Introduction

Bacteria are ubiquitous, inhabiting a wide range of environments, from commensals and symbionts to potential colonizers or pathogens. They exhibit a natural ability to resist chemical compounds and other microorganisms as part of their survival1. However, human intervention has exacerbated this resistance2.

Antimicrobial Resistance (AMR) is a phenomenon in which microorganisms become resistant to antimicrobial drugs that were previously effective against them3. AMR is recognized as one of the major threats to public health, with significant economic consequences and the potential to push millions of people into poverty4.

The excessive use of antimicrobials in various sectors, including medicine, agriculture, and food production, is one of the primary drivers of AMR5. Furthermore, improper waste management can result in the presence of antimicrobials in the environment, further increasing resistance. Some bacteria, such as Acinetobacter baumannii, Pseudomonas aeruginosa, and species of the Enterobacterales order, have become resistant to a wide range of antimicrobials. This includes resistance to carbapenems and, in the case of Enterobacterales, third-generation cephalosporins. Due to this resistance, A. baumannii and Enterobacterales have been classified as critical priority pathogens by the World Health Organization (WHO), while carbapenem-resistant Pseudomonas aeruginosa is considered a high-priority pathogen6.

Aquatic environments play a fundamental role in the dissemination of resistance, acting as reservoirs of genes where microorganisms can exchange these genes through horizontal transfer7. In urban areas, aquatic environments may present conditions that contribute to the maintenance of resistance, such as hospital and domestic sewage8. In both examples, they may contain resistant bacteria or agents that act as inducers of antimicrobial resistance, such as disinfectants and psychotropic compounds, in addition to antimicrobial drugs, especially after the COVID-19 pandemic9,10,11,12. Furthermore, the accumulation of microplastics in the environment can play a role not only as a means of transportation for these chemical pollutants but also as a vector in the spread of bacteria resistant to multiple drugs and their resistance genes13.

In a context where (AMR) is considered a serious global health problem14, this study investigates the presence of resistant Gram-negative bacteria and their resistance genes in water samples from the Carioca River (CR) and Rodrigo de Freitas Lagoon (RFL), two important anthropogenically impacted environments in Rio de Janeiro, comparing them with previously published data on carbapenem resistance in these environments from almost a decade ago15.

Materials and methods

Study sites

Surface water samples were obtained from the CR and RFL (Fig. 1), specifically selecting locations that were susceptible to anthropogenic impacts, including areas affected by rainwater and sewage runoff: CR 1 - upstream of the CR treatment plant (22° 56’ 0.413” S 43° 10’ 20.446” W), CR 2 - downstream of the treatment plant (22° 56’ 4.495” S 43° 10’ 19.553” W), CR 3 - the mouth of the CR at Flamengo Beach (22° 56’ 8.933” S 43° 10’ 11.615” W), RFL 1 - a recreational area on the shores of the lagoon (22° 58’ 27.278” S 43° 12’ 9.212” W), RFL 2 – near Caiçaras Island (22° 58’ 48.7” S 43° 12’ 43.5” W), RFL 3 - near the Jardim de Alah Canal (22° 58’ 48.324” S 43° 12’ 49.622” W), RFL 4 - outflow of the General Garzon Street canal into RFL, which has a watercourse extension from the Macacos River (22° 58’ 3.810” S 43°13’ 2.082” W), RFL 5 - near a hospital (22° 57’ 51.381” S 43° 12’ 51.492” W), and RFL 6 - before the Jardim de Alah Canal (22° 58’ 49.1” S 43° 12’ 34.8” W)15.

Sample collection and processing

Sampling took place in the morning on November 10, 2022. Each sample consisted of 500 mL of water collected in sterile glass containers. Subsequently, the samples were promptly chilled on ice and transported to the laboratory, following the guidelines established by the Standard Methods for the Examination of Water and Wastewater and ISO 19458:200616[,17.

Physicochemical analysis of water

During collection, a HI9829 model multiparameter meter from HANNA Instruments was used to conduct physicochemical analyses of the water at the sampling points to assess water quality.

