Antibiotic resistance profiles of sentinel bacteria isolated from aquaculture in Cambodia

Since the discovery of penicillin by A. Fleming in 1929, antibiotics from various classes have been manufactured and used globally to combat diseases caused by harmful bacteria in humans, animals, and plants. Consequently, significant amounts of antibiotics have been released into the environment through the release of wastes from households, hospitals, pharmaceutical companies, wastewater treatment plants, as well as aquaculture and livestock farms (Anh et al. 2021).

The consumption of fish and seafood is projected to rise by 27% by 2030, largely driven by the growth of the aquaculture industry, which is anticipated to expand by 62% in the same timeframe. Aquaculture is a crucial source of income for numerous households, supporting over 100 million individuals worldwide. Consequently, aquaculture plays a vital role in enhancing food security and reducing poverty (Reverter et al. 2020).

With the growing population in Cambodia, aquaculture is becoming economically important for ensuring food security. Fish is the most important source of animal protein for Cambodians, providing around 75% of the total animal protein intake for the population (Lang 2015). There are various and diverse production systems in the country, from floating cage culture, earthen pond culture, rice-fish culture, and other fish culture activities in small water bodies or aquaculture-based fisheries (FAO 2019). Aquaculture of Pangasianodon hypophthalmus or Trey pra was introduced to Cambodia and has since become a key species in the country's aquaculture due to its rapid growth, year-round production, high productivity, and high tolerance to unfavorable environmental conditions such as low dissolved oxygen, pH, and turbidity fluctuations (FiA 2019). On the other hand, giant snakeheads or Channa micropeltes or Trey chhdor in Khmer caught from the wild or imported from Vietnamese hatcheries are farmed in cages and ponds to convert low-value catches into high-value ones (Poulsen et al. 2008). The health of C. micropeltes is also a major concern, due to disease outbreaks caused by bacterial infections, especially in large-scale production, which can result in serious economic losses (Mohamad et al. 2020).

Antibiotics, antifungals, and other pharmaceutical drugs are administered to fish for disease prevention and treatment in aquaculture (Reverter et al. 2020) contributing to the global spread of antimicrobial resistance (AMR) (Done et al. 2015). Contamination from human and animal wastes in water effluents and, therefore, in the aquaculture environment also contributes to the increased risk (Hossain et al. 2022). This raises concerns about food safety and human health associated with aquaculture products. Irrational antimicrobial use, together with a lack of awareness, weak infection prevention and control, unregulated access, self-medication, inadequate training, and low-quality or counterfeit drugs, are some problems experienced in Cambodia (Om et al. 2016, 2017). There is no study on Knowledge Attitude and Practices (KAP) concerning the use of antibiotics for Cambodian aquaculture, but antibiotics are widely used in neighboring countries, particularly in Vietnam (Dang et al. 2021). A recent review found that 67 antibiotic compounds were utilized in 11 out of 15 countries between 2008 and 2018, with oxytetracycline (OT), sulfadiazine, and florfenicol (FFC) being used by 73% of these countries. On average, countries used 15 antibiotics, with top users including Vietnam, China, and Bangladesh (Lulijwa et al. 2020). Moreover, in Bangladesh, OT, ciprofloxacin (CIP), enrofloxacin, erythromycin (E), sulfadiazine, and trimethoprim (W) are being extensively used by fish farmers (Kawsar et al. 2022).

Several studies have shown that antibiotics accumulate in the culture environment, sediments, and buildup in farmed animal tissues, with consequences impacting human and environmental health (Lulijwa et al. 2020). For instance, the antibiotics accumulated in sediment could drive change in microbial communities through selection for antibiotic-resistant species to promote active medical antibiotics and antibiotic-resistant pathogens and genes due to increased selective pressure, rendering the drug increasingly useless (Lulijwa et al. 2020).

Emerging antibiotic-resistant bacteria (ARB) in aquaculture have been reported such as antibiotic-resistant Escherichia coli, Acinetobacter spp., Aeromonas spp., Streptococci and Enterococci, Salmonella spp., Edwardsiella spp, and Streptococcus spp. (Lulijwa et al. 2020). Due to the inherent connections between aquacultural systems with open water bodies, such as rivers and lakes, ARB have been reported to have side effects in open water systems (Lulijwa et al. 2020). At worst, dangerous pathogens eventually acquire resistance to all previously effective antibiotics, resulting in uncontrolled epidemics and epizootics that can no longer be treated (Lulijwa et al. 2020).

