Abstract
The environment is increasingly recognised as a hotspot for the selection and dissemination of antibiotic resistant bacteria and antibiotic resistance genes. These can be selected for by antibiotics and non-antibiotic agents (such as metals and biocides), with the evidence to support this well established by observational and experimental studies. However, there is emerging evidence to suggest that plant protection products (such as herbicides), and non-antibiotic drugs (such as chemotherapeutic agents), can also co-select for antibiotic resistance. This review aims to provide an overview of four classes of non-antibiotic agents (metals, biocides, plant protection products, and non-antibiotic drugs) and how they may co-select for antibiotic resistance, with a particular focus on the environment. It also aims to identify key knowledge gaps that should be addressed in future work, to better understand these potential co-selective agents.
Similar content being viewed by others
Introduction
Antimicrobials are agents which kill or inhibit the growth of microorganisms (bacteria, fungi, viruses, and parasites)1. Antimicrobial resistance (AMR) occurs when the organisms these agents target evolve to survive their toxic effects. AMR is a global health concern, with predictions that by 2050 it could be responsible for 10 million deaths per year2. Antibiotics are a subclass of antimicrobials, that are used to target bacterial infections in humans, animals, and plants. Resistance of bacteria to antibiotics is termed ‘antibiotic resistance’, and is of significant concern since antibiotics are used frequently in medicine to treat bacterial infections, e.g., tuberculosis3, and as prophylaxis, e.g., before major surgery4. In 2019, 4.95 million deaths were associated with drug resistant bacterial infections, of which 1.27 million deaths were directly attributed to drug resistant bacterial infections5. Antibiotic resistance can be acquired through de novo mutation, or through horizontal gene transfer (HGT) of antibiotic resistance genes (ARGs) encoded on mobile DNA elements (such as plasmids, integrons or transposons) that are passed between bacterial cells of the same or different species6. HGT rate can change depending on various factors such as exposure to stressors (e.g., antibiotics), changes in pH, and oxidative stress6, therefore, additional anthropological input of environmental contaminants that can increase HGT are of concern to environmental and human health.
Strategies to address antibiotic resistance have included attempting to reduce selection pressure by decreasing clinical and veterinary antibiotic consumption7. However, antibiotics are not the only agents capable of contributing to antibiotic resistance evolution. Agents other than antibiotics have antimicrobial properties, including biocides (e.g., quaternary ammonium compounds (QACs)) and metals (e.g., copper, zinc). These agents can indirectly select for antibiotic resistance and ARGs8,9,10, through a process known as co-selection. Consequently, exposure to these agents may increase resistance to antibiotic drugs and other antimicrobial compounds, even in the absence of antibiotic selective pressure.
Co-selection refers to the simultaneous selection for resistance to multiple agents and can occur through three different processes: co-resistance, cross-resistance and co-regulation (Fig. 1). Co-resistance occurs when multiple resistance genes are genetically linked, for example, on a conjugative plasmid8,9. Cross-resistance occurs when one mechanism provides resistance to more than one agent, for example, a multi-drug efflux pump removing more than one agent from the cell8,9. Co-regulation occurs when translational and transcriptional responses are linked and produce a co-ordinated response, such as the expression of multiple separate efflux pumps, triggered by the presence of one agent9. Co-selection has been acknowledged as a key mechanism that selects for ARGs in different microbial communities.
Cross-resistance occurs when a mechanism provides resistance to two or more agents (e.g., a multi-drug efflux pump). In this figure, the AcrAB-TolC efflux pump gene is shown, which can pump out of the cell multiple antibiotics and biocides137, non-antibiotic drugs86,138, and plant protection products32,139. Co-resistance occurs when two genes are physically linked on a piece of DNA so are inherited together e.g., an antibiotic and metal resistance gene located on a plasmid. In this figure, the metal resistance gene arsB and the antibiotic resistance gene blaTEM-1 are both present on a plasmid and can be inherited together139. Co-regulation occurs when the translational or transcriptional responses to one agent leads to a co-ordinated response to more than one agent (e.g., an antibiotic or biocide could lead to expression of a multi-drug efflux pump). In this figure, the transcriptional pathways of the mex operon and the czc operon are linked so that expression of either leads to the expression of both the czc and oprD efflux pumps140. Created with Biorender.
The relevance of the environment
The environment is increasingly recognised as an important reservoir of antibiotic resistance in which ARGs may spread between bacteria and be potentially selected for by micropollutants11,12,13. Antibiotic resistant bacteria in the environment and the genes they carry can then be transmitted to humans through food, drinking water, air, or through direct contact with the environment, such as through recreational use of coastal waters14.
Selection and co-selection can occur over wide concentration ranges of selective agents, illustrating the importance of the environment, since concentrations of antibiotics and other non-antibiotic agents can vary by orders of magnitude15,16,17. Clinical breakpoints are defined using determination of the Minimum Inhibitory Concentration (MIC), which is the lowest drug concentration where growth is inhibited18. Organisms that grow in concentrations of antibiotic higher than the MIC are considered resistant. However, subinhibitory concentrations can amplify resistance or increase persistence of resistance. The Minimal Selective Concentration (MSC) is the lowest concentration of an agent where the growth of resistant and sensitive strains is equal, and concentrations between the MSC and MIC of the resistant strain positively select for (i.e., amplify) resistant strains19. The Minimal Increased Persistence Concentration (MIPC) indicates the antibiotic threshold at which the rate of loss of the resistant strain is significantly slowed, compared to when no selective pressure is present. Concentrations between the MIPC and MSC lead to prolonged maintenance of resistant strains, even though there is no positive selection or amplification occurring20. Antibiotics and other potentially co-selective agents are present in human, animal, and environmental microbiomes at these sub-MIC concentrations, particularly in the environment where they are found within the ng-g/L range15, resulting in a large temporal and geographical antibiotic resistance selection window. There is evidence to suggest that antibiotic resistance can be selected for and/or maintained in the environment, due to selective agents that are present in municipal, industrial, and agricultural pollution21,22,23,24,25.
This non-systematic review aims to provide an overview of the co-selective effects of biocides, metals, non-antibiotic drugs (NADs), and plant protection products (PPPs), with a particular focus on the environment, all within a single paper. Previous reviews have covered some of these agents individually and we also direct the reader to these (e.g., for metals9,26,27; for biocides28,29; for NADs30,31; and for PPPs32). The review also aims to highlight some of the important concerns and knowledge gaps that have been identified with regards to these agents in the understanding of AMR evolution and dissemination in an environmental context. Suggestions as to how future research might address these points are also provided.
Metals
Metals naturally occur in the environment, with elevated concentrations resulting from pollution caused by anthropogenic activity (e.g., historical and current mining activity)33. Metals are utilised by humans for numerous purposes in different settings, including disinfection in human healthcare34; antifouling agents in aquaculture35; feed additives in animal husbandry36 and crop protection in agricultural practices37. The presence of metals is often beneficial to microorganisms as they are used as micronutrients and are required for survival38. However, some metals are toxic to bacteria and even essential metals can become toxic at high concentrations, or can inhibit bacterial growth39. Metals can select for chromosomal and plasmid-borne resistance mechanisms that ameliorate the toxic effects of metals in bacterial cells. These metal resistance mechanisms can also be genetically linked to antibiotic resistance genes (i.e., co-resistance)10,40; may share their mechanism of resistance with antibiotics (cross-resistance)41; or can be expressed alongside antibiotic resistance genes as a result of linked regulatory systems (co-regulation)42.
Metal presence and contamination is ubiquitous33, so co-selection for antibiotic resistance could occur in a wide variety of environments including wastewater, freshwater, manures, and soils, depending on bioavailability. The bioavailability of metals can be affected by factors including sorption to soil or sediment particles, pH, changes to ionic composition, or redox potential43. For these reasons, the presence of metals in some environments could result in differing risk of selection for ARGs than in others. For example, environments receiving mine waste are likely to have increased concentrations of metals, and therefore may have an increased selective pressure44. Other environments, e.g., manures and wastewater15, are more likely to harbour a range of pollutants including antibiotics, ARGs, and antibiotic resistant bacteria, which may increase risk of selection for antibiotic resistance through co-occurrence of multiple selective pressures and the genes/organisms upon which they may act. Concerns have long been expressed about the contribution of metals to the proliferation of antibiotic resistance via indirect selection pressures9, particularly as their stability may lead to long term selective pressures within a large temporal window13,45,46.
Numerous studies have investigated the relationship between antibiotic and metal resistance, and the topic is well reviewed elsewhere9,26,27. Studies that support co-selection for antibiotic resistance by metals have been conducted experimentally, in vitro using single species and complex communities of bacteria, and in situ in various environmental compartments including sewage sludge, aquaculture, and agricultural soils9,26,27. There is also evidence of co-selection for antibiotic resistance at sub-inhibitory concentrations of metals47,48,49,50,51.
