Identifying Bacteria with Public Health Significance from Farmed Nile Tilapia (Oreochromis niloticus), Zambia

Zambia has seen rapid development in aquaculture, and in recent years, the industry has experienced disease outbreaks where fish have increasingly become a potential contributor to emerging bacterial zoonotic diseases. The aim of this study was to identify bacterial pathogens with zoonotic potential in apparently healthy fish and water from their habitat. A total of sixty-three fish were sampled, and fifty-nine water samples were collected from the habitats of these fish. Bacteria were cultured from the internal organs of fish and water, and these were identified through standard bacteriological methods comprising morphological characterization, Gram-staining, and a panel of biochemical tests. The following bacterial pathogens with zoonotic potential were identified at a farm prevalence of Aeromonas (13.2%), Bacillus (2.1%), Clostridium (2.1%), Escherichia coli (0.7%), Klebsiella (6.9%), Lactococcus (2.1%), Listeria (0.7%), Staphylococcus (18.1%), and Streptococcus (0.7). Other bacteria with varying significance as fish pathogens identified included Acinetobacter (2.1%), Aequorivita (1.4%), Aerococcus (1.4%), Bordetella (2.1%), Carnobacterium (10.4%), Citrobacter (3.5%), Corynebacterium (1.4%), Dermatophilus (1.4%), Enterococcus (2.1%), Flavobacterium (4.2%), Micrococcus (6.9%), Planococcus (1.4%), Proteus (1.4%), Pseudomonas (6.3%), Rhodococcus (1.4%), Shewanella (1.4%), Streptococcus (0.7%), and Vagococcus (0.7%). The current study provides baseline information for future reference and the implementation of public health guidelines with regard to potential zoonotic diseases in fish.


Introduction
Fish is the most afordable protein for many people in densely populated countries, including Zambia [1]. Increased demand for fsh protein with escalating population growth has led to rapid growth in aquaculture farming in Zambia and in other countries worldwide [1][2][3]. Rapid aquaculture development in Zambia has created a potential danger of predisposing fsh consumers and workers on fsh farms and those in fsh processing plants to fsh zoonotic disease outbreaks (FZDOs). Worldwide, about 3 billion people derive almost 20 percent of their average per capita intake of animal protein from fsh [4]. Fish nutrition is an important source of energy and protein and provides a range of essential nutrients and vitamins to many households worldwide.
Te possibility of the emergence of risk factors for FZDOs through handling or ingestion of fsh and fsh products is ever-increasing [5]. Tere is a well-documented group of pathogens indigenous to the aquatic environments which have been associated with FZDOs. Tese pathogens have been isolated from open wounds in highly exposed fshermen and fsh handlers [6]. Te principal zoonotic fsh pathogens of concern are Aeromonas hydrophila, Edwardsiella tarda, Mycobacterium marinum, Streptococcus iniae, Vibrio vulnifcus, and Vibrio damsel [7]. Mycobacterium species, Streptococcus iniae, Clostridium botulinum, and Vibrio vulnifcus are of particular zoonotic importance and concern [8]. Intensive and confned fsh rearing in aquaculture predisposes fsh to a higher risk of bacterial load on their external surfaces, and contaminated fsh are therefore more likely to transmit the infection to humans [5].
Additionally, FZDOs are often related to management factors, such as the quality and quantity of nutrients in the water and high stocking densities, which can increase the bacterial load on the external surface of fsh.
Tis study was conducted on three fsh farms located in Chirundu district of Lusaka Province and Siavonga district of Southern Province. Te purpose of this study was to identify bacteria with zoonotic potential from fsh and water. Te bacteria were collected from apparently healthy fsh and water and identifed by standard bacteriological methods comprising morphological characterization, Gram-staining, and a panel of biochemical tests. Te farms involved in this study were assessed to ascertain their health status and whether there were any disease outbreaks.

Study Area.
Te study was conducted in Lusaka and the Southern Province of Zambia. Te study areas were selected based on the number of commercial fsh production activities. Te study areas included two districts: Chirundu (16.0271°S, 28.8509°E) and Siavonga (16.5323°S, 28.7111°E ( Figure 1).

