Multidrug-Resistance and Virulence-Related Properties of Diarrheagenic Escherichia Coli in Urban River: A Possible Source and Dissemination of Human Infections

The presence of multi-drug resistant (MDR) E. coli harboring virulence pathotypes in aquatic systems is a public health concern due to an increase number of cases of infections and outbreaks in industrialized and developing countries. The aim of the present study was to evaluate the microbiological quality of Joana river, located at Rio de Janeiro, by analyzing E. coli bacteria contamination and to investigate virulence properties and MDR proles by phenotypic and genotypic methods, including bacterial interaction with Caco-2 cells. A total of 34 E. coli were identied by MALDI-TOF and 20 E. coli were characterized as MDR when submitted to antimicrobial susceptibility test. Evaluation by multiplex-PCR of MDR E. coli demonstrated the presence of virulence pathotypes: EHEC (stx1, stx2, eae genes), STEC (stx2 gene) and EIEC/STEC (stx2, iaL genes). Virulence potential was demonstrated by the ability to adhere and survive within Caco-2 cells of MDR E. coli pathotypes (n = 4). In conclusion, this study demonstrates the presence of diarrheagenic MDR E. coli in river water at Rio de Janeiro. The possibility of aquatic environment dissemination of antimicrobial resistance and human contamination leading to community and nosocomial infections due to virulent MDR E. coli water-borne pathogens is a matter of concern. cephalosporins: cephalothin, cefazolin, cefoxitin, cefuroxime, cefotaxime, ceftriaxone, ceftazidime, cefepime; aminoglycosides: gentamicin, amikacin, kanamycin, tobramycin; ampicillin; piperacillin/taxobactam, amoxicillin/clavulanic acid, ampicillin/sulbactam; uorquinolones: ciprooxacin, noroxacin; carbapenems: imipenem, ertapenem, meropenem; aztreonam;

Several reports indicate ETEC as a major cause of diarrheal illness in poor areas of the world where they contribute to unacceptable mortality, particularly among young children (Fleckenstein; Kuhlmann, 2019); EHEC strains secret Shiga toxin (Stx) which can lead to hemolytic uremic syndrome; EPEC is an important cause of infant diarrhea and mortality worldwide; EIEC strains are involved in invasive intestinal infections and dysentery; EAEC strains cause persistent diarrhea due to a heatstable enterotoxin activity (Abe et al., 2008; Serapio-Palacios, Finlay, 2020); STEC may cause outbreaks, sporadic cases of hemorrhagic colitis and are associated with hemolytic uremic syndrome and thrombotic thrombocytopenic purpura (Beutin et al., 2004). Pathogenic E. coli strains may colonize human tissues by attaching and/or invading host cells. Clinical manifestations induced by each of these strains are associated with watery form of diarrhea or in ammatory presentation of the disease ( Population heterogeneity of E. coli has been linked to environmental changes and genome plasticity evolution of some lineages associated with human diseases due to new combinations of virulence genes and phenotypic diversity, contributing to survival, higher virulence pro les and multi-resistance dissemination (Ashbolt, 2004;Brito et al., 2008;Hartland and Leong, 2013).
Although bacterial interaction within gastrointestinal human tract is essential to maintain members of the normal micro ora, it is also considered a critical phase in all diarrheal infections caused by pathogenic E. coli strains (Kalita, Hu, Torres, 2014). Epithelial cell invasion is a virulence mechanism expressed by EIEC strains leading to dysentery-like illness. Invasive properties to cultured epithelial cells have been also reported for EPEC strains (Luck et al., 2005). Therefore, understanding the occurrence of pathogenic E. coli and their ability to adhere, invade and persist in the host cell will improve the knowledge of environmental transmission media to humans and its important role in emergence outbreaks creating potential threat and becoming a public health risk (Xiong et al., 2015). However, further studies remain necessary.
The aim of this present study was to evaluate the microbiological quality of Joana river, located at Rio de Janeiro metropolitan area, by analyzing E. coli bacteria contamination and to investigate virulence-related properties and multidrug-resistance pro les by phenotypic and genotypic methods, including host-cell interaction.