It is well established that antibiotic and metal resistance genes can co-occur within bacterial genomes and this has been explored in genomic studies which use publicly available, fully sequenced genome data from hundreds of bacterial species10,40. These co-occurrence patterns can arise as a result of the resistance genes simply co-existing within the bacterial cell (e.g., the metal resistance gene present on the chromosome and the antibiotic resistance gene on a plasmid), which one study found to be the most common route of co-occurrence10. Perhaps more alarmingly, although less common, the metal and antibiotic resistance genes can co-occur on mobile genetic elements such as plasmids10,40, and would lead to co-resistance under selective pressure by either agent but also potential HGT of both types of resistance. An example of this is bacitracin resistance genes, which can co-occur on plasmids with copper and zinc resistance genes40. Interestingly, plasmids that harbour metal resistance genes have been shown to be more likely to also contain ARGs than plasmids which do not carry metal resistance genes, and these plasmids were also more likely to be conjugative10. These genomic studies have shown several ARG classes are more likely to be associated with metal resistance genes, including beta-lactam, kasugamycin, bacitracin, aminoglycoside, polymyxin, and tetracycline resistance genes40. The most common metal or antibiotic resistance genes detected in these genomic based studies are often related to those agents which are most frequently used40, likely owed to a greater selection pressure and therefore a greater need for bacteria to evolve, maintain and disseminate resistance genes.
The evidence for co-selection for antibiotic resistance by metals is vast, however, less is known about how this translates to selection for antibiotic resistance in the environment, and further research is required to understand the selective windows across which metals can act. This is of particular concern since metals are highly persistent in the environment, suggesting that there could be a much larger temporal co-selective window than for other non-antibiotic agents.
Biocides
Biocides are defined by the European Commission as ‘any substance or mixture’ ‘with the intention of destroying, deterring, rendering harmless, preventing the action of, or otherwise exerting a controlling effect on, any harmful organism by any means other than mere physical or mechanical action’ (Regulation (EU) No 528/2012). This definition includes antibiotics, disinfectants, herbicides, pesticides, and other related compounds. For the purposes of this review, biocides used as disinfectants and detergents (such as quaternary ammonium compounds, ‘QACs’) are covered here, and biocides used in other circumstances are included elsewhere in this review.
The application concentration of biocides is many times higher than the MIC, and in-use concentrations are often thousands of times greater than the MIC, often in the g/L range52. However, they are also found in the environment at lower, sub-lethal concentrations (µg/L) following dilution or can be found as residues on many surfaces. These sub-lethal concentrations may select for antibiotic resistance.
There is mixed evidence regarding associations between antibiotic and biocide resistance. Several reviews have explored this topic, with some outlining increases in antibiotic resistance after exposure to disinfectants8,28,53, yet others have found no association between exposure to biocides and antibiotic resistance8,54,55. For example, one study demonstrated that there was no co-selection (tested by identifying MIC profiles for various antibiotics and disinfectants) between chlorhexidine and the antibiotics tested, in over one hundred Salmonella spp. isolates56. Whereas a later study showed that when Klebsiella spp. were exposed to chlorhexidine, resistance to colistin was co-selected for (with the increased expression of AMR genes, and a multi-drug efflux pump)57.
However, like metal resistance genes, biocide resistance genes are more likely to be genetically associated with ARGs, and are therefore likely to be co-selected for via co-resistance. Analysis of plasmid data from publicly available genome data has suggested that bacteria carrying plasmids with biocide resistance genes are more likely to also carry ARGs than those not carrying these plasmids10. QAC resistance genes have also been found on plasmids containing metal, beta-lactamase, trimethoprim, or aminoglycoside resistance genes, and are also found in conserved regions of integrons in both Gram-positive and Gram-negative bacteria8. The gene qacEΔ1 which confers resistance to several biocides, including QACs10, is found on class 1 integrons58 and Tn21 transposons59. These genetic data suggest that biocides, including QACs, have the potential to co-select for antibiotic resistance via co-resistance. Furthermore, the mobilisation potential of these genes provides an increased likelihood of this resistance passing between environmental and clinically relevant bacterial strains.
Genetic linkage of biocide resistance genes and ARGs does not confirm that co-selection has occurred. However, experiments in vitro have also demonstrated co-selection for antibiotic resistance by biocides in both single species experiments, and in bacterial communities. For example, an aquatic bacterial community exposed to benzalkonium chloride increased resistance to ciprofloxacin and penicillin G, through cross-resistance60. Exposure of Escherichia coli to 0.2 mg/L triclosan for 30 days increased mutation frequency, and antibiotic resistance, compared to non-exposed cells61. Exposure of E. coli to sodium chlorite and iodoacetic acid for 40 subculture cycles of 40 days increased resistance to amoxicillin and ciprofloxacin to ‘clinically relevant’ levels via upregulation of multidrug resistance efflux pumps associated with antibiotic resistance61,62. These demonstrate that co-selective effects of biocides can be detected both in single species, and in community studies.
However, other studies show biocides may not always co-select for antibiotic resistance. When a wastewater influent bacterial community was exposed to 8 mg/L benzalkonium chloride, there were no identifiable increases in ARG prevalence, but there were changes to the community composition according to reads-based metagenome analyses63. This may suggest that only intrinsically resistant species were enriched, that changes in regulation of genes involved in antibiotic resistance occurred which could not be detected with the approach used, or alternatively, that any genes selected for were not present in gene databases used for these analyses, or that there was simply no co-selective effect in this model community. The databases used to identify ARGs focus on clinically important strains and will therefore lack resistance genes that confer low-level resistance, or genes more prevalent in environmental strains. A recent functional metagenomic study illustrated that cross-resistance to biocides and antibiotics can be mediated by metabolic genes64, which would not be present on commonly used ARG databases.
In summary, in certain circumstances, exposure to biocides may increase resistance to some antibiotics, but it is not clear to what extent these effects may be consistent across all environments, whether these genes are stably heritable, and whether they confer high levels of resistance that would be of clinical concern. Additionally, different environments can contain large concentration ranges of biocides, often alongside antibiotics, and the consequences of co-occurrence are largely unknown. Further research is needed to understand these effects, for example, by testing a large concentration range (from point of use to residual and micropollutant concentrations). Additionally, longer-term experimental studies looking at both the selection for de novo mutants, and the sustained selection of these over time, are required. This is considered in more detail in the Discussion.
Non-antibiotic drugs (NADs)
NADs, for the purposes of this review, refer to any pharmaceutical used in human and veterinary medicine that is not an antibiotic. These compounds can be found at various concentrations in the human body and are also found at low concentrations in the aquatic environment65. There is increasing interest in some NADs as antibiotic adjuvants66,67. However, there is also increasing interest in their potential to co-select for antibiotic resistance.
There are several studies investigating reduction in growth by NADs with single species models68,69,70,71,72. One of these studies71 indicated that the AcrAB-TolC efflux pump was important in allowing the growth of E. coli when under stress of the NADs tested, since knockout mutants had inhibited growth compared to wildtype. Reduction in growth with antibiotics can be a good proxy for identifying MSCs for antibiotic resistance63,73. However, one study showed that whilst some chemotherapeutic agents reduced the growth of E. coli and Staphylococcus aureus, there was no corresponding increase in MIC, suggesting that there was no co-selective benefit74. Therefore, although reduction in growth could indicate co-selective effects, this cannot be confirmed without further experimental work.
Several NADs have also been shown to increase HGT rates between bacteria. Ibuprofen, diclofenac, naproxen, propranolol, and gemfibrozil can increase the rate of transformation (the acquisition of free DNA from the environment into bacterial cells) of a plasmid bearing ARGs in Acinetobacter baylyi75. Carbamazepine and acetaminophen, as well as some antidepressants, can increase conjugative transfer in E. coli76,77 which may allow for acquisition of plasmids bearing ARGs and genes involved in NAD resistance, or plasmids containing ARGs alone. Carbamazepine was shown to increase conjugation rates at environmental (0.05 mg/L) concentrations, raising concerns that there could be a large spatial window over which co-selection by this compound could occur78. Therefore, some NADs may promote the spread of AMR, even if they do not co-select for antibiotic resistance themselves. However, this may be pharmaceutical specific, since some anti-HIV drugs reduced the rate of ARG transfer through conjugation in E. coli and Klebsiella pneumoniae79.