Isolation of bacterial samples

To process the collected water samples, 200 µL were directly added to the CHROMagar™ KPC (Plastlabor) culture medium. The samples were further subjected to dilutions in saline solution (NaCl) at concentrations from 10⁻¹ to 10⁻³. Additionally, 1 mL of each sample was placed in a 50 mL sterile Falcon tube containing MacConkey broth and incubated at approximately 35 °C for 24 h. From this growth, 200 µL were spread on CHROMagar™ KPC (Plastlabor). All plates were incubated at approximately 35 °C for 24 h to allow bacterial growth, followed by counting the colony-forming units (CFU/mL). Different colonies of CFUs were subcultured on Nutrient Agar (NA) plates to obtain pure cultures. Subsequently, the isolates were stored in Brain Heart Infusion Broth (BHI - Merck) supplemented with 20% glycerol (Sigma) at −20 °C.

Fig. 1
figure 1

Location of Water Collection Points in the Carioca River and Rodrigo de Freitas Lagoon, Rio de Janeiro, Brazil. Data source: Brazilian Institute of Geography and Statistics - IBGE and Pereira Passos Institute - IPP. Datum: SIRGAS 2000.

Full size image

Identification of bacterial isolates

The isolates collected from CR and RFL sites were identified using Matrix-Assisted Laser Desorption/Ionization Time of Flight Mass Spectrometry (MALDI-TOF-MS), employing the MALDI-TOF MicroflexLT equipment from Bruker Daltonics (Bremen, Germany). The VITEK 2 system (BioMérieux, Nürtingen, Germany) was used to confirm the identification when MALDI-TOF-MS was unable to identify the bacterial isolate.

Antimicrobial susceptibility tests (AST)

Antimicrobial susceptibility tests were performed using the agar diffusion method, following Clinical and Laboratory Standards Institute (CLSI) guidelines18,19. The following antimicrobials were used, according to CLSI instructions for each bacterial group: Amoxicillin-clavulanate - AMC (20/10 µg); Ampicillin-sulbactam - ASB (10/10 µg); Piperacillin-tazobactam - PPT (100/10 µg); Cefepime - CPM (30 µg); Cefotaxime - CTX (30 µg); Ceftriaxone - CRO (30 µg); Cefoxitin - CFO (30 µg); Ceftazidime - CAZ (30 µg); Aztreonam - ATM (30 µg); Ertapenem - ETP (10 µg); Imipenem - IMP (10 µg); Meropenem - MER (10 µg); Gentamicin - GEN (10 µg); Tobramycin - TOB (10 µg); Amikacin - AMI (30 µg); Tetracycline - TET (30 µg); Minocycline - MIN (30 µg); Ciprofloxacin - CIP (5 µg); Levofloxacin - LEV (5 µg); Norfloxacin - NOR (10 µg); and Trimethoprim-sulfamethoxazole - SUT (1.25/23.75 µg).

When it was not possible to perform the comparison of breakpoints according to CLSI guidelines using the disk diffusion method on agar, the Minimum Inhibitory Concentration (MIC) assay was conducted using reagent strips - Etest®. The following antimicrobials were tested: Ceftriaxone, Imipenem, Meropenem, Gentamicin, and Levofloxacin.

Pseudomonas aeruginosa American Type Culture Collection (ATCC) 27,853 and Escherichia coli ATCC 25,922 and 35,218 were used as controls in the test.

Phenotypic detection of polymyxin resistance

Among the isolates that demonstrated resistance or intermediate results to at least one of the tested carbapenems, the drop test was performed as a screening method20. Isolates that exhibited resistance were subjected to the broth microdilution test to determine the minimum inhibitory concentration of polymyxin B21. An isolate of E. coli ATCC 25,922 was utilized as the negative control, while an isolate of Klebsiella pneumoniae, recognized for its resistance to polymyxin, served as the positive control.

Detection of resistance determinants and genetic context

Conventional Polymerase Chain Reaction (PCR) was carried out to detect genes encoding β-lactamases (blaSPM, blaSHV, blaTEM, blaCTX−M, blaKPC, blaNDM, blaOXA−48−like, blaGES, blaIMP, and blaVIM)22,23,24,25, mobile colistin resistance (mcr-1 to mcr-526 and mcr-6 to mcr-927), and intl128. PCR was performed on isolates that showed resistance or intermediate results to at least one type of tested carbapenem. In the case of Acinetobacter species, the presence of other oxacillinase-encoding genes (blaOXA−23, blaOXA−24, blaOXA−51, blaOXA−58, blaOXA−143, and blaOXA−235) was also investigated29. The controls used in the test are listed in Table S1.