Study Design and Sampling.
A cross-sectional study was carried out during September and December 2020 to investigate bacteria of zoonotic potential from healthy fsh and their environment (water). Healthy fsh ranging from fngerlings to out-growers with varying weights ranging from 1 gram to 600 grams from farms designated A, B, and C, reared in ponds and cages were purposively sampled. Te fsh were caught by dip-netting, sacrifced by stunning to the head, and put in sterile packs. Water samples were collected from the respective ponds and cages of these fsh and from Kafue River (water source for farm A) and Kariba dam (water source for farms B and C) in sterile 100 ml sterile bottles. A total of 44 samples each from farms A (15 fsh/29 water) and C (29 fsh/15 water) were collected, and 34 samples (19 fsh/15 water) from farm B (63 fsh and 59 water samples for the whole study) were collected. Te collected fsh and water collected were transported at 4°C to the University of Zambia, School of Veterinary Medicine (UNZA) in Lusaka for further analysis. From each fsh, the following organs were collected: gills, liver, spleen, and intestines. From each of these organs, swabs were collected for bacteriology during postmortem. Swabs from sampled water after being centrifuged at 2,500 rpm for 5 minutes were also collected.
Data regarding exposure factor locations in this study were collected with on-farm visits to observe the farming sites and operations as well as the surrounding environments. Trough this, the exposure factor locations were determined (Table 1).

Bacterial Isolation.
Bacterial swabs collected from fsh organs and water were aseptically streaked on MacConkey agar (HiMedia, India), nutrient agar (NA: HiMedia, India), and blood agar (HiMedia, India). Te plates were then incubated at 37°C for 48 hrs, and pure cultures were obtained by subculturing and incubating at 37°C for 48 hrs.

Biochemical Characterization.
Te isolates were identifed by determining colony morphology which included shape, color, pigmentation, hemolytic activity, size, edges, and elevation, and afterwards, the isolates were grouped accordingly. Two to three representative isolates from each group were subjected to Gram-staining [9,10]. Conventional biochemical tests were then used to characterize the bacteria. A loopfull of bacteria was aseptically added to 5 ml of phenol red broth containing 1% sugar and incubated at 37°C for 24 hrs to test for fermentation of diferent sugars. Sulphur reduction, indole production, and motility tests were determined by using sulphide indole motility (SIM) media again by adding a loopfull of bacteria and incubating at 37°C for 24 hrs. Further identifcation of isolates was according to Buller [11] (Table 2).

Data Analysis.
Te data obtained was entered in Microsoft Excel sheets 2007, cleaned, and exported to Stata SE 12 (https://www.stata.com/) for survey analysis. A descriptive statistical analysis of quantitative bacterial counts and measurements of location was used to describe the outcome of interest. Te results were presented in percentages/proportions, and the diference in the distribution of predictor variables was considered signifcant if the p value was less than 0.05 at a 95% confdence interval. Pearson's chi-square was used to test the signifcance. A spider web analysis was used to analyze the overlap between exposure factors generated under the overall descriptive epidemiology from key thematic areas. Both Table 1 and Figure 2 present basic data based on descriptive epidemiology. Table 1 shows the four main thematic areas: demographical, biological, environmental, and management factors, and each had several descriptors listed under each thematic area (Table 1). Figure 2, on the other hand, used the data from Table 1 to visualize the extent of interaction, overlap, and, to a greater extent, dominance of the factors within these epidemiological descriptors that had a greater infuence. Using the spiderweb analysis, the list of factors within each thematic area was collapsed within the data points in the polar or spider chart. Tus, each of the thematic area was grouped together with other existing thematic areas to describe their interrelationship (Table 3).

Results
A total of 122 samples were collected including fsh 51.6% (n � 63) and water 48.4% (n � 59) from farms A, B and C. Te details are shown in Table 2.
Te identifed exposure factor locations from visual determination of the three farms under study are shown in Table 1. Te cage production system comprised the outgrower cage and the juvenile cage, while the other types of production fell under the pond production system.
Te four factor location aggregates identifed from data collected were as follows: demographical, biological, environmental, and management factors. When the analysis was undertaken, controlling for production types, demographical factor, and location covered a wider area, followed by biological factors locations ( Figure 2). Environmental factors and location were closely associated with biological factors, while management factors and locations overlapped across all factors and locations, albeit as an outlier ( Figure 2).