Additionally, some studies have investigated phenotypic resistance to antibiotics after exposure to NADs. One study showed that E. coli exposed to 5–100 mg/L of the antidepressant fluoxetine for 30 days exhibited increased antibiotic resistance to chloramphenicol, amoxicillin, and tetracycline, which may have been mediated by increased upregulation of multidrug resistant efflux pumps80. Fluoxetine also increased the mutation rate in E. coli, via ROS-mediated mutagenesis80. The chemotherapeutic drug methotrexate also co-selected for trimethoprim resistance at concentrations 40 to 320 fold lower than its MIC, with MSCs experimentally determined for chromosomal (200 µg/mL) and plasmid borne (25 µg/mL) trimethoprim resistance using tagged isogenic E. coli strains81. Similarly, exposure to eight different chemotherapeutic drugs (daunorubicin, epirubicin, mitomycin C, gemcitabine, bleomycin, dacarbazine, and azacitidine) at therapeutic concentrations increased mutation rates in E. coli through increased activation of the SOS response82. Additionally, this same study found that three chemotherapeutic drugs selected for imipenem resistant Pseudomonas aeruginosa, ciprofloxacin resistant S. aureus, and cefotaxime resistant Enterobacteria cloacae82. The non-steroidal anti-inflammatory drug diclofenac at 80 µg/L has been shown to lead to changes in gene expression, leading to increases, and also decreases in phenotypic resistance to antibiotics in S. aureus83. Diclofenac led to increased resistance to oxacillin and vancomycin, decreased resistance for ciprofloxacin, orfloxacin, and norfloxacin, but had no effect on susceptibility to tetracycline or chloramphenicol, indicating that collateral sensitivity may limit extensive multi-drug resistance acquisition after exposure to diclofenac, or other NADs83. The atypical antipsychotic medication quetiapine, at concentrations likely to occur in the human gut, has been shown to increase the expression of marA, acrA, and tolC, and can reduce ompF expression, also increasing antibiotic resistance in E. coli84. Finally, exposure to the antidepressants amitriptyline, fluoxetine, and sertraline can increase phenotypic antibiotic resistance in Acinetobacter baumanni85.
A large co-selective effect of NADs may be through cross-resistance, through increased expression of efflux pumps. For example, diclofenac has been shown to be a substrate of the multi-drug efflux pump AcrAB-TolC86. Although not tested, exposure to these NADs may increase expression of AcrAB-TolC and increase resistance to antibiotics. This could be a target for future research.
These studies (and most studies testing NADs) often use concentrations that are much higher than those present in the environment. However, concentrations of NADs present in hospital and municipal wastewater treatment plants have also been shown to increase ciprofloxacin resistance in Salmonella enterica serotype Typhimurium by increasing mutation rates87. Therefore, there is some evidence to suggest that effects seen at high concentrations may also occur at lower concentrations, suggesting selection could occur in wastewater, or potentially freshwater. However, there is some conflicting evidence, with a recent publication88, indicating that 30 day exposure of E. coli to acetaminophen, ibuprofen, TiO2, metformin, and propranolol at environmentally relevant concentrations did not lead to increased phenotypic resistance to antibiotics, nor to increases in ARGs.
Finally, there is emerging research testing mixtures of antibiotics and NADs, although there are few studies specifically investigating this. One recent study has shown that E. coli exposed to duloxetine and chloramphenicol had a synergistic response compared to exposure to one of these agents alone, with the E. coli becoming resistant eight antibiotics, with increased expression of acrA, acrB and marA (genes involved in efflux pumps)30. Most research focuses on mixtures as adjuvants and use growth reduction as an endpoint to show efficacy for this application67,89,90,91. There is a large knowledge gap in how combinations of pharmaceuticals can select for antibiotic resistance, particularly at sub-MIC concentrations.
In summary, there is potential for NADs not only to have antibacterial effects, but to have mechanisms of activity (such as upregulation of efflux pumps, i.e., cross-resistance and/or co-regulation) that overlap those of antibiotics. Thus, any prolonged exposure to these drugs may lead to co-selection for ARGs, though some studies (e.g., 79) have shown this may not always be the case. As several of these compounds can increase rates of HGT, there is also the opportunity for an increase in ARG transfer within microbiomes. Research is needed on the effects of NADs on gene expression, such as whether they induce a specific expression profile, lead to co-regulation, or induce mutations that confer cross-resistance (e.g., through upregulation of the SOS pathway92). Additionally, since most research into NADs has focused on the human health effects (e.g., on the gut microbiome) and benefits of using NADs as adjuvants, further work is needed to investigate if NADs can also co-select for antibiotic resistance at a range of concentrations relevant to different contexts.
Plant protection products (PPPs)
PPPs, also referred to as pesticides, are any agent used in agriculture/horticulture that prevent or treat infection or infestation by unwanted organisms. They include herbicides, fungicides (some of which have the same mechanism of action as clinical fungicides93), and insecticides, which target weeds, fungi, and insects, respectively. They are biologically active ingredients which are applied in commercial formulations that contain a variety of other ingredients to increase the efficacy of the active chemical94.
Increasing human population growth has resulted in higher demands for food and as a result, increased reliance on PPPs for food security. This trend is likely to continue, yet in 2012, world usage of the chemicals at the producer level already totalled over 2.72 billion kg95. PPPs are applied at high concentrations (often >g/L of active ingredient as calculated from product labels), repeatedly, to crops and soils, exposing the bacteria in the environment to significant quantities of PPPs over large temporal windows. When conducting environmental risk assessments for these agricultural chemicals, limited testing is carried out on microorganisms. Environmental risk assessments only require testing of carbon and nitrogen turnover, at limited PPP concentrations, over a limited time period, and do not consider the potential impacts on community diversity and antibiotic resistance96,97. Due to the vast number of target organisms PPPs act upon, and PPP mechanisms of action, this section investigating co-selection by PPPs will be split according to the major groups: herbicides, fungicides, and insecticides.
Herbicides
A few studies have investigated co-selection for antibiotic resistance by herbicides. One study investigated the effects of herbicide exposure in soil microcosms using phenotypic testing, metagenomics and qPCR98. Selective pressures exerted by glyphosate, glufosinate, and dicamba elevated the relative abundance of a range of ARG classes and MGEs in the soil bacterial communities. These increases occurred irrespective of changes to the community, as only glufosinate had a significant effect on community composition. The effects on relative abundance of ARGs and MGEs were observed at agriculturally relevant concentrations of 10 mg/kg98, demonstrating increases in ARGs could potentially occur in the field during herbicide application. Furthermore, de novo mutations were found in genes linked to multiple antibiotic resistance mechanisms (e.g., efflux pumps, N-acetyltransferases), and conjugation frequency of a multidrug resistance plasmid increased with exposure to the herbicides, thought to be a result of increased cell membrane permeability and up-regulation of stress-related genes98. Glyphosate-induced increases in conjugation frequency were also observed in another study which investigated the transfer of a resistance plasmid between donor and recipient E. coli species, which was due to upregulation of genes involved in cell membrane permeability and conjugation99.
Isolates of P. aeruginosa showed increased phenotypic resistance to aztreonam but did not change in resistance to colistin or polymyxin B, after exposure to the widely used herbicide active ingredients atrazine and diuron100. Atrazine has also been shown to significantly increase phenotypic resistance to ciprofloxacin, kanamycin, and streptomycin (but not tetracycline) in single species broth microcosm experiments using E. coli101. The mechanisms of co-selection in these experiments were not identified. In similar experiments, sublethal concentrations of formulations ‘Roundup’ (active ingredient glyphosate), ‘Kamba’ (active ingredient dicamba) and ‘2,4-D’, also resulted in increased phenotypic resistance to ciprofloxacin in both Salmonella enterica sv. Typhimurium and E. coli102,103. However, they, also reported decreases in phenotypic resistance to other antibiotics (collateral sensitivity), along with variable changes in resistance to other antibiotics that were largely dependent on herbicide, bacterial species, and antibiotic102. Efflux was shown to play a role in the increased tolerance of E.coli to chloramphenicol and kanamycin in the presence of dicamba and glyphosate respectively102. Similar variable results were observed in another study by the group using the active ingredients of the herbicides104, and multidrug efflux pump AcrAB-TolC was shown to play a role in the changes in phenotypic resistance, suggesting cross-resistance could be involved104. The variation in co-selective potential observed in these studies may be due to class specific effects of these herbicides, and their modes of action, particularly on bacteria, which have not been comprehensively characterised as they are not the target organism.
Fungicides
Fungicide exposure has also been associated with an increased abundance of ARGs, observed using metagenomic sequencing. The active ingredients carbendazim, azoxystrobin, and chlorothalonil increased the abundance of ARGs including sul1, sul2, aadA, tetX2, and tet(L) in microcosm experiments containing greenhouse and mountain soil105. These genes are involved in antibiotic target replacement, drug inactivation, and efflux pump mechanisms106.
Exposure of a soil community to azoxystrobin at agriculturally relevant concentrations also showed increased community phenotypic antibiotic resistance to aminoglycosides (streptomycin, kanamycin, and gentamicin)107, and commercial formulations of azoxystrobin were shown to increase phenotypic resistance to ampicillin, and chloramphenicol, but have no effect on tetracycline resistance in two soil types (loamy sandy soil, and clay loam soil)108. An increase in resistance to streptomycin was also observed in the loamy sandy soil, but sensitivity to this antibiotic was observed in the clay loam soil, demonstrating how soil properties may impact response to the fungicide. Again, the mechanisms involved in these increases in resistance were not determined108.
In addition to these findings in soil communities, azoxystrobin across a range of concentrations (0.1–5 mg/kg), altered the gut microbiome of the soil-dwelling worm Enchytraeus crypticus and increased the relative abundance of unclassified ARGs, analysed by Illumina amplicon sequencing and high-throughput qPCR109. However, the authors concluded it was unclear whether the increase in ARGs was a result of the changing gut community, or due to direct selection for ARGs.