The investigation of the genetic context of blaKPC and blaNDM genes was conducted through PCR in isolates identified as potential producers of the enzymes encoded by these genes. Specific primers were used for the Tn4401 region, which encompasses the ISKpn6, ISKpn7, TnpA, and IR regions30,31, to determine the possible presence of the transposon Tn4401 associated with the blaKPC−2 gene. Primers directed at the IS3000 region were employed to seek the region adjacent to the blaNDM gene32.

Analysis of multiple antimicrobial resistance phenotype (MARP) and multiple antimicrobial resistance index (MARI) of bacterial isolates

The isolates that demonstrated resistance to three or more antimicrobials were subjected to MARP pattern analysis. The multiple antimicrobial resistance index was calculated using the mathematical equation proposed by Krumperman33,34,35.

$$MARindex = a/b$$

Where ‘a’ denotes the number of antimicrobials to which the isolate exhibited resistance, while ‘b’ represents the total number of antimicrobials to which the isolate was exposed. An MAR index greater than 0.20 is indicative of intensive antimicrobial usage, characterizing a “high-risk” contamination source33,34,35,36.

Statistical data analysis

Initially, variables were standardized on a scale from 0 to 1.0 to facilitate statistical analysis. Each variable was assigned a score of 1.0 for the highest observed value, with other values scaled proportionately. Spreadsheets and figures were then generated using Past 4.0 and Microsoft Excel.

“Antimicrobial Non-susceptibility” was calculated by dividing non-susceptible isolates for each antimicrobial by the total bacterial isolates recovered at each site, converting the result to a 0–1 scale.

Non-susceptible Bacteria Density was computed by summing non-susceptible isolates across Enterobacterales, Acinetobacter spp. and Pseudomonas spp. for each antimicrobial, then dividing by the total isolates in each taxonomic group at the site.

Principal component analysis considered Parts Per Million of Dissolved Oxygen (ppmDO), turbidity, and Non-Susceptible Bacteria Density for clustering sampling sites. Clustering used the UPGMA (Unweighted Pair Group Method with Arithmetic Mean) method, employing Rho similarity.

Resistance gene occurrence via conventional PCR was assessed by dividing positive gene counts by total isolates at each site, compared to Araújo et al.15 data.

Results were organized into a matrix to detect patterns among clustering groups, based on 0–1 scaled variables.

Results

Identification and distribution of bacterial isolates in water samples

A total of 101 isolates of Gram-negative bacilli (GNB) were recovered from water samples: 45 from the CR and 56 from RFL. Among these isolates, 65% were non-fermenting GNB, with the vast majority (78.8%, 52/66) identified as Acinetobacter species, followed by 18.2% (12/66) identified as Pseudomonas species, along with one isolate of Burkholderia cepacia complex and one of Stenotrophomonas maltophilia. The remaining 30% of isolates were from the Enterobacterales order, with the overwhelming majority (83.9%, 26/31) being Enterobacter species. In smaller proportions, 9.7% (3/31) were E. coli, 3.2% (1/31) K. pneumoniae and 3.2% (1/31) Kluyvera georgiana. Additionally, among the 101 isolates, two were identified as Aeromonas caviae, one as Delftia acidovorans and one as Comamonas testosteroni (Table S2).

Evaluation of antimicrobial susceptibility profile of target groups to a panel of tested antimicrobials and phenotypic antimicrobial resistant pattern of individual bacterial isolates

The results from the antimicrobial susceptibility tests showed 45.5% non-susceptibility and 54.5% susceptibility to at least one of the tested carbapenems. For D. acidovorans and C. testosteroni isolated from point 1 of the RFL, the MIC was determined due to the absence of defined breakpoints by the disk-diffusion method. These isolates demonstrated resistance to gentamicin, with MIC values of > 1024 mg/L and 32 mg/L, respectively.

Among the 46 isolates non-susceptible to carbapenems, 19% (9/46) demonstrated resistance to polymyxin (Table 1), with point CR2 having the majority of these isolates.