Materials And Methods
Study area and sample collection.
Water samples were collected from Joana River located in front of a University Hospital at Rio de Janeiro metropolitan area, Southeast Brazil, which receives water from different sources including residences, State University, rain and hospital sewage.
Joana River has a total length of 3.412 meters, with 2.400 meter of tunnel, which is considered the largest urban drainage tunnel in Brazil. A ow limiter was built that allows a passage of up 7m³/s of water from the Joana River to the Macaranã River until it ows directly into Guanabara Bay ( Figure 1).
Approximately 100 mL of water samples were collected, stored in sterilized bags and transported to laboratory for analysis.
Samples were inoculated in 100mL of Brain Hearth Infusion broth -BHI (2x), incubated for 24h at 37ºC, and subsequently cultivated onto MacConckey ágar (24h at 37ºC), as previously described (Nogueira et al., 2015). E. coli strains were identi ed using Matrix Assisted Laser Desorption Ionization Time-Of-Flight (MALDI-TOF) mass spectrometry. This method analyzes the pro les of bacterial macromolecule that are obtained from whole bacteria. The procedure provides a unique mass spectral ngerprint of the microorganisms, biopolymer molecules normally present in the condensed phase be converted into intact, isolated ionized molecules in the gas phase. Then, ions are separated according to their molecular weight after migration in an electric eld. Each molecule detected is characterized by molecular mass, the charge, the ratio mass/charge and the relative intensity of the signal (Carbonelle et al., 2011).
Bio lm formation on hydrophobic polystyrene surface.
Bio lm assays on polystyrene surfaces were performed for all E. coli strains. The optical density (OD) of the stained attached bacteria and control wells were read at λ = 570 nm. The cut-off OD (ODc) was de ned as the mean OD of the negative control (TSB only). Based on the ODs of the bacterial bio lms, all strains were classi ed into the following categories: non-adherent (-: OD ≤ ODc), weakly adherent (+: ODc> OD ≤ 2x ODc), moderately adherent (++: 2x ODc> OD ≤ 4x ODc), or strongly adherent (+++:OD>4x ODc). Each assay was performed in triplicate and repeated three times. S. epi dermidis strain ATCC 35984 was used as a positive con troll (Van Belkum et al., 2007).
Bio lm formation on a hydrophilic glass surface.
Mi croorganisms were inoculated in glass tubes (15x100 mm) containing 5 mL of TSB medium and incubated at 37ºC for 48 h.
The supernatants containing non-adherent bac terial cells were discarded. Fresh sterile TSB (5 mL) was added to the test tubes and re-incubated for 48 h. This procedure was repeated twice. Glass-adherent bacteria created a con uent coat of cells on the sides of the tube. Microorganisms were classi ed as non-adherent (-: ab sence of adherence), weakly adherent (+: adherent bacte ria appeared as a ring at the interface between the medium and the air), moderately adherent (++: bacteria attached on the side of the glass tubes), or strongly adherent (+++: bacteria attached on the side of the glass tubes and at the interface between the medium and the air). S. epidermidis strain ATCC 35984 was used as a positive control (Mattos-Guaraldi and Formiga, 1991).

cells per well) in Minimum Essential
Medium Eagle (Sigma-Aldrich) supplemented with 10% bovine fetal serum. Mid-log-phase bacteria were cultured in Trip Soy Broth -TSB for 24h at 37ºC and reach OD 580 nm of 0.2 were then added to each well with MOI of 10 and 100 bacteria per epithelial cell to test the in uence of the amount of inoculated bacteria on the number of internalized bacteria. Internalization assays was allowed to occur for 2 h and 4 h at 37°C in an atmosphere of 94% air-6% CO2.
To determine the level of bacterial adhesion, 96-well plates containing epithelial cells incubated with mid-log-phase bacteria had been prewashed three times with PBS and lysed with 100 ul PBS-triton (Sigma-Aldrich) to enumerate adherent bacteria added. All strains were shown to be susceptible to ≤ 150 mg/mL of gentamicin to the invasion experiments and were incubated for 1 h to determine the bacterial invasion. After incubation period, monolayers were washed three times with saline and lysed with 0.1% Triton X-100 to determine the viable counts of released intracellular bacteria. Invasion ability was expressed as the percentage of inoculum that survived 150 mg of gentamicin per ml treatment. Following the invasion period as described above, assay of rsistence was demonstrated after 24 h. The infected cells were incubated at 37°C in 5% CO2 using 150 mg of gentamicin per ml.
The results were recorded as percentage of the original inoculum. All assays were conducted in triplicate and were repeated independently at least three times (Hirata Jr. et al., 2004;Sahly et al., 2000).