There is little research investigating changes in phenotypic resistance to antibiotics after, or with exposure to fungicides in single species experiments. However, there are studies investigating how fungicides may increase HGT of ARG-bearing plasmids110,111. The fungicide ‘mancozeb’ was shown to facilitate plasmid mediated ARG transfer in experiments exploring intra-species transfer (E. coli donor to E. coli recipient) and inter-species transfer (E. coli donor to Pseudomonas putida recipient), alongside altering the expression of conjugation and stress response genes110. In another study that used the same donor and recipient species, the fungicides azoxystrobin and carbendazim enhanced the expression of conjugation related genes, while chlorothalonil was able to enhance reactive oxygen species, activate the stress response and increase membrane permeability, all resulting in the transfer of ARG-containing plasmid RP4111.
Insecticides
To date, insecticides are the least researched of the major PPP groups with regards to co-selection for antibiotic resistance. However, some evidence for potential co-selection exists. Insecticide exposure (‘Pyrethrum’, formulation containing pyrethrin active ingredients) increased phenotypic resistance to streptomycin and ciprofloxacin in single species experiments with E. coli101. The organophosphate insecticide chlorpyrifos also increased ARGs and intI1 in bulk soil, but not rhizosphere soil, when tested in microcosms followed by qPCR112. The authors speculated that the difference between the soils could be due to reduced availability of the insecticide in the two soil types, or differences in bacterial species present, highlighting how bioavailability (mentioned previously with regards to metals) is also a concern for other non-antibiotic agents.
Furthermore, studies have shown correlations between tolerance to insecticides and antibiotic resistance in isolates obtained from contaminated fields113,114,115. Rangasamy et al., demonstrated that Bacillus strains isolated from soils contaminated with lindane, α-endosulfan, and β-endosulfan insecticides, displayed phenotypic resistance to ampicillin, cefotaxime, chloramphenicol, streptomycin, and tetracycline, however, there was no comparison to control isolates113. Additionally, removal of a plasmid with insecticide degrading properties from these Bacillus strains increased their susceptibility to the antibiotics, suggesting a link between insecticide degradation genes and antibiotic resistance or degradation113. Similarly, isolates obtained from fields contaminated with unspecified pesticides were found to be phenotypically tolerant to insecticides, fungicides, and antibiotics, with isolates often containing ampC, tetM, ermD, mecA, and ermG, genes115. However, these results are correlative, and environmental isolates in such studies are often exposed to a myriad of stressors that may increase resistance evolution and HGT, so it may not be the sole action of the insecticide that resulted in antibiotic resistance.
To conclude, research into co-selection for antibiotic resistance by PPPs is in its early stages, with many of the studies showing evidence of co-selective effects, yet others show varied results. In addition to herbicides, fungicides, and insecticides, there are other groups of PPPs including molluscicides, acaricides, rodenticides, and plant growth regulators. To our knowledge, there are no studies on these classes of PPPs and co-selection for antibiotic resistance.
Furthermore, the active ingredients of PPPs are often applied alongside other ingredients as part of a chemical formulation, which improve the efficacy of the active chemical (i.e., adjuvants, which act to increase surface area and have surface active properties). Therefore, large quantities of adjuvants, solvents, and other ingredients may also be applied to crops and soils and should be investigated for their co-selective potential. The 2017 study by Kurenbach et al., did investigate some common herbicide co-formulants (‘Pulse Penetrant carboxymethyl cellulose’ and ‘Tween80’) for their co-selective potential. They found varied effects on phenotypic resistance that were dependent on species, antibiotic of interest, and co-formulant104.
Knowledge gaps
There is clear evidence for co-selection by various non-antibiotic agents as summarised in this review (Table 1), and the potential for co-selection by others which have shown varied results, but there are still significant knowledge gaps, including data specifically using environmentally relevant concentrations, species, or microbial communities. Particularly, there is limited research on co-selection for antibiotic resistance by NADs and PPPs and key knowledge gaps are discussed below. A theme common to all the recommendations below, is that despite there being strong evidence for co-selection in many of the studies presented, the underlying co-selection mechanisms remain unclear. Further studies should include experiments that specifically confirm the mechanism of co-selection e.g., co-resistance, cross-resistance, co-regulation, or any combination of these, where possible. This may not always be feasible, if for example, a multidrug efflux pump (cross-resistance) was located on a plasmid with an antibiotic resistance gene (co-resistance), with the expression of both genes under control of the same promoter (co-regulation). No one mechanism could be ascribed to co-selection in this case. However, for simpler studies with less genetic complexity, the co-selection mechanism may be easier to identify. This would be beneficial to not only improve the robustness of results, but also to begin to discern the most common/concerning mechanism of co-selection, which could direct future studies.
Where does co-selection occur?
Although co-selection has been shown to occur in laboratory experiments, the extent and implications of co-selection are largely unknown. For example, in which bacterial microbiomes can co-selection occur? Further research should aim to identify these, so their study can be prioritised. This could be addressed by testing a wider range of matrices (e.g., sewage/river sediment/soil/surface water communities, in addition to laboratory strains).
Furthermore, it is important to consider the vast number of agents or mixtures already in use and consider the future development of new agents or mixtures. Testing of new agents is clearly recognised to be important but re-testing of agents previously not considered to have co-selective effects may also be necessary. For example, some NADs are now being recognised as having antibacterial effects, but may also have co-selective effects, despite being in clinical use for decades81,116. In addition, a better understanding of the usage of non-antibiotic agents would help to determine hotspots and mitigation strategies. This is easier for some non-antibiotic agents than others. Prescription records are particularly well developed for NADs, but over-the-counter use is less well monitored. Usage of other non-antibiotic agents such as biocides and PPPs face more challenges in monitoring. For example, pesticide active ingredients are present on their own, and in combination with other active ingredients in a wide range of commercial products, which are often easily purchased over the counter, making monitoring of use difficult. Many of the decisions as to how, when, and where the products will be used (within the advised restrictions) are made by the consumer. Only when we have a full understanding of where each agent acts as a co-selecting agent can we implement mitigation strategies.
At which concentrations does co-selection occur?
Following the identification of co-selective agents, it is important to investigate the concentration ranges capable of selecting for antibiotic resistance. Many non-antibiotic agents will be present in different settings at a range of concentrations, for example, as residues on surfaces, and in the environment as diluted contaminants. More studies are needed to understand potential selective hotspots, especially for compounds that are applied at high concentrations (such as biocides and PPPs). However, as detailed in the introduction, there has been increasing interest in selection for ARGs by antibiotics at sub-MIC concentrations, and a growing understanding that this increases the window of selection or persistence of resistance mechanisms in bacterial communities19,20. Worryingly, lower selective pressures are more likely to result in selection and persistence of resistance genes with a low fitness cost to the bacterium. This means they may persist in the absence of selection far longer than ARGs with higher fitness cost, which would only be selected for by higher concentrations of co-selective agents117,118,119. Some research has already demonstrated that certain non-antibiotic agents are capable of a similar effect, increasing resistance at sub-inhibitory concentrations47,49. There needs to be focused study of the co-selective windows for other non-antibiotic agents, as without data it is impossible to pinpoint where and when there could be fixation and dissemination of ARGs, and transfer to opportunistic pathogens resulting from co-selective agent contamination.
Appreciating chemical complexities
It is of paramount importance to appreciate chemical complexities, including their potency and behaviour that is driven by ever-changing natural environments. Experimental designs on co-selection studies tend to focus on individual chemicals at certain concentration levels. The assumption is made that these chemicals are constant in the context of dynamic changes in microbial communities, however, this is not the case. Chemicals are in a dynamic equilibrium between physical and biological environments. They also continuously partition between liquid and solid interfaces (e.g., suspended particulate matter or sediments) and this affects their bioavailability. In general, chemicals with lower polarity (higher hydrophobicity, usually described by logKow or logP constants) tend to be more persistent and bioaccumulate in biota and sediments. The potency and bioavailability of chemicals are also directly linked with their state-of-charge. Chemicals can be either non-charged (neutral, e.g., pure water), positively/negatively charged (e.g., negatively charged anti-inflammatory drugs or positively charged beta-blockers), or zwitterionic (e.g., quinolone antibiotics) in their surrounding environment, and their charge will be driven by changing pH. The pH of natural environments can change significantly from neutral to acidic which might be a result of natural processes as well as contamination, e.g., acid mine drainage. This change in pH might have an impact on chemicals in question. As an example: ibuprofen with pKa (acid-base constant) of 4.85 will be largely protonated in a heavily polluted environment (e.g., rivers affected by acid mine discharges), but it will deprotonate (to gain negative charge) in a neutral, healthy environment. This change of charge state has critical consequence to its bioavailability, potency, and transport across the cell wall. Large scale monitoring campaigns clearly show variable presence of chemicals that is driven not only by usage but also by their properties120, e.g., plant uptake is to a large extent driven by hydrophilicity and charge121. Chemicals, when they pass through human/animal bodies, are subject to metabolism that leads to the formation of metabolites. Phase II metabolites such as glucuronides, are relatively unstable in wastewater and are subject to microbial cleavage of parent compounds122. This is of key consequence to risk assessment of chemicals in the environment, as it is commonly assumed that metabolites are less potent than parent chemicals123. Lastly, many chemicals (including most antibiotics) have stereocenters (e.g., chiral centres) which enables them to exist in several configurational forms, each potentially triggering different biological responses124.