Table 1 Bacterial isolates resistant or intermediate to carbapenems and polymyxins recovered from water sample collection points in the carioca river and Rodrigo de Freitas Lagoon, Rio de Janeiro, on November 10, 2022.
Full size table

Assessment of MARP and MARI for all identified groups

Among the isolates in the present study, 36% (37/101) exhibited a MAR index above 0.20 (Tables 2 and 3, and 4). Of these, 46% (17/37) were species of the order Enterobacterales, 43.2% (16/37) were Acinetobacter species, 5.4% (2/37) were Pseudomonas guariconensis and 5.4% (2/37) were A. caviae. The isolates with the highest MAR index were two E. coli (0.77), one K. georgiana (0.66), one A. caviae (0.66) and one A. johnsonii (0.64).

Table 2 MAR phenotypes and MAR index patterns in enterobacterales isolated from the carioca river and Rodrigo de Freitas lagoon in the City of Rio de Janeiro, on November 10, 2022.
Full size table
Table 3 MAR phenotypes and MAR index patterns in Acinetobacter species isolated from the carioca river and Rodrigo de Freitas lagoon in Rio de Janeiro, on November 10, 2022.
Full size table
Table 4 MAR phenotypes and MAR index patterns in Pseudomonas species and Aeromonas caviae isolated from the carioca river and Rodrigo de Freitas lagoon in Rio de Janeiro, on November 10, 2022.
Full size table

PCR of genetic determinants of resistance

Among the 46 carbapenem non-susceptible isolates, 41% (19/46) exhibited blaKPC gene amplification by PCR, 26% (12/46) exhibited the blaGES gene, 6% (3/46) exhibited the blaNDM gene, 6% (3/46) exhibited the blaCTX−M gene, 2% (1/46) exhibited the blaVIM gene, and 32% (15/46) exhibited intl1 amplification (Table 5). Among the 19 carbapenem non-susceptible isolates that exhibited the blaKPC gene, the following observations were made: 4 isolates showed amplification of the surrounding insertion sequences ISKpn6 by PCR; 1 isolate showed amplification of ISKpn6, ISKpn7 and tnpA; 1 isolate showed amplification of ISKpn7 and tnpA; 1 isolate showed amplification of the left IR; and 12 isolates did not amplify regions of Tn4401. Conversely, all isolates potentially carrying the blaNDM gene amplified the IS3000 region (Tn3000).

Table 5 Bacterial isolates and β-Lactamase genes recovered from water sample collection points in the carioca river and Rodrigo de Freitas Lagoon, Rio de Janeiro, on November 10, 2022.
Full size table

Contextualization of phenotypic and genotypic analysis of antimicrobial resistance, physicochemical, and biochemical properties of water at collection points

Antimicrobial-resistant isolates carrying resistance-associated genes were identified at nearly all sampling points, except at point RFL 6 (Fig. 2).

Fig. 2
figure 2

(a) Proportional profile of isolated non-susceptible—resistant and intermediate—to common antimicrobials tested per sampling point. (b) Proportional profile of antimicrobial resistance gene occurrence per sampling point. Legend: CR – Carioca River; RFL – Rodrigo de Freitas Lagoon.

Full size image

Non-susceptibility per site collection

Figure 3a illustrates the percentage of isolates that are either non-susceptible and susceptible to carbapenems at each sampling site. The statistical analysis of biological and physicochemical variables revealed that sampling points CR1, CR2, RFL3, and RFL4 are grouped by similarity (Fig. 3b). This group exhibits the highest percentage of non-susceptibility to the tested antimicrobials (Fig. 3c).

When comparing the results of the variables in a matrix, we observed that Group 1 (RFL4, RFL3, CR2 and CR1) showed a lower proportional score of isolates belonging to the genus Acinetobacter that were non-susceptible to antimicrobials and a higher number of isolates from the order Enterobacterales compared to Group 2 (RFL1, RFL2, RFL5 and CR3). However, among the common antimicrobials tested in the different taxonomic groups, both Group 1 and Group 2 showed high resistance to cephalosporins: cefotaxime, ceftazidime and ceftriaxone. Additionally, Group 2 exhibited a high level of resistance to the monobactam aztreonam, while Group 1 showed high resistance to the carbapenem imipenem (Fig. 4).