Knowledge of antimicrobial use in Cambodia is still limited and there are few published AMR datasets in this country. There is a clear research gap, and data on antibiotic-resistant profiles of bacteria isolated from aquaculture farms are lacking in Cambodia.

Sentinel bacteria are a group of bacteria that are monitored for their potential to cause infections and their resistance to antibiotics. Considering that Aeromonas spp. are useful bacterial indicators of water quality in aquaculture and a potential indicator of antimicrobial susceptibility for the aquatic environment (Naviner et al. 2011), we attempt to investigate these organisms in fish and their culture environment. The incidence of AMR in fish pathogens and the aquaculture environment needs to be monitored at regular intervals to undertake effective measures for the timely prevention of bacterial diseases. In this context, the study aimed to evaluate the level of AMR of Aeromonas spp. and other bacteria associated with aquaculture ponds from giant snakehead and Pangasius fish farms.

This study aimed at providing insights into the status of AMR in Cambodia's aquaculture, contributing to a dataset for controlling and reducing AMR emergence. This study is also a baseline study for enlightening future research.

Three types of samples, including fish, water, and sediment, were collected in this study. Two fish species, striped catfish and giant snakehead, were collected from the ponds in monoculture and polyculture systems. Water samples were collected from the same cultures (ponds) as the fish samples. One water sample from each culture system was analyzed. A list of fish farmers in different locations was provided by respective fisheries officers. The first sampling was conducted on 14 June 2021 in KT, and the second sampling was conducted on 26 August 2021 in KS.

Fish were randomly collected at the sampling locations and placed into a sterile bag and killed by immersion in fusing ice, then placed in expanded boxes, and transported to the laboratory under refrigeration. They were processed within 4–6 h after collection (Scarano et al. 2018).

From approximately 15 cm below the water surface, 250 mL water samples were collected and poured into a sterile plastic bottle. The water samples were stored in an ice box and immediately transported to the laboratory (4–6 h).

From the same locations, approximately 100 g of sediment was collected from the pond bottom using a sediment grab (Cai et al. 2019). The collected sediment was placed in a sterile falcon tube and then put into an ice box during transportation to the laboratory (4–6 h). Samples were immediately processed after arrival at the laboratory.

Skin, gills, and intestinal content samples were aseptically collected, weighed (3 g), and diluted with 27 mL of phosphate buffered saline (PBS) (a combination of NaCl: 8 g/L, KCl: 0.2 g/L, Na₂HPO₄: 1.44 g/L, and KH₂PO₄: 0.24 g/L) in a 1:10 (w/v) ratio. Serial dilutions up to 10⁻⁵ were performed. Of the homogenized sample, 0.1 mL were spread on Chromocult Coliform ES agar plates (Merck, Germany) incubated at 37 °C for 24 h and Aeromonas Medium Base (Ryan's medium) supplemented with ampicillin (AMP) at 5 mg/L (Oxoid, United Kingdom) and incubated at 30 °C for 48 h.

Water samples were diluted into PBS in a 1:10 (v/v) ratio and vortexed for 5 s before serial dilution up to 10⁻³ in the same buffer. Volumes (0.1 ml) of each dilution were spread and treated as reported above for fish samples.

Sediment samples were suspended in a 1:10 (w/v) ratio, vortexed for 5 min, and serially diluted up to 10⁻³ in PBS (Cai et al. 2019). The treatment of these samples was the same as that described for fish and water samples.

The Aeromonas Medium Base was used to detect the presence of Aeromonas species. The opaque dark green presumptive colonies with darker centers were selected for further characterization. All isolates were stored in tubes containing Luria-Bertani (LB) (Trypton: 10 g/L, yeast extract: 5 g/L, and NaCl: 5 g/L) broth with 30% v/v glycerol at −80 °C for further analysis.

Each bacterial isolate was tested for antibiotic susceptibility. Antibiotic susceptibility was determined by the disk diffusion method (Baron et al. 2017). Inoculum was prepared by mixing three or more culture colonies in 4 mL of PBS and adjusted to the turbidity equivalent to a 0.5 McFarland standard (Remel™, United States).

The culture suspension was spread onto Mueller-Hinton agar (MHA) (HiMedia, India) (4 mm depth equivalent to 25–30 mL) in three directions using sterile swabs (∼3 rotations of 60° angle). Plates were allowed to dry for a few minutes and antibiotic disks were placed at 3 cm intervals on the agar surface using a disk dispenser (Oxoid, United Kingdom). Plates were then incubated at 35 ± 2 °C for 16–18 h, and the diameter of inhibition zones was measured.