Discussion
Te prevalence of bacteria isolates in fsh with zoonotic potential reported in this study provides important preemptive baseline data as well as an early warning system concerning the formulation and implementation of public health guidelines for the management and control of       zoonotic diseases in fsh. Tis present study has identifed twenty-seven diferent genera of bacteria from the apparently healthy tilapia Oreochromis niloticus. Some of these isolates are well-known etiological pathogens that cause diseases in both fsh and humans. Seong-Joon et al. [12], in a similar study, reported ffteen bacteria genera isolated from eels, three of them Aeromonas, Citrobacter, and Pseudomonas species which have also been identifed in this present study. In another study from Tilapia in Trinidad, they also isolated thirteen bacteria genera with fve of these genera being Pseudomonas, Staphylococcus, Enterobacter, Bacillus, and Aeromonas species also reported in our present study [13]. Aeromonas sobria and A. hydrophila are well-known virulent pathogens of fsh worldwide, isolated from clinically sick fsh associated with high mortalities [14][15][16]. Regarding proportional representation, in our current study, Staphylococcus spp. had the highest prevalence rate at 18.1%. Aeromonas species also had a higher prevalence of 13.2%, a fnding similar to reports in Uganda [17] and other parts of the world [12]. Te other most common bacteria isolated in this present study included Klebsiella, Pseudomonas, Carnobacterium, Streptococcus, and Lactococcus spp. Te prevalence rates of the other bacteria genera were relatively low. Although theoretically, streptococcosis afects all sizes of fsh [18], bigger fsh tend to be more susceptible to other infections such as Pseudomonas [19], Flavobacterium [20], and other Streptococcus species [21][22][23]. Aeromonas, Staphylococcus, Lactococcus, and Streptococcus species are well-known zoonotic bacterial pathogens of fsh frequently isolated from diseased fsh. Te symptoms of fsh diseases associated with these bacterial pathogens are skin ulceration, abnormal swimming (swimming in circles), blindness (whitish appearance of the eyes), and exophthalmos. Other authors have also reported similar symptoms with Aeromonas species infections in fsh [24], Lactococcus garvieae [25][26][27], Streptococcus iniae [23,28], and Streptococcus agalactiae [22,29].
Bacteria from the family streptococcus are well-known opportunistic bacteria in natural aquatic environments. When poor husbandry and excessive stocking density are practiced on a farm, it predisposes the fsh to clinical diseases caused by this family of bacteria [27]. Other bacteria species isolated in this study have also been associated with a few cases of disease outbreaks in fsh with varying pathogenicity which include Pseudomonas aeruginosa [30], Citrobacter species [31], and Klebsiella species [32]. Tese bacteria species were isolated from apparently healthy fsh and cannot be directly linked to any disease. In this study, no information was readily available on the pathogenicity of Planococcus spp., Shewanella spp., Dermatophilus spp., and Micrococcus spp. in fsh. Te pathogenicity of most bacteria, although ubiquitous in the aquatic environment, depends on stress in the fsh host to cause disease [33][34][35]. Intensive fsh Table 4: Occurrence of the diferent bacteria genera on farms A, B, and C.