Fig. 3
figure 3

(a) Percentage of occurrence of isolates non-susceptible and susceptible to tested carbapenems per sampling site. (b) Clustering of sampling sites based on antimicrobial resistance per taxonomic group/per site, dissolved oxygen concentration (ppmDO), and turbidity - Cophenetic Coefficient of 0.85. (c) Percentage of all isolates non-susceptible and susceptible to carbapenems within the clustering of sampling site groups. Legend: CR – Carioca River; RFL – Rodrigo de Freitas Lagoon.

Full size image
Fig. 4
figure 4

Matrix comparing variable results (rows) and sampling points (columns) aligned by the clustering outcome: Group 1 (Carioca River 1 and 2, Rodrigo de Freitas Lagoon 3 and 4) on the left and Group 2 (Carioca River 3 and Rodrigo de Freitas Lagoon 1, 2, and 5) on the right.

Full size image

Legend: Achieve this, we translated the variable values, observed on a scale from 0 to 1, into a color scale where values near 1 are displayed in shades of red, while those near zero appear in shades of blue. The four columns on the left correspond to group 1, whereas the four on the right represent group 2. The variables set up were: ppmDO – Parts Per Million of Dissolved Oxygen; FNU (Turbidity - Formazin Nephelometric Units); CFU – Colony Forming Units; AMI – Amikacin; ASB – Ampicillin-Sulbactam; ATM – Aztreonam; CPM – Cefepime; CTX – Cefotaxime; CAZ – Ceftazidime; CRO – Ceftriaxone; CIP – Ciprofloxacin; GEN – Gentamicin; IMP – Imipenem; MER – Meropenem; PPT – Piperacillin-Tazobactam; SUT – Sulfamethoxazole-Trimethoprim; TET – Tetracycline; TOB – Tobramycin; Taxa – taxon number; CR – Carioca River; RFL – Rodrigo de Freitas Lagoon.

Comparison between the present study with Araújo et al., 201615

In comparison to the data from the 2016 publication15, which collected samples at the same sites, the present study observed the continued presence of the blaKPC gene at the CR collection points (Fig. 5).

Fig. 5
figure 5

Percentage occurrence of the blaKPC gene along the Carioca River (a) and Rodrigo de Freitas Lagoon complex (b) - Rio de Janeiro, Brazil, in the current study (collection in November 2022) compared to a study that collected samples in July 201315,

Full size image

Legend: The percentage of blaKPC occurrence per sampled site was plotted in Fig. 5. We included Araújo et al.. (2016) data on blaKPC occurrence when they sampled the same site in the percentage graph15. CR – Carioca River; RFL – Rodrigo de Freitas Lagoon.

Discussion

The mechanisms and epidemiology of AMR remain a significant challenge for human, animal, and environmental health. The aim of this study was to update an AMR inventory established nearly a decade ago to better understand how bacterial resistance persists in the environment.

The study focused on a river that traverses various neighborhoods in Rio de Janeiro, receiving effluents and waste from domestic and hospital sources, as well as rainwater. This river drains sewage from different communities before flowing into a popular beach in the city15,37. Additionally, we investigated a lagoon that collects water from multiple channels passing through various neighborhoods and is also used for sports, leisure, and tourism activities.

Collection occurred on a sunny day during low tide, although it had rained the day before. Recent studies suggest that rainfall can increase the presence of antimicrobial resistance genes (ARGs) in water38. Low tide and reduced inflow of saltwater may lead to increased isolation of microorganisms due to reduced dilution, which might enhance contact with sediments. These sediments can serve as reservoirs for ARGs, and rainfall can disturb these sediments39. Furthermore, low tide has been associated with increased concentrations of carbapenemase producers, with polluted water influx from Guanabara Bay affecting the concentration of these producers37.

At sites CR1 and CR2, we observed a higher number of different species, particularly from Enterobacterales (Fig. 4), and greater diversity of antimicrobial resistance genes (Fig. 2b). This may indicate contamination from antimicrobials, hospital waste, and human excreta.

Our results also revealed the presence of bacterial isolates categorized as critically prioritized pathogens by the WHO6, as well as members of the ESKAPE group (Enterococcus faecium, Staphylococcus aureus, K. pneumoniae, A. baumannii, P. aeruginosa, and Enterobacter spp.), which are hospital pathogens known for their antimicrobial resistance and virulence40,41. Notably, we found a large number of non-fermentative Gram-negative bacteria isolates, such as Acinetobacter species. Research suggests that the adaptive capacity of Acinetobacter species extends beyond their high metabolic potential encoded in their chromosomes to continuous gene transfer through their plasmids. Many of these plasmids are mobilizable, and mega-plasmids can not only move themselves but also mobilize other plasmids, potentially facilitating resistance gene transfer between habitats42,43.