Table 2 shows the nine antibiotics and their sensitivity concentrations. The zone diameter breakpoints in this study followed the Enterobacteriaceae breakpoints (CLSI 2018) for Enterobacteriaceae and Pseudomonadaceae species tested against seven antibiotics such as AMP (10 μg), OT (30 μg), cefpodoxime (CPD) (10 μg), CIP (5 μg), W (1.25 μg), gentamicin (CN) (10 μg), and FFC (30 μg). Aeromonas spp. tested against six antibiotics such as OT (30 μg), oxolinic acid (OA) (2 μg), W (1.25 μg), E (15 μg), CN (10 μg), and FFC (30 μg), and Interpretive Categories and Zone Diameter ECVs followed Aeromonas salmonicida as described in VET04 (CLSI 2020). As FFC belongs to the same class as chloramphenicol and has similar characteristics, the zone diameter breakpoint of FFC follows that of chloramphenicol for Enterobacteriaceae and Pseudomonadaceae species. E. coli ATCC 25922 was used as a quality control strain for antibiotic disc as recommended by CLSI (2018). The multiple antibiotic resistance (MAR) index was calculated as the ratio between the number of antibiotics that an isolate is resistant to and the total number of antibiotics the organism is exposed to. A MAR greater than 0.2 means that the high-risk source of contamination is where antibiotics are frequently used.

Matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) is used to quickly identify microorganisms by acquiring a spectrum that is characteristic of each bacterial species (Croxatto et al. 2012). First, each bacterial isolate was grown on MHA agar (Oxoid, UK) for 24 h at 30 °C. Then, a sterilized loop was used to pick up a small portion of the colonies, and this inoculum was directly smeared as a thin film directly onto a sample position on the MALDI target plate. Next, 1 μL of 70% formic acid was overlaid and dried at room temperature. Then, 1 μL of HCCA matrix solution was overlaid for 30 min at room temperature. The target was ready for analysis using the MALDI-TOF bench (Bruker Daltonics, Microflex) (Kanak & Yilmaz 2019).

Forty-eight (17 from fish samples, 23 from water samples, and 8 from sediment samples) were screened from all 17 ponds of the study sites. Those appearing dark green, opaque, and with the dark center were confirmed to be Aeromonas spp. by MALDI-TOF. The results from MALDI-TOF confirmed that only nine isolates identified as Aeromonas spp., accounting for 19% of the 48 strains. Four Aeromonas spp. were identified, including Aeromonas caviae (n = 2), Aeromonas hydrophila (n = 2), Aeromonas ichthiosmia (n = 1), and A. salmonicida (n = 4) (Table 3). In addition, 34 isolates belonging to Enterobacteriaceae were found, accounting for 70%, including Citrobacter freundii (n = 8), Citrobacter braakii (n = 2), Enterobacter cloacae (n = 1), E. coli (n = 4), Klebsiella oxytoca (n = 1), Klebsiella pneumoniae (n = 3), Plesiomonas shigelloides (n = 1), Proteus hauseri (n = 3), Proteus vulgaris (n = 3), Rahnella aquatilis (n = 3), and Rahnella inusitata (n = 2). In addition, only one isolate of Providencia alcalifaciens was detected in the water. Two species of Serratia, Serratia marcescens (n = 1) and Serratia nematodiphila (n = 1), were also detected.

Four species of Pseudomonas, accounting for 10% of overall isolates, were identified, namely, Pseudomonas fragi (n = 1), Pseudomonas guariconensis (n = 2), Pseudomonas putida (n = 1), and Pseudomonas stutzeri (n = 1).

According to the manufacturer of Aeromonas medium supplemented with AMP (5 mg/L), presumptive Aeromonas will appear to be dark green and opaque. A. hydrophila must be dark green and have a dark center. Adding AMP is supposed to inhibit other non-resistant bacteria including Enterobacteriaceae but not Aeromonas spp. which are intrinsically resistant to AMP. However, results showed that 70% of species isolated in Aeromonas media were Enterobacteriaceae species, in numbers higher than those of Aeromonas spp. (n = 9). The addition of 5 mg/L AMP is not sufficient to inhibit the isolated Enterobacteriaceae when they are developing resistance to AMP.

Total isolates were 28 from Mgi and 9 from Pgi, followed by 8 from Mpa and 3 from Ppa. Overall, 37 of the total 48 isolates were obtained from giant snakehead environment, where Enterobacteriaceae predominated. However, more investigation is needed to confirm this assumption.