The complexities of chemicals, and chemical metabolites, is an important consideration in their potential co-selective effects, and the extent to these effects in different environments. For example, a wastewater treatment plant output to the freshwater system will be dispensing parent compound, metabolites, and several forms of these compounds into an environment that could be changeable in terms of pH, and the bacteria that are present there, so the effects of this input to the freshwater system may vary over season/time. Further work should try to understand and incorporate some of this complexity into experimental design to determine how important these chemical properties are in co-selection.
What are the co-selective effects in complex (microbial and chemical) contexts?
Most research published to date has focused on the incidence of co-selection at the single species level. However, bacteria exist in the environment within complex microbial communities. Previous research has shown that selection for antibiotic resistance by some aminoglycoside antibiotics is reduced when a bacterium is in a community, compared to when grown in isolation125. Similar effects could occur with exposure to non-antibiotics in complex communities, but this is yet to be researched. Additionally, when considering the effects of selective agents on microbial communities, research should consider that different microbiomes exist in different environments and may respond differently to the same selective pressure.
Similarly, these agents are rarely present in isolation. In environmental compartments e.g., wastewater and soils, antibiotics are found alongside metals, biocides, NADs, and other potentially selective agents. However, the research literature on the selective effects of antibiotic mixtures, and of antibiotics alongside non-antibiotic agents, is limited. MICs and MSCs are usually experimentally determined using single species models, and most often for single antibiotic stressors18,126,127. Therefore, it is likely that combined exposure risks (of multiple agents) are underappreciated30. Mixture effects can be both positive (additive or synergistic), or negative (antagonistic). There is some evidence that suggests particular combinations of agents, such as metals and antibiotics, or antibiotics alongside NADs, can increase antibiotic resistance. For example, it was shown that the combined exposure of E. coli to duloxetine (an antidepressant) and chloramphenicol increased the transcription of ARGs compared to exposure to each compound alone119. Interactions may be agent specific, for example, a tetracycline and arsenic mixture decreased the MSC of tetracycline47, whereas the MSC of ciprofloxacin increased in the presence of zinc128.
Not only are bacterial communities exposed to a mixture of agents, but those agents degrade at different rates, and into a variety of different metabolites. This results in the communities being exposed to various metabolites at variable concentrations, in addition to parent compounds, that could have co-selective effects129,130. Future research should aim to address this, by exposing bacterial communities present in ecological compartments (contaminated and pristine soil environments, the aquatic environment, marine, the gut, clinical environments, veterinary environments, wastewater etc.) to combinations of antibiotic and non-antibiotic agents and context-relevant metabolites.
Is there evidence of increased resistance to non-antibiotic agents, after exposure to antibiotics?
Research concerning increased resistance to antibiotics after exposure to non-antibiotics is increasing, but research into the reverse is lacking. However, there is some evidence that antibiotics can co-select for non-antibiotic agent resistance. After adaptation to erythromycin, E. coli demonstrated increased resistance to the biocide triclosan131. Conversely, adaptation to colistin has been shown to have no effect on cross-resistance to chlorhexidine57. Efflux pumps may be important in this process. For example, the AcrAB-TolC efflux pump can provide resistance to a range of antibiotics132 and remove biocides from within the cell28. It could be argued that after exposure to antibiotics, these broad range resistance mechanisms may also confer resistance to biocides. Despite this, there is a lack of understanding of this phenomenon, which might be a concern for hospitals and other clinical settings since disinfectant biocide use, and even NAD use, could lead to increased antibiotic resistant nosocomial infections. Future studies should test a range of concentrations of these agents and use bacterial communities that are representative of the various compartments that might be affected.
The problem of bias in gene databases
ARG databases contain genes mostly identified in clinical isolates after exposure to antibiotics. It is likely that non-antibiotic agents could be selecting for novel or as-yet unannotated resistance genes, which are therefore not detected in current studies. Novel genes could be of great concern, particularly if there is a high risk of transfer of these novel genes into human, animal, or plant pathogens. Methods such as functional metagenomics, and predictive methods (e.g., machine learning133), alongside confirmatory experimental work could be used to identify these novel genes134.
Challenges for risk management
Current environmental risk assessments for antibiotics do not consider the risk of selection for antibiotic resistance in the environment. The only tests on microorganisms focus on aerobic capacity and community functioning, and are tested on algae96, or on activated sludge communities97. Recent research calls for the inclusion of risk of selection for resistance126. Although the inclusion of this in environmental risk assessment would be a positive step towards risk mitigation, it is important to also consider the co-selective properties of non-antibiotic agents. Simply addressing the risk posed by antibiotics may not be sufficient to remove the selective pressure if co-selective agents are present in the environment. Mixture effects are also not routinely identified in terms of environmental risk, and therefore incorporating these into environmental risk assessment would be a step forward in protecting the environment in terms of selection, maintenance and dissemination of ARGs and resistant bacteria.
Conclusions
Due to the widespread presence of potentially co-selective agents, and in an ever increasingly connected world, reducing antibiotic usage to tackle antibiotic resistance may not have the desired effects if no mitigations are put in place to reduce co-selective agents. There are around 355,000 chemicals or mixtures of chemicals thought to be in use135. Anthropogenic production of these chemicals has led to some considering that we have “crossed a planetary boundary” for pollution by novel entities meaning that we are “outside of the safe operating space” for chemical pollution136. The number of chemicals and agents that have been tested for their co-selective effects is merely a fraction of this total number. Even among those that have been tested (e.g., PPPs), their co-selective potential clearly varies under different contexts and more research is required to fully understand the threat that these agents, and mixtures of agents may have in terms of AMR. Our understanding of the individual and mixture effects of agents and their behaviour in different environments is also poor, and so we may be underestimating their effects. The One Health perspective highlights the connectivity between different environmental compartments (e.g., between human and veterinary health, the soil, and the aquatic environment). When considering this, and the vast numbers of agents that are present globally, there is likely no area immune to the co-selective effects of agents that can select for antibiotic resistance.
References
WHO. Global Action Plan On Antimicrobial Resistance. (World Health Organisation, Geneva, Switzerland, 2015).
O’Neill, J. Antimicrobial resistance: tackling a crisis for the health and wealth of nations. The Review on Antimicrobial Resistance. (2014).
Zumla, A., Raviglione, M., Hafner, R. & von Reyn, C. F. Tuberculosis. N. Eng. J. Med. 368, 745–755 (2013).
Bratzler, D. W. et al. Clinical practice guidelines for antimicrobial prophylaxis in surgery. Am. J. Health Syst. Pharm. 70, 195–283 (2013).
Murray, C. J. L. et al. Global burden of bacterial antimicrobial resistance in 2019: a systematic analysis. Lancet 399, 629–655 (2022).
Gillings, M. R. Lateral gene transfer, bacterial genome evolution, and the Anthropocene. Annal. N. Y Acad. Sci. 1389, 20–36 (2017).
O’Neil, J. Tackling drug-resistant infections globally: final report and recommendations. (Review on Antimicrobial Resistance, London, 2016).
Chapman, J. S. Disinfectant resistance mechanisms, cross-resistance, and co-resistance. Int. Biodeterior. Biodegrad. 51, 271–276 (2003).
Baker-Austin, C., Wright, M. S., Stepanauskas, R. & McArthur, J. V. Co-selection of antibiotic and metal resistance. Trends Microbiol. 14, 176–182 (2006).
Pal, C., Bengtsson-Palme, J., Kristiansson, E. & Larsson, D. G. Co-occurrence of resistance genes to antibiotics, biocides and metals reveals novel insights into their co-selection potential. BMC Genom. 16, 964 (2015).
Alonso, A., Sánchez, P. & Martínez, J. L. Environmental selection of antibiotic resistance genes. Environ. Microbiol. 3, 1–9 (2001).
Forsberg, K. J. et al. The shared antibiotic resistome of soil bacteria and human pathogens. Science 337, 1107–1111 (2012).
Dickinson, A. W. et al. Heavy metal pollution and co-selection for antibiotic resistance: a microbial palaeontology approach. Environ. Int. 132, 105117 (2019).
Leonard, A. F. C. et al. Exposure to and colonisation by antibiotic-resistant E. coli in UK coastal water users: Environmental surveillance, exposure assessment, and epidemiological study (Beach Bum Survey). Environ. Int. 114, 326–333 (2018).
aus der Beek, T., Weber, F.-A., Bergmann, A., Grüttner, G. & Carius, A. Pharmaceuticals in the Environment. (2016).
Stephanie Graumnitz, D. J. The database “Pharmaceuticals in the Environment” Update for the period 2017-2020. (2021).