The horizontal transfer of ARGs is often mediated by plasmids from environmental, animal, or human sources. Such plasmid transfer can occur at environmental interfaces, such as in sewage systems or wastewater treatment plants (WWTPs), where environmental and human bacteria converge43. Thus, the points along the CR are conducive to this occurrence. Additionally, aquatic environments may contain sub-inhibitory concentrations of antimicrobials, which can induce bacterial tolerance, persistence, dependency, or survival44. Sub-inhibitory concentrations can also promote mutations and gene transfer, contributing to increased antimicrobial resistance45.

Analysis of distribution at collection points showed that Group 1 had the highest percentage of antimicrobial resistance (Fig. 3), consistent with the assessment of MARP and MARI, identifying that points from Group 1 (CR1, CR2 and RFL4) had the highest number of isolates with a MAR index above 0.20 (Tables 2 and 3, and 4). This suggests a potential relationship between observed resistance and anthropogenic impacts. CR1 and CR2 receive residential and hospital effluents, including from a disused wastewater treatment plant, while RFL4 is near a canal connecting the lagoon to the beach. Point RFL4 also receives effluents from the Macacos River and a horse racing club, which may introduce residual flora and pathogenic load. Overflow during rainfall can lead to contamination with animal excreta and antimicrobials46.

Point CR1 was notable for presenting E. coli isolates resistant to carbapenems with the highest recorded MAR index (0.77) and the presence of blaNDM and intl1 (Table 5). Additionally, another carbapenem-resistant isolate with one of the highest MAR index, K. georgiana (0.66), was found at this point. Besides possessing the blaCTX−M gene, a likely precursor47, this bacterium also demonstrated the presence of blaKPC, having amplified the ISKpn6 and ISKpn7 regions of Tn4401, in addition to intl1 (Table 5).

Although CR3 and RFL1 did not show the highest percentages of antimicrobial resistance or MAR index above 0.20, isolates with high MAR index were found. For example, A. caviae at CR3, resistant to carbapenems and polymyxin (Table 1), had a MAR index of 0.66 and amplified blaKPC and blaGES (Table 5). Similarly, RFL1 had a carbapenem-resistant A. johnsonii with a MAR index of 0.64, carrying blaKPC and amplified ISKpn6 and blaGES. These findings indicate that both group 1 and group 2 points could pose ‘high-risk’ environmental contamination33.

Several hypotheses explain the higher number of antimicrobial-resistant bacteria at CR1, CR2, and RFL4 compared to other sites. One possibility is that high concentrations of sewage bacteria mix with contaminants and native bacteria along the river, reducing their proportion in samples. Additionally, RFL4 receives effluents from the Macacos River and a horse racing club, which could recolonize the lagoon with high pathogenic load. Some authors question whether pathogenic bacteria persist similarly to autochthonous bacteria adapted to environmental stresses48,49. In this study, sewage bacteria may suffer synergistic effects of salinity and solar radiation, reducing their cultivability.

Similarities between collection points showed high resistance to cephalosporins (Fig. 5). These β-lactam antimicrobials have been widely used in human medicine since the 19th century and are currently utilized as second- or third-line treatments50. These antimicrobials are also employed in veterinary medicine50. These compounds can be excreted in feces and urine by humans and animals51,52. Studies report persistence of cephalosporin-resistant isolates in manure treatment products53. Enterobacteriaceae resistant to carbapenems and all tested cephalosporins have been found in fish, water, and aquaculture workers54.

Resistance to imipenem was also higher (Fig. 5), which has a high potential for selecting antimicrobial-resistant bacteria, even at low concentrations55.

Comparing this study with Araújo et al.. (2016), conducted nearly a decade ago in the same environments15, it is evident that previous samples from RFL did not detect blaKPC. The earlier study suggested potential clonal dissemination of KPC-carrying bacteria at CR1 but not at CR2. The KPC enzyme is widespread and considered endemic in Brazil, with aquatic environments possibly continuing to act as genetic reservoirs and dissemination routes for these bacteria56. The persistence of genes encoding β-lactamase enzymes, such as blaKPC, blaNDM, and blaGES, complicates the treatment of bacterial infections, especially those that are multidrug-resistant14.