Aeromonas spp. was detected in the three matrices (fish, water, and sediment). A recent study also found that a high frequency of Aeromonas spp. is responsible for fish diseases, with the interesting thing being that some strains have been associated with human disease (Huddleston et al. 2013). Almost 95% of human isolates are in decreasing order belonging to the species A. caviae, Aeromonas veronii, Aeromonas dhakensis, and A. hydrophila (Figueras & Beaz-Hidalgo 2015). Enterobacteriaceae are Gram-negative, catalase-positive, function with facultative aerobics, and are non-glucose fermenters (Kaper et al. 2004). These bacteria are usually associated with the gastrointestinal tract of fish. Some studies have shown that some of their species, such as E. coli, Enterobacter spp., and K. pneumonia, are isolated from fish in pisciculture (Yagoub 2009). Although in most cases Enterobacteriaceae are part of normal microbiota from fish when colonizing human organs and tissues, they can cause some diseases, like urinary tract infections (Guzmán et al. 2004; Nagamatsu et al. 2015). In other studies, Enterobacteriaceae such as Edwardsiella tarda, S. marcescens, Klebsiella aerogenes, Proteus penneri, P. hauseri, E. cloacae, Enterbacter cancerogenus, Enterbacter ludwigii, C. freundii, E. coli, Kluyvera cryocrescens, P. shigelloides, and Providencia vermicola were recovered from infected freshwater goldfish (Preena et al. 2021).

Aeromonas spp. are frequently isolated from diseased fish and cause infection in humans in rare cases (Huddleston et al. 2013). Aeromonas species have been found in diseased fish, including those with conditions like motile Aeromonas septicemia, hemorrhagic septicemia, ulcer disease, and red-sore disease (Patil et al. 2016). Aeromonas outbreaks commonly occur in aquaculture facilities, as these bacteria are highly ubiquitous in freshwater bodies. A. salmonicida and A. hydrophila are known to cause ulcerative and hemorrhagic skin ulcers in fish under stress, which is often associated with poor sanitation and nutritional deficiencies (Igbinosa et al. 2012; Preena et al. 2020). A. hydrophila and other Aeromonas species (i.e., A. caviae, A. veronii, Aeromonas sobria, and Aeromonas schubertii) are the main causative agents of hemorrhagic septicemia in warm-water fish (Gilani et al. 2021). Moreover, E. coli of the Enterobacteriaceae family is an indicator microorganism; its presence in environmental samples, food, or water usually indicates recent fecal contamination or poor sanitation practices, and some are even pathogenic. The important thing to consider is that both microorganisms not only infect fish but can also cause human disease (Jamal et al. 2020).

All Enterobacteriaceae isolates were 100% resistant to AMP, followed by W (19%). A few isolates were resistant to OT, CPD, CN, and FFC. Two among four isolates of E. coli were found to be multidrug-resistant isolates (MAR > 0.2) and were resistant to AMP, CPD, W, and FFC. Unlike in our study, Reed et al. (2019) found that E. coli isolates from human samples, pigs, and chickens were resistant to AMP, fluoroquinolone, and CN. Eight out of nine Aeromonas isolates were resistant to W and seven out of nine were resistant to OA. Five isolates were resistant to OT, and a few were resistant to CN, FFC, and erythromycin.

Most of Enterobacteriaceae isolates in this study were susceptible to CIP. CIP is a second-generation fluoroquinolone antibiotic used to treat a number of bacterial infections in humans. A high prevalence of CIP-resistant Salmonella spp. isolates from human bloodstream infections in Cambodia have been reported (Reed et al. 2019). E. coli and K. pneumoniae isolated from Thailand showed high resistance rates to third-generation cephalosporins and CIP (Reed et al. 2019). The majority of Salmonella and Campylobacter species sampled from chicken meat retailers in Phnom Penh markets were resistant to nalidixic acid, amoxicillin, CIP, and cephalothin (Reed et al. 2019).

Although there was no resistance to CIP among isolates in our study, it is necessary to strictly control the use of this antibiotic, which is not recommended to use in aquaculture (Reed et al. 2019). A strong resistance rate for CIP was observed in E. coli recovered from several hospital effluents and in neighboring aquaculture sites (Girijan et al. 2020). This emphasizes the strong link between aquaculture and human activities through water and highlights the need for careful selection of aquaculture sites.