Results
In this study, a total of 34 E. coli strains were identi ed by MALDI-TOF mass spectrometry with ≥ 2 score. Analysis of data showed that all strains presented resistance to at least one group of antimicrobial agents tested, and expressed a resistant (R) pro le.
Interestingly, 20 of 34 (58.8%) E. coli river isolates presented resistance to three or more antimicrobial group of interest and were considered as multi-drug resistance (MDR). Multi-drug resistance of E. coli river isolates included resistance to 3 rd and 4 th cephalosporin, aminoglycosides, uorquinolones and carbapenems ( Table 1).
Analysis of bio lm formation ability of resistant and MDR E. coli strains demonstrated that all strains were able to promoted bio lm formation on polyestirene surface in different levels. From the 20 MDR E. coli strains, 40% (n=8) were considered as strongly adherent on both polyestinere and glass surfaces, followed by 20% (n=7) moderated adherent, 20% (n=4) weakly adherent and 5% (n=1) non-adherent. Resistant E. coli also demonstrated heterogeneity among both polyestinere and glass bio lm formation: 42.8% (n=6) were able to promote bio lm formation and were considered as weakly adherent, followed by, 35.7% (n=5) as moderated adherent and 7.2% (n=1) as strongly adherent, while 14.3% (n=2) of resistant E. coli strains were considered as nonadherent (Table 2).
Further analysis of virulence pathotypes of MDR E. coli strains were displayed in Table 3 and Figure 1. Data showed that 20% (n=4) of MDR E. coli river isolates expressed virulence pathotypes: EHEC (Ec31); STEC and EIEC -hybrid (Ec21) and STEC (Ec30 and Ec10) while 80% (n=16) did not presented any virulence pathotypes analyzed. Evaluation of the presence of virulence genes demonstrated heterogeneity among the isolates. All four strains presented stx2 gene followed by stx1 (n=1), iaL (n=1) and eaeA (n=1). Multiplex-PCR for Lt, St and aegg genes presented negative results in all opportunities (Figure 2).
The results of the quantitative cell-associated of MDR E. coli river isolates (n=4) harboring virulence patothypes were shown in Table 2. The highest level of adherence to Caco-2 cells was observed with STEC (Ec10 and Ec30) strains presenting stx2 gene: 63.1% and 45.9% respectively. EIEC/STEC (Ec21) and EHEC (Ec31) strains expressed lower ability of adherence to human epithelial cell line: 35.4% and 27.8% respectively. Viable internalized bacteria were detected at 1 h post-infection of the monolayers, regardless of the E. coli pathotypes but at different levels. The highest percentages of viable intracellular bacteria deduced from Caco-2 cell-associated bacteria were observed for both STEC (Ec10 and Ec30) strains: 3.5% and 0.1% respectively. EIEC/STEC (Ec21) and EHEC (Ec31) strains presented lower percentages of viable intracellular bacteria: 0.007% and 0.01% respectively. Bacterial persistence following a longer period of incubation (24h) was displayed in Table 2. All four MDR E. coli analyzed strains were able to survive in the presence of Caco-2 cells at different levels: EIEC/STEC (Ec21) 1.8%; EHEC (Ec31) 3.14%, STEC (Ec30) 18.3% and STEC (Ec10) 145.8%. Interestingly, the STEC (Ec10) isolate not only persisted viable but also was capable to multiply within the Caco-2 cell. In this present study, we documented the presence of resistant and MDR E. coli strains isolated from Joana river, located at Rio de Janeiro metropolitan area, Brazil. Heterogeneity among virulence pathotypes as well as host pathogen interaction with Caco-2 cells was veri ed among MDR river isolates. The presence of antimicrobial resistant E. coli in aquatic systems released from anthropogenic sources such as community, industries, veterinary and hospitals, is a public health concern in industrialized and developing countries due to relevance on environmental dissemination of antimicrobial resistance (Djordjevic, Stokes, Chowdhury, 2013;Berendonk, et al., 2015).
Although generally harmless, E. coli strains may express virulence potential properties that account for human localized and invasive infections in both communities and hospital enviroments (Hall-Stoodley, Costerson, Stoodley, 2004). The plasticity of the E. coli genome has hindered the identi cation of certain E. coli isolates as a pathotype, because some isolates combine the main virulence characteristics of different pathotypes and are thus considered potentially more virulent hybrid pathogenic strains. In this study, MDR E. coli river isolates presented the following distinct pathotypes: STEC, EHEC, EIEC/STEC. MDR E. coli Ec31 isolated strain was classi ed as EHEC due to the ability to produces stx1 and stx2 Shiga toxin (Stx) cytotoxins associated with eaeA gene. MDR E. coli strains Ec21, Ec30 and Ec10 presented stx2 Shiga toxin (Stx) cytotoxins. The association of Stx cytotoxins, especially stx2, with severe diseases has been extensively studied by using endothelial cell lines and their ability to adhere is related to EHEC/STEC pathogenesis (Rivas et al., 2016). In addition, bio lm may act as bacterial protection against adverse environmental conditions, especially in aquatic envionment. A study conducted by Biscola and coworkers (2011) evaluated the capacity of bio lm formation in EHEC/STEC strains isolated from different reservoirs and demonstrated strongly ability to adhere on both glass and polystyrene surfaces. In fact, cell invasion and survival of EHEC/STEC strains in cultured human intestinal epithelial cells has been previously described (Cordeiro et al., 2013) and may be related to bio lm strongly adherence. It should be mentioned that this invasive characteristic has been identi ed in EHEC/STEC serotypes, Our studies also demonstrated the presence of MDR enteroinvasive E. coli Ec21 strain. EIEC is a causative agent of dysentery in humans, especially in developing countries, due to their ability to invade and penetrate cells by endocytosis, as shown in table 2.
