1. Introduction
A total of five sea turtle species have been recorded in Taiwan, including the green turtle (
Chelonia mydas; endangered), hawksbill turtle (
Eretmochelys imbricata; critically endangered), olive ridley turtle (
Lepidochelys olivacea; vulnerable), loggerhead turtle (
Caretta caretta; vulnerable), and leatherback turtle (
Dermochelys coriacea; vulnerable) [
1]. All of the species of sea turtles are included in the Red List of Threatened Species, which is maintained by the World Conservation Union (IUCN Red List). In addition, they are listed under the Schedule of Protected Marine Species, which is managed by the Ocean Affairs Council of the Executive Yuan in Taiwan. Nevertheless, sea turtles have been adversely affected by anthropogenic activities, including non-target bycatch, coastal development, marine debris, global environmental change, marine pollution, and anthropogenic-exacerbated diseases such as fibropapillomatosis (FP), among others [
1,
2,
3,
4,
5,
6,
7]. As sea turtles play a pivotal role in maintaining the health of marine ecosystems [
8,
9], previous studies have utilised these animals as biological indicators to determine pollution levels in local marine environments [
10,
11,
12,
13,
14]. Due to their high fidelity to coastal feeding grounds and nesting sites, their longevity, and their frequent use of nearshore habitats affected by anthropogenic activities, marine turtles are frequently exposed to antimicrobial-resistant bacteria, antimicrobial residues, and heavy metals. This is particularly prevalent in areas with a high level of environmental pollution [
6,
11,
15,
16,
17,
18].
Antimicrobials are regarded as one of the most significant emerging contaminants, due to the paucity of data regarding their cumulative toxic effects on aquatic organisms and the fact that their continued presence leads to the development of antimicrobial-resistant bacteria [
19,
20]. In fact, it has been demonstrated that anthropogenic activities resulting in the generation of agricultural and aquaculture wastes, in addition to water run-off, can facilitate the introduction of multiple antimicrobial resistance organisms and their associated resistance factors into the environment and water bodies [
21,
22,
23,
24]. In particular, the excessive and inappropriate use of antimicrobials (including nitrofurantoin, amikacin, and gentamicin) in human and veterinary medicine, as well as in agriculture and aquaculture, has led to an increase in selective pressure on bacteria, resulting in the emergence of antimicrobial-resistant strains and the dissemination of antimicrobial resistance throughout various environments [
19,
23,
25,
26,
27,
28,
29,
30,
31,
32,
33,
34]. Hence, the use of antimicrobials in the aquaculture industry has been prohibited or severely regulated in numerous countries across Europe and North America due to the potential for adverse effects, the presence of residues in animal tissue, the emergence of bacterial resistance, and the risk of environmental contamination [
23,
30,
35,
36]. Nevertheless, they are used in other regions, particularly in Asia [
19,
30,
37,
38,
39].
Additionally, the development of bacterial resistance to antimicrobials can be influenced not only by the antimicrobials themselves but also by other pollutants, such as heavy metals. Previous studies have demonstrated a correlation between heavy metals and the selection of antimicrobial resistance genes, indicating that these pollutants can induce resistance in metal-contaminated environments [
40,
41,
42,
43,
44,
45]. A study by Fang et al. (2020) [
46] revealed that 22.32% of the
V. parahaemolyticus isolates identified in Pacific mackerel exhibited multi-heavy metal resistance. A study conducted in southwest Nigeria demonstrated that all
Vibrio isolates from water in aquaculture ponds with elevated metal contamination levels exhibited tolerance to five metals (copper, zinc, lead, nickel, and chromium) [
47]. It has been demonstrated that antimicrobial-resistant
Vibrio isolates demonstrate tolerance to heavy metals, including cadmium, zinc, and copper [
41,
48,
49]. Furthermore, in 2018 Kang et al. reported antibiotic and heavy metal resistance (Ba
2+, Co
3+, Cd
2+, and Cu
2+) of
V. ahemolyticus (
n = 59) isolated from oysters in Korea [
50]. However, there are currently no published data on the heavy metal resistance patterns of
Vibrio strains in sea turtles in Taiwan.