Bacteria genera
Occurrence n (%) Overall (n � 153) Farm A (n � 60) Farm B (n � 51) Farm C (n � 42) Te table shows occurrence of the diferent bacteria genera isolated from the three farms. 6 International Journal of Microbiology farming, high stocking density, and increased human activities in intensive fsh farming afect water quality and lead to environmental deterioration that may give rise to the emergence of rare zoonotic bacterial fsh diseases in the future [17]. Tis hypothesis is supported by having isolated the following well-documented zoonotic bacteria: Species in fsh: Aeromonas, Klebsiella, Bacillus [36,37], Staphylococcus, Listeria [37], Clostridium [20,38], E. coli [39], Lactococcus, and Streptococcus. Humans get infected through the consumption of raw or undercooked fsh, although Lactococcus, Streptococcus, and Staphylococcus spp. can be transmitted through abraded, wounded skin or through injuries caused by fsh fns or fsh spikes during handling and processing. Water samples from ponds had a signifcantly higher prevalence of bacterial contamination (p > 0.0001) than water from cages. Tis diference can partly be associated with the high stocking density of fsh in ponds compared to cages which are placed in lakes with fresh running water continuously. Water recirculation, stock movement contamination, and high organic matter deposition are more rampant in the earthen pond production system as compared to the cage production system. Transmission of bacteria from water to fsh and/or humans in aquaculture systems can also easily be facilitated by the high stocking density of fsh and by direct contamination of the pond soil lining which may be responsible for the high prevalence and diversity of bacteria observed in earthen ponds as compared to cage production systems. Certain bacterial isolates could not be identifed against any of the bacteria profles used as presented by Buller [11].
In this study, four factors that could give rise to a hazard in food safety were identifed: demographical, biological, environmental, and management factors. Tese factors can occur at various stages of the food chain (farm to table) and involve the contamination of food by diferent causal agents that can be of biological origin (parasites, viruses, bacteria, fungi, or prions) or chemicals (heavy metals, natural toxins, or organic compounds), risking the health of consumers [40,41]. Some of the causes of the loss of food safety are inadequate or absent hygiene conditions and practices, high degree of handling, and the use of contaminated water or raw materials [40][41][42].
Te exposure factors that could give rise to hazards in food safety and disease were observed from various locations on the fsh farms in this study (Table 1) during onfarm visitations and observations. Te results indicated that water source, stocking density, seining practices, control of piscivorous birds, and disposal of dead fsh were the main exposure factors to bacterial pathogens in fsh/ aquatic environments and humans. Other factors included pond preparation and treatment and excess vegetation around ponds for fsh farms that used ponds. Farm A would drain their ponds completely dry after the end of the harvesting cycle without removing the mud. Te ponds were treated with lime and left to dry out before restocking the fsh. Tis method of pond preparation and treatment, if not done correctly, increases the chances of exposing new fsh stocks to infectious pathogens from the previous fsh stocks [43].
Water supply and the frequency of changing the water is another important factor. Farms A and C with the pond production system had their water sources from the Kafue river and Kariba dam, respectively. Water was made available to the ponds via dam liners, and this water was treated before flling the ponds. Tis would at times take longer for the water to be changed at times, leading to compromised water quality due to accumulation of fsh fecal material [44]. Poor water quality leads to low dissolved oxygen levels, low pH values, and high nitrate and ammonia levels [45]. Poor water quality is also a result of overstocking. Overstocking either in cages and ponds may also lead to cannibalism behavior in fsh due to food competition. When this happens, the skin may break, injure, or get wounded, enhancing pathogen entry. In addition, food competition due to overstocking leads to stunted growth and reduced immunity [43].
Seining practices are also part of exposure factors that lead to the spread of pathogenic factors on fsh farms, as observed on the fsh farms in our study. Farmers used the same fshing nets between ponds in farms A and C and between cages in farm B between fsh harvesting without prior disinfection. Te seining equipment may act as a vehicle to transfer infectious agents from one pond/cage to another. A correlation between seining activities and transmission or outbreaks of Aeromonas hydrophila in cultured catfsh was reported in Alabama, USA [46]. Te use of separate equipment in diferent fsh ponds/cages and prior disinfection are recommended.
One of the ways to reduce the chances of an infection spreading on a fsh farm is to promptly remove and appropriately dispose of moribund and dead fsh. In this study, dead fsh were left in ponds/cages for longer before removal or left to be feasted upon by other fsh, and this perpetuates the cycle of infection. Cannibalism of infected fsh acts as a possible route for the transmission of streptococcal infection in fsh. Te farmers who used disposal methods buried the dead fsh in pits, though at some farms, it was observed that workers would eat the moribund or recently dead fsh. Piscivorous birds can also feast on moribund or dead fsh and transmit infection from one pond or cage to another. Tese birds are also involved in the transmission of digenean parasites, Francisella spp., Edwardsiella tarda, and viral pathogens [43]. In this study, the most common methods of controlling these birds include the use of bird nets and physical chasing, though these are not very efective as birds could still have access to fsh.
As with many research activities, limitations were present. In our case, time and resources (fnancial) were among the limiting factors. A longitudinal study was going to be suitable instead of the cross-sectional study done. It would account for environmental changes (temperature, pH) and seasons that afect the microbiota in the water where fsh live and the microbiota of fsh. Failure to identify some of these bacteria isolates could be attributed to gaps in the diagnostic capability of the techniques used, and this could be a group of bacteria representing new genera or species in the fsh samples we collected. More time and fnancial resources are needed. Sampling and analysis biases International Journal of Microbiology were the two forms of biases encountered in this study. To lower the sampling bias, three diferent farms were involved in the sampling process. Sampling was spread across the diferent production systems (cage and pond systems), and it was nonprobabilistic; hence, our analysis did not involve the use of more statistical tools, thereby reducing the level of biases in this study.
In conclusion, given its limitations and strengths, this current study has been able to provide baseline information with regard to future reference as well as being an important stepping stone with regard to the implementation of public health guidelines which have potential zoonotic implications arising from fsh.

Data Availability
All data used for the study are available upon request from the corresponding author.

Ethical Approval
Ethical approval was granted by the University of Zambia Biomedical Research Ethics Committee (UNZABREC) (C0000474). Fish safety and welfare was considered such that appropriate collection methods were used and aseptic techniques were observed when collecting water and sampling fsh.

Consent
Consent was obtained from all study sites as well as participants in the study. Fish farmers consented (written/oral) before farm visitation and sample collection was done.

Conflicts of Interest
Te authors declare that they have no conficts of interest.