that adaptation to antimicrobials is correlated with an increase in biofilm biomass in E. coli, which is believed to be due to an increase in protein and polysaccharide content within the extracellular polymeric substance (EPS). This change in EPS composition may lead to reduced Ampicillin susceptibility, as observed in our Ampicillinadapted isolates. Moreover, Microarray analysis performed by Niba et al. [27] revealed that antimicrobial exposure resulted in an increase in the expression of fimDFHI, which encodes proteins involved in fimbrial biosynthesis, which has been shown to be positively associated with an increase in biofilm formation. This may provide a potential link between the increase in biofilm formation and the resistance caused by ampicillin exposure in E. coli M17.
In general, the probiotic E. coli M17 has low virulence and moderate toxicity [1]. However, the attenuation of the relative pathogenicity observed in the G. melonella model following exposure to cefazolin is one of the most interesting findings of this study. This presents prospects for attenuating virulence, not only for probiotics but also for pathogenic bacteria, and should be further investigated and extended to more microorganisms. Furthermore, further studies should be conducted to elucidate the mechanism of action involved in this attenuation of pathogenicity.

CONCLUSION
In this investigation we evaluated the changes induced in the probiotic E. coli M17 following exposure to ampicillin, kanamycin, Cefazolin and silver nanoparticles. The results revealed that these antimicrobials may induce significant changes on this probiotic, including an increase in the relative pathogenicity (assessed here on Galleria melonella waxworm model), except the strain exposed to cefazolin. Therefore, as far as possible, the probiotic E. coli M17 should not be used in combination with antibiotics and further investigations are required to expand similar work on more probiotics in order to avoid resistance build-up which might be transmitted by horizontal transfer.

ETHICAL APPROVAL
Not applicable