Despite the similarities invasion mechanism and symptoms of the disease, the infectious dose of EIEC appears to be a milder and self-limiting form when compared to Shigella, who leads to an exacerbation of proin ammatory response. EIEC was responsible for several outbreaks, but there are few reports on routes of transmission and distribution of this bacterium in nature, including water and cheese (Marier et al., 1973;Valentini et al., 1992) as well as the direct transmission through person-to-person contact. The isolation of EIEC in Brazil has ranged from 0.5 to 15%, depending on the population investigated (Moreno et al., 2010;Lozer et al., 2013). Toledo and Trabulsi (1990) investigated the presence of this microorganism in different areas of the city of São Paulo. This bacterium has been found related to children with diarrhea (15.9%). Studies performed outside the city of São Paulo showed a low prevalence of these bacteria, 0.5-2.5% (Oliveira et al., 1989).
Few studies have been conducted to investigate pathogenic E. coli strains in urban rivers, although pollution of surface waters with these pathogens has been implicated in an increased number of disease outbreaks and consequent deaths (Masters et al., 2011).
In attempt to investigate virulence potential of E. coli environmental isolates expressing MDR pro les and virulence genes, were investigated for ability of interaction with Caco-2 human intestinal epithelial cells. All MDR E. coli strains of STEC, EHEC and EIEC/STEC pathotypes were able to adhere to epithelial cells surfaces. MDR E. coli (Ec 10 and Ec30) isolates, classi ed as STEC pathotypes and presenting stx2 gene, expressed the higher ability of adherence, internalization and persistence within Caco-2 epithelial cells. Previous report documented that STEC annually was responsible to 2,801,000 cases of acute illness, 3890 HUS Moreover, one MDR E. coli (Ec31) river isolate was characterized as EHEC, presenting lower levels of adherence, internalization and persistence within Caco-2 epithelial cells. Virulence potential of this pathotypes is partially demonstrated by the ability of attaching intimately and effacing microvilli of epithelial intestinal cells that can directly induce renal and endothelial lesions due to expression of eaeA, Stx1 and Stx2 genes (Donnenberg, 1993;Maule, 2000;Gomes et al., 2016). Survival and persistence of EHEC in contact with surfaces and exposure to water environments among other conditions should be recognized as important risk factors in the spread of this pathogen, including rivers located at metropolitan areas. Data that deserves attention In this study, one MDR E. coli (Ec31) river isolate was characterized as EHEC, presenting lower levels of adherence, internalization and persistence within Caco-2 epithelial cells, possibly related to previously described cytotoxicity abilities -whether apoptosis and/or necrosis (Donnenberg, 1993;Maule, 2000;Gomes et al., 2016;Abul-Milh et al., 2001). Data reinforce the fact that survival and persistence of EHEC in contact with surfaces and exposure to water environments among other conditions should be recognized as important risk factors in the spread of this pathogen, including rivers located at metropolitan areas. Information that deserves attention concerning the virulence potential and risk of contamination by EHEC pathotypes is the ability of causing acute infections with only ten bacterial cells indicating a high virulence level (Maule, 2000).

Conclusion
In conclusion, this study demonstrates the presence of diarrheagenic E. coli strains in river water source at Rio de Janeiro metropolitan area, Brazil. However, a subset of these strains demonstrated a high pathogenic potential as they exhibited a multidrug resistant phenotype and virulence genes. The possibility of contamination leading to human infection and cause gastrointestinal disease due to MDR E. coli presenting virulence pathotypes water-borne pathogens is a matter of concern. The presence of diarrheagenic E. coli in river waters warrants the implementation of environmental safety strategies in order to avoid the dissemination of clones to people leaving in the area but particularly those more vulnerable communities who utilize these waters for domestic purposes, including Rio de Janeiro.

Con ict of interest
There are no con icts of interest among the author E. coli strains Antimicrobial resistance pro les  Legend: SA/ +++: strongly adherent, MA/++: moderated adherent; +: adherent bacte ria appeared as a ring at the interface between the medium and the air. Fi gu re s Figure 1 Collection site of water samples from Joana River located in front of a University Hospital at Rio de Janeiro metropolitan area, Southeast Brazil