Wilkinson, J. L. et al. Pharmaceutical pollution of the world’s rivers. Proc. Natl. Acad. Sci. 119, e2113947119 (2022).
EUCAST. Methods for the determination of susceptibility of bacteria to antimicrobial agents. Terminology. Report No. 1198-743x, 291 (1998).
Gullberg, E. et al. Selection of resistant bacteria at very low antibiotic concentrations. PLoS Pathog 7, e1002158 (2011).
Stanton, I. C., Murray, A. K., Zhang, L., Snape, J. & Gaze, W. H. Evolution of antibiotic resistance at low antibiotic concentrations including selection below the minimal selective concentration. Commun. Biol. 3, 467 (2020).
Bielen, A. et al. Negative environmental impacts of antibiotic-contaminated effluents from pharmaceutical industries. Water Res. 126, 79–87 (2017).
Milaković, M. et al. Pollution from azithromycin-manufacturing promotes macrolide-resistance gene propagation and induces spatial and seasonal bacterial community shifts in receiving river sediments. Environ. Int. 123, 501–511 (2019).
Chen, L., Todd, R., Kiehlbauch, J., Walters, M. & Kallen, A. Notes from the Field: Pan-Resistant New Delhi Metallo-Beta-Lactamase-Producing Klebsiella pneumoniae. 2022. https://doi.org/10.15585/mmwr.mm6601a7external (2016)
Hayes, A. et al. Predicting selection for antimicrobial resistance in UK wastewater and aquatic environments: Ciprofloxacin poses a significant risk. Environ. Int. 169, 107488 (2022).
Gaze, W. H. et al. Impacts of anthropogenic activity on the ecology of class 1 integrons and integron-associated genes in the environment. ISME J. 5, 1253–1261 (2011).
Pal, C. et al. in Advances in Microbial Physiology Vol. 70 (ed Poole R. K.) 261-313 (Academic Press, 2017).
Vats, P., Kaur, U. J. & Rishi, P. Heavy metal-induced selection and proliferation of antibiotic resistance: A review. J. Appl. Microbio. n/a. https://doi.org/10.1111/jam.15492
Buffet-Bataillon, S., Tattevin, P., Bonnaure-Mallet, M. & Jolivet-Gougeon, A. Emergence of resistance to antibacterial agents: the role of quaternary ammonium compounds—a critical review. Int. J. Antimicrob. Agents 39, 381–389 (2012).
Basiry, D. et al. The effect of disinfectants and antiseptics on co- and cross-selection of resistance to antibiotics in aquatic environments and wastewater treatment plants. Front. Microbiol. 13, 1050558 (2022).
Shi, D. et al. Combined exposure to non-antibiotic pharmaceutics and antibiotics in the gut synergistically promote the development of multi-drug-resistance in Escherichia coli. Gut Microbes 14, 2018901 (2022).
Jiang, Q., Feng, M., Ye, C. & Yu, X. Effects and relevant mechanisms of non-antibiotic factors on the horizontal transfer of antibiotic resistance genes in water environments: a review. Sci. Total Environ. 806, 150568 (2022).
Qiu, D. et al. Response of microbial antibiotic resistance to pesticides: an emerging health threat. Sci. Total Environ. 850, 158057 (2022).
Briffa, J., Sinagra, E. & Blundell, R. Heavy metal pollution in the environment and their toxicological effects on humans. Heliyon 6, e04691 (2020).
Schneider, G. et al. Textiles impregnated with antimicrobial substances in healthcare services: systematic review. Front. Public Health 11. https://doi.org/10.3389/fpubh.2023.1130829
Boxall, A. B. A., Comber, S. D., Conrad, A. U., Howcroft, J. & Zaman, N. Inputs, monitoring and fate modelling of antifouling biocides in UK Estuaries. Marine Pollut. Bull. 40, 898–905 (2000).
Food Standards Agency. Feed additive authorisations, https://data.food.gov.uk/regulated-products/feed_authorisations (2022).
Lewis, K. A., Tzilivakis, J., Warner, D. & Green, A. An international database for pesticide risk assessments and management. Human and ecological risk assessment. Int. J. 22, 1050–1064 (2016).
Murdoch, C. C. & Skaar, E. P. Nutritional immunity: the battle for nutrient metals at the host-pathogen interface. Nat. Rev. Microbiol. 20, 657–670 (2022).
Lemire, J. A., Harrison, J. J. & Turner, R. J. Antimicrobial activity of metals: mechanisms, molecular targets and applications. Nat. Rev. Microbiol. 11, 371–384 (2013).
Li, L. G., Xia, Y. & Zhang, T. Co-occurrence of antibiotic and metal resistance genes revealed in complete genome collection. ISME J. 11, 651–662 (2017).
Adhikary, A., Biswal, S., Chatterjee, D. & Ghosh, A. S. A NiCoT family metal transporter of Mycobacterium tuberculosis (Rv2856/NicT) behaves as a drug efflux pump that facilitates cross-resistance to antibiotics. Microbiology 168. https://doi.org/10.1099/mic.0.001260 (2022)
Lee, L. J., Barrett, J. A. & Poole, R. K. Genome-wide transcriptional response of chemostat-cultured Escherichia coli to zinc. J. Bacteriol. 187, 1124–1134 (2005).
Dean, J. R. in Bioremediation: Methods and Protocols (ed Cummings, S. P.) 15-36 (Humana Press, 2010).
Zou, H.-Y. et al. Antibiotic resistance genes in surface water and groundwater from mining affected environments. Sci. Total Environ. 772, 145516 (2021).
Vhahangwele, M. & Khathutshelo, L. M. in Heavy Metals (ed Hosam El-Din M. S. and Refaat F. A.) (IntechOpen, 2018).
Wuana, R. A. & Okieimen, F. E. Heavy metals in contaminated soils: a review of sources, chemistry, risks and best available strategies for remediation. ISRN Ecol. 2011, 1–20 (2011).
Gullberg, E., Albrecht, L. M., Karlsson, C., Sandegren, L. & Andersson, D. I. Selection of a multidrug resistance plasmid by sublethal levels of antibiotics and heavy metals. mBio 5, e01918–01914 (2014).
Chen, S. et al. Heavy metal induced antibiotic resistance in bacterium LSJC7. Int. J. Mol. Sci. 16, 23390–23404 (2015).
Li, X. et al. Sub-lethal concentrations of heavy metals induce antibiotic resistance via mutagenesis. J. Hazard. Mater. 369, 9–16 (2019).
Li, J., Phulpoto, I. A., Zhang, G. & Yu, Z. Acceleration of emergence of E. coli antibiotic resistance in a simulated sublethal concentration of copper and tetracycline co-contaminated environment. AMB Express 11, 14 (2021).
Zhang, Y. et al. Sub-inhibitory concentrations of heavy metals facilitate the horizontal transfer of plasmid-mediated antibiotic resistance genes in water environment. Environ. Pollut. 237, 74–82 (2018).
Maillard, J.-Y. & Pascoe, M. Disinfectants and antiseptics: mechanisms of action and resistance. Nat. Rev. Microbiol. https://doi.org/10.1038/s41579-023-00958-3 (2023)
Kampf, G. Acquired resistance to chlorhexidine – is it time to establish an ‘antiseptic stewardship’ initiative? J. Hospital Infect. 94, 213–227 (2016).
Wales, A. D. & Davies, R. H. Co-Selection of resistance to antibiotics, biocides and heavy metals, and its relevance to foodborne pathogens. Antibiotics 4, 567–604 (2015).
Cheng, G. et al. Selection and dissemination of antimicrobial resistance in Agri-food production. Antimicrob. Resist. Infect. Control 8, 158 (2019).
Beier, R. C. et al. Characterization of Salmonella enterica isolates from turkeys in commercial processing plants for resistance to antibiotics, disinfectants, and a growth promoter. Foodborne Pathog. Dis. 8, 593–600 (2011).
Wand, M. E., Bock, L. J., Bonney, L. C. & Sutton, J. M. Mechanisms of increased resistance to chlorhexidine and cross-resistance to colistin following exposure of klebsiella pneumoniae clinical isolates to chlorhexidine. Antimicrob. Agents Chemother. 61. https://doi.org/10.1128/aac.01162-16 (2017)
Gillings, M. et al. The evolution of class 1 integrons and the rise of antibiotic resistance. J. Bacteriol. 190, 5095–5100 (2008).
Liebert, C. A., Hall, R. M. & Summers, A. O. Transposon Tn21, flagship of the floating genome. Microbiol. Mol. Biol. Rev. 63, 507–522 (1999).
Tandukar, M., Oh, S., Tezel, U., Konstantinidis, K. T. & Pavlostathis, S. G. Long-term exposure to benzalkonium chloride disinfectants results in change of microbial community structure and increased antimicrobial resistance. Environ. Sci. Technol. 47, 9730–9738 (2013).
Lu, J. et al. Non-antibiotic antimicrobial triclosan induces multiple antibiotic resistance through genetic mutation. Environ. Int. 118, 257–265 (2018).