The aforementioned circumstances have the consequence that aquatic environments act as reservoirs of antimicrobial-resistant bacteria from a variety of sources. These include human wastewater, hospital effluents, and animal and plant agricultural runoff [
23,
24,
51,
52]. It is well documented that sea turtles typically spend a significant portion of their lives submerged beneath the surface of the aquatic environment, with the majority of their time spent in this state [
53,
54]. Furthermore, it has been demonstrated that antimicrobial-resistant bacteria have been detected in both healthy wild sea turtles and those that are injured or unwell and undergoing rehabilitation; as a result, it can be postulated that the effectiveness of clinical treatments for microbial infections in sea turtles that are rescued may be severely limited [
5,
6,
11,
18,
55,
56,
57]. Although previous studies have indicated a potential association between antimicrobial-resistant strains derived from sea turtles and coastal effluent pollution [
6,
11,
12,
15,
58], further investigation is required to elucidate this hypothesis. Because sea turtles exhibit high site fidelity throughout their various life stages (neritic and oceanic) [
18,
59,
60,
61], they are optimal candidates for sentinel programs investigating the sources of antibiotic-resistant bacteria in sea turtles in Taiwan. The early juvenile stage, which is observed in the open ocean, is referred to as the oceanic stage. In contrast, the juvenile stage that is found in coastal waters is designated as the neritic stage [
18,
59,
60,
61].
In recent decades, there has been a notable rise in antimicrobial resistance among various bacterial species, including
Vibrio. This phenomenon may be attributed to the pervasive misuse of antimicrobials in both agricultural and human systems [
20,
24,
50,
62,
63]. Vibriosis, the most prevalent bacterial disease affecting a wide range of marine and estuarine fish species, is primarily caused by
V. harveyi,
V. parahaemolyticus, and
Vibrio vulnificus, which are among the most frequent fish pathogens, leading to significant economic losses in marine or estuarine aquaculture practices [
24]. Furthermore,
Vibrio species have often been found in sea turtles around the world [
6,
14,
16,
17,
64]. Tsai et al. (2021) reported
Vibrio spp. as the most dominant (31.91%) species in Taiwanese sea turtles, with
V. alginolyticus (46.66%),
V. harveyi (20.00%),
V. vulnificus (20.00%),
V. cholerae (6.66%), and
V. metschnikovii (6.66%) being the most prevalent
Vibrio species identified in that study [
6].
It has been proposed that the assessment of antimicrobial resistance and the identification of an association between different antimicrobial resistance characteristics can prevent the further selection of resistance, thus constituting a critical tool in the development of efficient control guidelines [
24,
65]. Given the aforementioned circumstances, this study aimed to investigate further evidence of anthropogenic impact on sea turtles by comparing the levels of phenotypic antimicrobial resistance and heavy metal resistance genes from
Vibrio species found in different life stages of these reptiles.