In our study, Pseudomonadaceae were resistant to AMP (5/5 isolates), CPD, W, and FFC (3/5 isolates) and sensitive to OT, CIP, and CN.

When the MAR index score is greater than 0.2, this indicates that bacteria are multidrug-resistant, namely, they are resistant to three or more antibiotics, and they have a high-risk potential (Figure 2(c)). Fourteen out of 34 (41%) bacterial isolates belonging to Enterobacteriaceae were found to be high-risk isolates according to the MAR index score higher than 0.2, and 2 out of 9 Aeromonadacea, with a score lower than 0.2. However, in most of the isolated strains (8/9 isolates), their MAR index score was higher than 0.2.

In total, 24 of the 48 isolates (50%) had an MAR index score greater than 0.2 and were multidrug-resistant. This is an alarming sign of antibiotic resistance in the aquaculture system. The presence of multiple ARB in the aquaculture environment can pose ecological risks to aquatic organisms and is responsible for the spread of antibiotic-resistant genes and ARB in aquatic ecosystems. These resistant bacteria can be transmitted to humans through food consumption and direct contact with animals and their environment. This situation raises serious concerns regarding human health and the health of the aquatic environment.

Antibiotics used in aquaculture and in human medicine are very similar (Sapkota et al. 2008; Naviner et al. 2011), and, therefore, the development of resistance of pathogens to antimicrobials important in human medicine is of utmost concern. The misuse and overuse of antibiotics have led to the emergence of ARB in the environment, an increase in antibiotic resistance in fish pathogens, the transfer of these resistance determinants to bacteria and then to terrestrial farmed animals, and finally, the development of human pathogens along with changes of the bacterial flora both in sediments and in the water. In addition, the misuse of antimicrobial agents in aquaculture can increase the prevalence of resistant bacteria that can be transmitted to humans and cause infections. Such a direct transfer of resistance from aquatic environments to humans may occur through (1) consumption of aquaculture food products or through drinking water and (2) direct contact with water or aquatic organisms or through the handling of aquaculture food products (López-bueno et al. 2020).

As mentioned earlier, there is no KAP for aquaculture in Cambodia, but a study conducted on duck farming (Om & McLaws 2016) reported the misuse and overuse of antibiotics at the farm level as well as a lack of awareness, weak infection prevention and control, unregulated access, self-medication, inadequate training of community health workers, and low-quality counterfeit drugs as problems rife in Cambodia. It is likely that these factors are responsible for, or contribute to, the high prevalence of AMR observed in our study.

Therefore, education on the prudent use of antibiotics in food animals and regulations are urgently needed in food animal farming in Cambodia (Om & McLaws 2016). We believe that this need is also urgent for Cambodian aquaculture.

This study highlights the widespread occurrence of AMR of sentinel bacteria associated with striped catfish (P. hypophthalmus) and giant snakehead (C. micropeltes), the two fish most reared in the pond culture environment in Cambodia. The detection of multiple ARB belonging to the families Aeromonadaceae, Enterobacteriaceae, and Pseudomonadaceae emphasizes the need for action to mitigate the risks associated with improper antibiotic use in aquaculture. Antibiotic susceptibility testing revealed antibiotic resistance patterns in 24 of 48 isolates, accounting for 50% with higher MAR index scores (>0.2). Overall, this study contributes to our understanding of the AMR status in aquaculture farms in the study area of Cambodia and provides valuable insights for management and reduction of antibiotic resistance in this sector.

The authors are grateful to the Ministry of Europe and Foreign Affairs and the French Embassy through the AQUACAM-FSPI project for financial support for this research, to the Platform for Aquatic Ecosystem Research and Cambodia Higher Education Improvement Project (Credit No. 6221-KH) for providing laboratory facilities, and to the Department of Aquaculture Development of Fisheries Administration, Cambodia for their support with sample collection.

CP contributed to the conceptualization (equal), data curation (equal), formal analysis (equal), methodology (lead), validation (equal), visualization (equal), and writing the original draft (lead). SM contributed to the investigation (equal) and validation (equal). PK contributed to the investigation (equal) and validation (equal). SChe contributed to the investigation (equal) and validation (equal). CT contributed to the investigation (supporting), resources (equal), and supervision (supporting). OH contributed to data curation (equal), formal analysis (equal), and methodology (Supporting). SS contributed to the supervision (equal) and validation (equal). SCh contributed to the validation (equal). DC contributed to the conceptualization (equal), formal analysis (equal), methodology (equal), supervision (equal), validation (equal), and review and editing the manuscript writing (equal).

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.