Li, D., Zeng, S., He, M. & Gu, A. Z. Water disinfection byproducts induce antibiotic resistance-role of environmental pollutants in resistance phenomena. Environ. Sci. Technol. 50, 3193–3201 (2016).
Murray, A. K., Zhang, L., Snape, J. & Gaze, W. H. Comparing the selective and co-selective effects of different antimicrobials in bacterial communities. Int. J. Antimicrob. Agents 53, 767–773 (2019).
Murray, A. K., Zhang, L., Snape, J. & Gaze, W. H. Functional metagenomic libraries generated from anthropogenically impacted environments reveal importance of metabolic genes in biocide and antibiotic resistance. Curr. Res. Microb. Sci. 4, 100184 (2023).
Aus der Beek, T. et al. Pharmaceuticals in the environment–Global occurrences and perspectives. Environ. Toxicol. Chem. 35, 823–835 (2016).
Ashburn, T. T. & Thor, K. B. Drug repositioning: identifying and developing new uses for existing drugs. Nat. Rev. Drug Discov. 3, 673–683 (2004).
Ejim, L. et al. Combinations of antibiotics and nonantibiotic drugs enhance antimicrobial efficacy. Nat. Chem. Biol. 7, 348–350 (2011).
Alsterholm, M., Karami, N. & Faergemann, J. Antimicrobial activity of topical skin pharmaceuticals - an in vitro study. Acta Derm Venereol 90, 239–245 (2010).
Sud, I. J. & Feingold, D. S. Action of antifungal imidazoles on Staphylococcus aureus. Antimicrob. Agent. Chemother. 22, 470–474 (1982).
Li, L. et al. RapidAIM: a culture- and metaproteomics-based Rapid Assay of Individual Microbiome responses to drugs. Microbiome 8, 33 (2020).
Maier, L. et al. Extensive impact of non-antibiotic drugs on human gut bacteria. Nature 555, 623–628 (2018).
Younis, W., Thangamani, S. & Seleem, M. N. Repurposing non-antimicrobial drugs and clinical molecules to treat bacterial infections. Curr. Pharm. Des. 21, 4106–4111 (2015).
Greenfield, B. K. et al. Modeling the emergence of antibiotic resistance in the environment: an analytical solution for the minimum selection concentration. Antimicrob. Agents Chemother. 62. https://doi.org/10.1128/aac.01686-17 (2018)
Campbell, O., Gagnon, J. & Rubin, J. E. Antibacterial activity of chemotherapeutic drugs against Escherichia coli and Staphylococcus pseudintermedius. Lett. Appl. Microbiol. 69, 353–357 (2019).
Wang, Y. et al. Non-antibiotic pharmaceuticals enhance the transmission of exogenous antibiotic resistance genes through bacterial transformation. ISME J. https://doi.org/10.1038/s41396-020-0679-2 (2020)
Jia, Y. et al. Acetaminophen promotes horizontal transfer of plasmid-borne multiple antibiotic resistance genes. Sci. Total Environ. 782, 146916 (2021).
Ding, P., Lu, J., Wang, Y., Schembri, M. A. & Guo, J. Antidepressants promote the spread of antibiotic resistance via horizontally conjugative gene transfer. Environ. Microbiol. n/a. https://doi.org/10.1111/1462-2920.16165 (2022)
Wang, Y. et al. Antiepileptic drug carbamazepine promotes horizontal transfer of plasmid-borne multi-antibiotic resistance genes within and across bacterial genera. ISME J. 13, 509–522 (2019).
Buckner, M. M. C. et al. HIV drugs inhibit transfer of plasmids carrying extended-spectrum β-Lactamase and carbapenemase genes. mBio 11, e03355–03319 (2020).
Jin, M. et al. Antidepressant fluoxetine induces multiple antibiotics resistance in Escherichia coli via ROS-mediated mutagenesis. Environ. Int. 120, 421–430 (2018).
Guðmundsdóttir, J. S. et al. The chemotherapeutic drug methotrexate selects for antibiotic resistance. eBioMedicine 74. https://doi.org/10.1016/j.ebiom.2021.103742 (2021)
Meunier, A. et al. Enhanced emergence of antibiotic-resistant pathogenic bacteria after in vitro induction with cancer chemotherapy drugs. J. Antimicrob. Chemother. 74, 1572–1577 (2019).
Riordan, J. T. et al. Alterations in the transcriptome and antibiotic susceptibility of Staphylococcus aureus grown in the presence of diclofenac. Ann. Clin. Microbiol. Antimicrob. 10, 30 (2011).
Kyono, Y. et al. The atypical antipsychotic quetiapine promotes multiple antibiotic resistance in Escherichia coli. J. Bacteriol. 204, e0010222 (2022).
Gurpinar, S. S., Kart, D. & Eryilmaz, M. The effects of antidepressants fluoxetine, sertraline, and amitriptyline on the development of antibiotic resistance in Acinetobacter baumannii. Archiv. Microbiol. 204, 230 (2022).
Laudy, A. E., Mrowka, A., Krajewska, J. & Tyski, S. The influence of efflux pump inhibitors on the activity of non-antibiotic NSAIDS against gram-negative rods. PLoS One 11, e0147131 (2016).
Bírošová, L., Lépesová, K., Grabic, R. & Mackuľak, T. Non-antimicrobial pharmaceuticals can affect the development of antibiotic resistance in hospital wastewater. Environ. Sci. Pollut. Res. Int. 27, 13501–13511 (2020).
Hall, R. J. et al. Non-antibiotic pharmaceuticals exhibit toxicity against Escherichia coli at environmentally relevant concentrations with no evolution of cross-resistance to antibiotics. bioRxiv. https://doi.org/10.1101/2023.08.21.554069 (2023).
Masadeh, M. M., Alzoubi, K. H., Masadeh, M. M. & Aburashed, Z. O. Metformin as a potential adjuvant antimicrobial agent against multidrug resistant bacteria. Clin. Pharmacol. 13, 83–90 (2021).
Chan, E. W. L., Yee, Z. Y., Raja, I. & Yap, J. K. Y. Synergistic effect of non-steroidal anti-inflammatory drugs (NSAIDs) on antibacterial activity of cefuroxime and chloramphenicol against methicillin-resistant Staphylococcus aureus. J. Glob. Antimicrob. Resist. 10, 70–74 (2017).
Gil, D. et al. Synergistic antibacterial effects of analgesics and antibiotics against Staphylococcus aureus. Diagn. Microbiol. Infect. Dis. 96, 114967 (2020).
Crane, J. K., Alvarado, C. L. & Sutton, M. D. Role of the SOS response in the generation of antibiotic resistance in vivo. Antimicrob. Agents Chemother. 65, e0001321 (2021).
Lamberth, C. Latest research trends in agrochemical fungicides: any learnings for pharmaceutical antifungals? ACS Med. Chem. Lett. 13, 895–903 (2022).
DEFRA, The Official Controls (Plant Protection Products) Regulations 2020: policy statement, F.a.R.A. Department for Environment (2020).
Atwood, D. & Paisley-Jones, C. Pesticides Industry Sales and Usage, 2008-2012 Market Estimates., (U.S. Environmental Protection Agency, Washington DC, 2017).
OECD Test No. 201: Freshwater Alga and Cyanobacteria, Growth Inhibition Test, OECD Guidelines for the Testing of Chemicals, Section 2, OECD Publishing, Paris, https://doi.org/10.1787/9789264069923-en (2011).
OECD Test No. 209: Activated Sludge, Respiration Inhibition Test (Carbon and Ammonium Oxidation), OECD Guidelines for the Testing of Chemicals, Section 2, OECD Publishing, Paris, https://doi.org/10.1787/9789264070080-en (2010).
Liao, H. et al. Herbicide selection promotes antibiotic resistance in soil microbiomes. Mol. Biol. Evol. 38, 2337–2350 (2021).
Zhang, H., Liu, J., Wang, L. & Zhai, Z. Glyphosate escalates horizontal transfer of conjugative plasmid harboring antibiotic resistance genes. Bioengineered 12, 63–69 (2021).
Braz, V. S., Moretto, J. A. S., Fernandes, A. F. T. & Stehling, E. G. Change in the antimicrobial resistance profile of Pseudomonas aeruginosa from soil after exposure to herbicides. J. Environ. Sci. Health B 54, 290–293 (2019).
Jun, H. et al. Effects of sub-lethal concentrations of copper ammonium acetate, pyrethrins and atrazine on the response of Escherichia coli to antibiotics. F1000Research 8, 32–32 (2019).
Kurenbach, B. et al. Sublethal exposure to commercial formulations of the herbicides dicamba, 2,4-dichlorophenoxyacetic acid, and glyphosate cause changes in antibiotic susceptibility in Escherichia coli and Salmonella enterica serovar Typhimurium. mBio 6. https://doi.org/10.1128/mBio.00009-15 (2015)
Kurenbach, B., Hill, A., Godsoe, W., Hamelsveld, S. & Heinemann, J. Agrichemicals and antibiotics in combination increase antibiotic resistance evolution. PeerJ 6, e5801 (2018).