3. Discussion
Although no significant differences in MAR index values were observed between neritic-stage and oceanic-stage groups (which may be attributed to the relatively limited sample size in our study), it was found that
Vibrio spp. isolates with a multiple antimicrobial resistance (MAR) pattern and a MAR index value greater than 0.2 were more likely to be observed in neritic-stage sea turtles. Furthermore, isolates from neritic sea turtles were more resistant to most of the antimicrobial agents tested (except beta-lactams and macrolides) than isolates from oceanic sea turtles. It was also observed that isolates derived from neritic sea turtles exhibited greater resistance to nitrofurans (nitrofurantoin) and aminoglycosides (including amikacin and gentamicin) than those derived from oceanic sea turtles. We further found that 35% of isolates (all derived from neritic-stage turtles) were resistant to nitrofurans (nitrofurantoin) and aminoglycosides (including amikacin and gentamicin). Nevertheless, previous studies conducted over the past few decades have indicated that the majority of
Vibrio spp. isolates were sensitive to nitrofurantoin, gentamicin, and amikacin. For example, Amaro et al. (1999) [
66] isolated several
Vibrio vulnificus strains that were sensitive to nitrofurantoin, amikacin, and gentamicin from aquatic habitats in Taiwan [
66]. Moreover, Liu et al. (2004) [
67] also reported that
Vibrio alginolyticus isolates obtained from diseased white shrimp (
Litopenaeus vannamei; also known as
Penaeus vannamei) in Taiwanese culture ponds were all sensitive to nitrofurantoin [
67]. In a retrospective analysis of 84 clinical case studies of patients with
Vibrio vulnificus infection in Taiwan over a period of six years, from 1995 to 2000, Hsueh et al. (2004) [
68] found that all isolates demonstrated susceptibility to gentamicin and amikacin. However, in recent decades, a significant number of bacteria, including
Vibrio, have emerged with an unprecedented resistance to a wide range of antimicrobials. This could be a consequence of their misuse in both agricultural and human systems [
20,
62]. As reported by Baralla et al. (2021) [
19], aminoglycosides and nitrofurans were used in aquaculture in China from 1996 to 2013. Nitrofurantoin was commonly employed in veterinary medicine as a treatment for protozoan and bacterial infections [
31]. Aftabuddin et al. (2009) [
38] reported the use of nitrofurans as a prophylactic agent for
Penaeus monodon vibriosis in Bangladesh. Tonguthai (2000) [
37] reported that nitrofurans were employed as a chemotherapeutic agent for the treatment of shrimp hatchery vibriosis in Thailand. Furthermore, amikacin and gentamicin are frequently employed in both human and veterinary medical industries to treat a spectrum of gram-negative bacterial infections [
32,
33,
34]. The correlation between the emergence of aminoglycoside resistance and the extensive use of antimicrobials, particularly in the context of the prevention and treatment of
Vibrio cholerae, as well as other applications, is well documented [
20]. A study by Redpath et al. (2021) [
69] has indicated that the veterinary use of aminoglycoside antimicrobials is currently under increasing scrutiny. In particular, gentamicin was commonly used empirically without bacterial culture and susceptibility testing [
69]. Furthermore, the presence of
Vibrio spp. that are non-susceptible to antimicrobial agents represents a significant concern in the context of plastic pollution. The presence of
Vibrio spp. on microplastics has been demonstrated in a number of studies. These bacteria are found in high abundance on these particles [
70,
71]. In light of these observations, Canellas et al. (2021) [
48] postulate that
Vibrio spp. may be acquiring antimicrobial resistance genes (ARGs) from bacteria released into marine environments through sewage, and subsequently dispersed via plastic debris or in the form of plankton throughout the bay. This may have the potential to endanger the health of those exposed to this environment. A previous study by Kim et al. [
33] indicates that a reduction in the consumption of aminoglycosides is associated with a reduction in the prevalence of resistance. This indicates the necessity of the implementation of an antimicrobial cycling strategy at the national level.
In this study, 48.9% of
Vibrio species from sea turtles were found to have a MAR pattern, indicating resistance to at least three different classes of antimicrobials. Additionally, a statistically significant difference was observed between the nine
Vibrio species and their respective MAR patterns (
p = 0.033; Fisher’s exact test). The rank order of MAR patterns of the
Vibrio spp. isolates evaluated was as follows:
V. rotifetianus (1/1; 100.0%) >
V. harveyi (7/9; 77.8%) >
V. vulnificus (3/4; 75.0%) >
V. parahaemolyticus (4/6; 66.7%) >
V. alginolyticus (8/18; 44.4%) >
V. campbellii (0/5; 0.0%),
V. fluvialis (0/2; 0.0%),
V. mediterranei (0/1; 0.0%), and
V. cyclitrophicus (0/1; 0.0%) (
Table 1). Among the 47
Vibrio isolates identified in this study,
V. alginolyticus (18/47) and
V. harveyi (9/47) were the most prevalent and second most prevalent species, respectively (
Table 1). Among these, a total of eight out of 18 isolates of
V. alginolyticus exhibited a MAR phenotype.
V. alginolyticus is frequently found in sea turtles’ lesions [
18,
72,
73]. Furthermore,
V. alginolyticus also plays a multitude of roles in marine turtles. For instance, it is a constituent of the common flora in the oral cavity and cloaca of healthy sea turtles captured from foraging grounds [
74]. Additionally, it has been identified as an opportunistic pathogen [
75,
76]. It has been linked to the development of dermatological conditions in an aquarium-reared loggerhead turtle [
77]. The collective findings of the current and previous studies indicate that
V. alginolyticus with MAR should be regarded as an important pathogen in sick and injured neritic-stage sea turtles in rehabilitation centres. With regard to public health, the species
V. alginolyticus,
V. harveyi,
V. parahaemolyticus, and
V. vulnificus are the major species often associated with human infections [
24,
78]. It is therefore recommended that veterinarians working in sea turtle rehabilitation facilities should be aware of the risk of exposure to these bacteria.
A total of 40.4% of all isolates in this study exhibited MAR index values exceeding 0.2, indicating a high risk of antimicrobial contamination. A MAR index score above 0.2 signifies a notable level of antimicrobial use in the area and predicts a high-risk environment for the spread of antimicrobial resistance. [
24,
30,
79,
80]. In this study, the MAR index exhibited a range between 0.00 and 0.73. This is a cause for concern, given that a study by Mohamad et al. (2019) [
81] found that approximately 75% of
Vibrio isolates from diseased fish in Malaysia also had a MAR index exceeding 0.2. Fernandes et al. (2021) [
82] reported a low prevalence of multiple-antimicrobial-resistant gram-negative bacteria isolated from loggerhead sea turtles (
Caretta caretta) in Cape Verde. Only two isolates (10%) with MAR index values greater than 0.2 were detected in that study: 0.25 for
E. cloacae and 0.33 for
A. hydrophila/caviae. We also found that the mean MAR index of
V. alginolyticus identified in our study (ranging from 0.07 to 0.73) was significantly higher than that reported in wild loggerhead turtles (MAR index ranging from 0.00 to 0.08) [
82]. Compared to studies by Sony et al. (2021) [
24], who reported MAR indexes ranging from 0.05 to 0.47 in isolates from live diseased fishes, Ha et al. (2023) [
30], who reported indexes ranging from 0.25 to 0.67 in farmed
Litopenaeus vannamei, and Mohamad et al. (2019) [
81], who reported indexes ranging from 0.06 to 0.56 in marine fishes, the results of our study imply a higher MAR index for
Vibrio spp. in our study area.
Although our study found no significant difference in the MAR index between isolates from turtles found in different sea turtle stranding areas, the proportion of
Vibrio spp. isolates with a MAR index greater than 0.2 in neritic-stage turtles (42.5% of isolates) was higher than the proportion found in oceanic-stage turtles (28.6% of isolates). These results suggest that sea turtles in neritic foraging habitats are at higher risk of antimicrobial contamination than those in pelagic environments [
24,
30,
79,
80,
81,
83]. As noted in a previous study, the number of oceanic-stage juveniles (curved carapace length < 30 cm) [
59,
61] identified in stranding reports in Taiwan is limited [
1]. In other words, most of the sea turtle strandings found in Taiwan were neritic-stage individuals [
84]. Indeed, Taiwan’s coastal waters are known to contain green turtle feeding and migration habitats [
85,
86,
87]. Furthermore, several studies have reported that sea turtles are often exposed to antimicrobial residues and antimicrobial-resistant bacteria, which are common in polluted areas, due to their high frequency of coastal foraging, their long lifespans, and their frequent use of nearshore habitats affected by anthropogenic activities [
6,
11,
15,
16,
17]. Although wild sea turtles are unlikely to have been exposed to antimicrobial therapy under natural conditions, resistance in bacteria isolated from wild sea turtles has become an increasing concern [
5,
6,
13,
14,
58,
82]. It is possible that the multiple-antimicrobial-resistant strains isolated from wild sea turtles are related to coastal wastewater pollution. Specifically, selection pressure may cause bacteria in marine habitats to develop resistance after exposure to antimicrobial-containing environmental effluents (e.g., agricultural, aquaculture, human, and veterinary waste effluents) [
11,
12,
13,
14,
15,
43,
82,
88,
89,
90,
91,
92].