Currently, ceftazidime-avibactam is available as a therapeutic option but does not cover metallo-β-lactamases, often requiring combination therapy with aztreonam57. Cefiderocol is a new antimicrobial effective against MDR Gram-negative bacteria producing metallo-β-lactamases but is not yet available in Brazil. Reports indicate NDM-producing isolates resistant to cefiderocol and heteroresistant subpopulations of A. baumannii58,59. Polymyxins, such as polymyxin B and colistin, are last-resort treatments for multi-resistant Gram-negative isolates; however, resistance is increasing, limiting their efficacy60. In this study, nine isolates were resistant to carbapenems and polymyxins (Table 1). While mobile colistin resistance genes were not detected, other mechanisms, such as lipopolysaccharide alterations, efflux pumps, and capsule formation, may be involved61.

The intl1 gene, a marker for environmental pollution, was present in isolates from CR and RFL. The intl1 gene encodes an integrase of a subclass of integrons widely distributed in Gram-negative bacteria and contributes to the emergence of multidrug-resistant clones62. The qacEΔ1 gene, an efflux pump located in the 3’ conserved sequence of class 1 integrons, might explain disinfectant resistance, particularly given the increased use of disinfectants during the COVID-19 pandemic63,64.

Possible horizontal transfer of resistance genes was suggested by the presence of intl1 in isolates with antimicrobial resistance genes. Isolates of Enterobacter species containing only intl1 were found at CR2 and RFL4 (Table 5), potentially making them susceptible to acquiring resistance genes.

In all E. coli isolates with the blaNDM gene, the IS3000 region of Tn3000 was successfully amplified, highlighting its role in gene dissemination. For blaKPC-positive isolates, only one showed amplification of all Tn4401 transposon parts. Research suggests that in Brazil, blaKPC−2 may be present in NTEKPC elements carried by IncQ1-type plasmids65.

Environmental conditions differed between the collections conducted in 2013 and 2022. The 2013 sampling took place in July, which is typically a cold and dry period, while the 2022 sampling occurred in November, a time characterized by hot and rainy weather. Furthermore, the environmental conditions during these years were likely affected by the COVID-19 pandemic and climate change. Even short periods of selection can enhance plasmid persistence, and the bacterial community context is crucial for plasmid dissemination and resistance gene maintenance43. The persistence of resistance genes like blaKPC, blaNDM, and blaGES indicates ongoing environmental dissemination15. While the direct impact on human or animal health is not yet proven, continued horizontal gene transmission remains a concern.

As of the time this study was written, the Carioca River no longer flows into Flamengo Beach; instead, both the Carioca River and Rodrigo de Freitas Lagoon have been redirected to a submarine outfall. However, overflow events from these effluents may pose significant risks to the population, as they contain multidrug-resistant and opportunistic bacteria that carry β-lactamase-encoding genes. It is important to note that these effluents still reach the treatment plant, but due to insufficient treatment, they are discharged into an outfall located near Ilha de Palmas, which is part of the Cagarras Natural Monument. Instances of probable contamination from this outfall have already been reported in the surrounding environment66.

Conclusion

In conclusion, the data from this study demonstrate the presence of the blaKPC gene in both previously identified and new locations, along with other resistance genes such as blaGES. This finding, combined with environmental changes and potential human exposure, highlights the need for ongoing monitoring and mitigation of antimicrobial resistance in these ecosystems. The detection of resistant microorganisms in environmental samples raises significant concerns for human and animal health. It illustrates the expansion of antimicrobial resistance beyond clinical settings and emphasizes the interconnectedness of health in the One Health approach to addressing environmental dissemination threats. These findings underscore the necessity of a holistic and interdisciplinary strategy to confront the growing challenges posed by antimicrobial resistance. Continuous surveillance of both natural and human-impacted environments is crucial for understanding the extent of the issue and providing insights to guide prevention and control strategies. Additionally, proper treatment of effluents is essential due to the potential negative impact of submarine discharges on the surrounding islands and marine life. Preventing risks to human and animals health relies on the early identification of sources and mechanisms of antimicrobial resistance dissemination.