Kurenbach, B. et al. Herbicide ingredients change Salmonella enterica sv. Typhimurium and Escherichia coli antibiotic responses. Microbiology 163, 1791–1801 (2017).
Zhang, H. et al. Fungicides enhanced the abundance of antibiotic resistance genes in greenhouse soil. Environ. Pollut. 259, 113877 (2020).
Alcock, B. P. et al. CARD 2020: antibiotic resistome surveillance with the comprehensive antibiotic resistance database. Nucleic Acids Res. 48, D517–d525 (2020).
Aleksova, M., Kenarova, A., Boteva, S. & Radeva, G. Azoxystrobin Impact on a Selection of Soil Bacterial Resistance to Amynoglicoside Antibiotics. Comptes rendus de l’Académie bulgare des Sciences 72, 1359–1365 (2019a).
Aleksova, M., Kenarova, A., Boteva, S. & Radeva, G. Induced Bacterial Antibiotic Resistance under the Application of Fungicide Azoxystrobin. Comptesrendus de l’Académie bulgare des Sciences 72, 1592 (2019b).
Zhang, Q. et al. The fungicide azoxystrobin perturbs the gut microbiota community and enriches antibiotic resistance genes in Enchytraeus crypticus. Environ. Int. 131, 104965 (2019).
Song, J. et al. A novel bidirectional regulation mechanism of mancozeb on the dissemination of antibiotic resistance. J. Hazardous Mater. 455, 131559 (2023).
Zhang, H. et al. Fungicide exposure accelerated horizontal transfer of antibiotic resistance genes via plasmid-mediated conjugation. Water Res. 233, 119789 (2023).
Guo, A., Pan, C., Ma, J. & Bao, Y. Linkage of antibiotic resistance genes, associated bacteria communities and metabolites in the wheat rhizosphere from chlorpyrifos-contaminated soil. Sci. Total Environ. 741, 140457 (2020).
Rangasamy, K., Athiappan, M., Devarajan, N. & Parray, J. A. Emergence of multi drug resistance among soil bacteria exposing to insecticides. Microb. Pathog. 105, 153–165 (2017).
Nawab, A., Aleem, A. & Malik, A. Determination of organochlorine pesticides in agricultural soil with special reference to γ-HCH degradation by Pseudomonas strains. Bioresour. Technol. 88, 41–46 (2003).
Anjum, R., Grohmann, E. & Malik, A. Molecular characterization of conjugative plasmids in pesticide tolerant and multi-resistant bacterial isolates from contaminated alluvial soil. Chemosphere 84, 175–181 (2011).
Wei, Z. et al. Emerging pollutant metformin in water promotes the development of multiple-antibiotic resistance in Escherichia coli via chromosome mutagenesis. J. Hazard. Mater. 430, 128474 (2022).
Andersson, D. I. & Hughes, D. Antibiotic resistance and its cost: is it possible to reverse resistance? Nat. Rev. Microbiol. 8, 260–271 (2010).
Andersson, D. I. & Hughes, D. Evolution of antibiotic resistance at non-lethal drug concentrations. Drug Resis. Updates 15, 162–172 (2012).
Andersson, D. I. & Hughes, D. Microbiological effects of sublethal levels of antibiotics. Nat. Rev. Microbiol. 12, 465–478 (2014).
Proctor, K. et al. Micropollutant fluxes in urban environment – A catchment perspective. J. Hazard. Mater. 401, 123745 (2021).
Petrie, B. et al. Biotic phase micropollutant distribution in horizontal sub-surface flow constructed wetlands. Sci. Total Environ. 630, 648–657 (2018).
Kasprzyk-Hordern, B. et al. Estimation of community-wide multi-chemical exposure via water-based chemical mining: Key research gaps drawn from a comprehensive multi-biomarker multi-city dataset. Environ. Int. 147, 106331 (2021).
Cwiertny, D. M., Snyder, S. A., Schlenk, D. & Kolodziej, E. P. Environmental designer drugs: when transformation may not eliminate risk. Environ. Sci. Technol. 48, 11737–11745 (2014).
Elder, F. C. T., Feil, E. J., Snape, J., Gaze, W. H. & Kasprzyk-Hordern, B. The role of stereochemistry of antibiotic agents in the development of antibiotic resistance in the environment. Environ. Int. 139, 105681 (2020).
Klumper, U. et al. Selection for antimicrobial resistance is reduced when embedded in a natural microbial community. ISME J. 13, 2927–2937 (2019).
Murray, A. K., Stanton, I., Gaze, W. H. & Snape, J. Dawning of a new ERA: environmental risk assessment of antibiotics and their potential to select for antimicrobial resistance. Water Res. 200, 117233 (2021).
WHO. Antimicrobial Resistance Global Report on Surveillance. (2014).
Vos, M. et al. Zinc can counteract selection for ciprofloxacin resistance. FEMS Microbiol. Lett. 367. https://doi.org/10.1093/femsle/fnaa038 (2020)
Holton, E. & Kasprzyk-Hordern, B. Multiresidue antibiotic-metabolite quantification method using ultra-performance liquid chromatography coupled with tandem mass spectrometry for environmental and public exposure estimation. Analyt. Bioanalyt. Chem. 413, 5901–5920 (2021).
Holton, E. et al. Spatiotemporal urban water profiling for the assessment of environmental and public exposure to antimicrobials (antibiotics, antifungals, and antivirals) in the Eerste River Catchment, South Africa. Environ. Int. 164, 107227 (2022).
Braoudaki, M. & Hilton, A. C. Adaptive resistance to biocides in Salmonella enterica and Escherichia coli O157 and cross-resistance to antimicrobial agents. J. Clin. Microbiol. 42, 73–78 (2004).
Okusu, H., Ma, D. & Nikaido, H. AcrAB efflux pump plays a major role in the antibiotic resistance phenotype of Escherichia coli multiple-antibiotic-resistance (Mar) mutants. J. Bacteriol. 178, 306–308 (1996).
Anahtar, M. N., Yang, J. H. & Kanjilal, S. Applications of Machine Learning to the Problem of Antimicrobial Resistance: an Emerging Model for Translational Research. J. Clin. Microbiol. 59, https://doi.org/10.1128/jcm.01260-20 (2021)
Schmieder, R. & Edwards, R. Insights into antibiotic resistance through metagenomic approaches. Future Microbiol. 7, 73–89 (2012).
Wang, Z., Walker, G. W., Muir, D. C. G. & Nagatani-Yoshida, K. Toward a global understanding of chemical pollution: a first comprehensive analysis of national and regional chemical inventories. Environ. Sci. Technol. 54, 2575–2584 (2020).
Persson, L. et al. Outside the safe operating space of the planetary boundary for novel entities. Environ. Sci. Technol. 56, 1510–1521 (2022).
Li, X. Z., Plésiat, P. & Nikaido, H. The challenge of efflux-mediated antibiotic resistance in Gram-negative bacteria. Clin. Microbiol. Rev. 28, 337–418 (2015).
Elkins, C. A. & Mullis, L. B. Mammalian steroid hormones are substrates for the major RND- and MFS-type tripartite multidrug efflux pumps of Escherichia coli. J. Bacteriol. 188, 1191–1195 (2006).
Kariuki, S. et al. Ceftriaxone-resistant Salmonella enterica serotype typhimurium sequence type 313 from Kenyan patients is associated with the blaCTX-M-15 gene on a novel IncHI2 plasmid. Antimicrob. Agents Chemother. 59, 3133–3139 (2015).
Perron, K. et al. CzcR-CzcS, a two-component system involved in heavy metal and carbapenem resistance in Pseudomonas aeruginosa. J .Biol. Chem. 279, 8761–8768 (2004).
Acknowledgements
L.M.M. was supported by a BBSRC/AstraZeneca iCASE PhD studentship (BB/T508330/1). A.H. was supported by FRESH CDT/AstraZeneca PhD Studentship (NE/R011524/1). W.H.G. was supported by a NERC Knowledge Exchange Fellowship (NE/S006257/1). A.K.M. was supported by a NERC Industrial Innovation Fellowship (NE/R01373X/1) and a NERC New Investigator Grant (NE/R01373X/1). The funders had no role in the conception nor writing of this paper.
Author information
Authors and Affiliations
Contributions
L.M.M: writing - original draught, writing - review and editing. A.H: writing - original draught, writing - review and editing, visualisation. J.S: writing – review and editing. B.K-H: writing – review and editing. W.H.G: writing - review and editing, supervision, funding acquisition. A.K.M: conceptualisation, writing – review and editing, supervision, funding acquisition.
Corresponding author
Ethics declarations
Competing interests
J.S. is a former employee and shareholder of AstraZeneca PLC. AKM is currently an Associate Editor for npj Antimicrobials & Resistance. All remaining authors declare no competing interests.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Murray, L.M., Hayes, A., Snape, J. et al. Co-selection for antibiotic resistance by environmental contaminants. npj Antimicrob Resist 2, 9 (2024). https://doi.org/10.1038/s44259-024-00026-7
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s44259-024-00026-7