Bacterial resistance to antimicrobials can also be influenced by other pollutants, such as heavy metals. For example, heavy metals may induce antimicrobial resistance in metal-contaminated environments, as previous studies have shown that heavy metals are correlated with the selection of antimicrobial resistance genes [
41,
42,
43,
44]. Although antimicrobial resistance genes can occur naturally in various environments in low abundance, studies have demonstrated that their abundance increases in the presence of several pollutants, including crude oil, sewage, and heavy metals [
43,
44,
93,
94]. In addition, the presence of heavy metal-resistant bacteria can indicate the degree of contamination in the environment and is frequently linked to the co-selection of antimicrobial resistance [
48,
95,
96]. Mechanisms of antimicrobial and heavy metal resistance co-selection in environmental bacteria include cross-resistance, co-resistance, and co-regulation [
97,
98,
99]. In our study, the rank order of the metal resistance gene detection rate of the
Vibrio spp. isolates evaluated was as follows:
arsC (70.2%; 33/47) >
chrA (68.1%; 32/47) >
merA (34.0%; 16/47) >
copA (25.5%;12/47) >
czcA (0.0%; 0/47). The analysis revealed a statistically significant differentiation between the nine
Vibrio species and the rates of detection of their respective metal resistance genes, specifically
chrA. (
p = 0.021; Fisher’s exact test) and
copA (
p = 0.022; Fisher’s exact test) (
Table 4). The most prevalent species with
chrA and
copA, respectively, were
V. parahaemolyticus (6/6; 100.0%) and
V. vulnificus (3/4; 75.0%) in our study. In addition, the presence of heavy metal resistance genes in isolates was more common in neritic-stage turtles than in oceanic-stage turtles (CCL < 30 cm), except for
arsC, which was more common in oceanic-stage turtles. Fang et al. (2020) [
46] reported in their study that 22.32% of
Vibrio parahaemolyticus isolates identified from Pacific mackerel showed multi-heavy metal resistance (MHMR). A study conducted in southwest Nigeria showed that all
Vibrio isolates from water in aquaculture ponds with high levels of metal contamination were able to tolerate five metals (copper, zinc, lead, nickel, and chromium) [
47]. The heavy metal resistance gene (
copA) has also been detected in
Vibrio spp. isolates isolated from clams (
Meretrix meretrix) in China [
49]. Additionally, the majority of antimicrobial-resistant isolates in the same study showed tolerance to heavy metals, including Cd, Zn, and Cu [
49]. A study conducted in China found that
V. parahaemolyticus isolates from
Penaeus vannamei at freshwater farms, seawater farms, and markets exhibited both multidrug resistance (MDR) and MHMR [
41]. On the other hand, in a study conducted in Brazil, the detection of heavy metal resistance genes (
copA and
merA) was observed in antimicrobial-resistant
Vibrio isolates [
48]. In our study, a significant association was observed between MAR and the presence of
merA (mercuric reductase gene) (
p = 0.010). Similarly, a significant association was also observed between the presence of the
merA gene and MAR index values greater than 0.2 (
p = 0.027). In previous research on trace elements, methylmercury, and polybrominated diphenyl ethers in sea turtles in the South China region and Taiwan, liver methylmercury (
MeHg) levels were 6–750 times higher than the range of
MeHg levels observed in green turtles in Baja California, Mexico, and similar to those estimated for green turtles in the Mediterranean, a historically industrialised area [
53]. Furthermore, a marginally significant association was also identified between the presence of the
chrA gene and MAR index values exceeding 0.2 (
p = 0.051) in our study. These findings suggest that heavy metal accumulation in the marine environment may induce bacterial resistance to these metals and co-select for resistance to antimicrobials [
41,
44,
94,
100].