Antibiotic resistance is a well-recognized major threat to global public health. Many studies have surveyed antibiotic resistant bacterial strains or their corresponding genes in hotspots or reservoirs of antimicrobial resistance such as hospitals, waste-water treatment plants (WWTP), sewage systems, and animal feeding operations, where antibiotics are commonly found at high concentrations (Bengtsson-Palme et al., 2018; Burcham et al., 2019; Kraemer et al., 2019; Maestre-Carballa et al., 2019; Xiang et al., 2020). In particular, attempts have been made to correlate the concentration of antibiotics with the occurrence of antibiotic resistant bacterial strains or genes in these hotspots (Rodriguez-Mozaz et al., 2015; Pärnänen et al., 2019).

The complete set of resistance-related genes in a particular system is termed the environmental resistome, and the determination of its risk has been proposed by Martínez et al. (2015), where antibiotic resistance genes (ARGs) are ranked depending on the associated public health risks they might pose (Bürgmann et al., 2018; Asante and Osei Sekyere, 2019). It is known that constant exposure of microbial communities to higher concentrations of antibiotics exerts a selective pressure on the environmental resistome (Ebmeyer et al., 2021), and thus it is also expected that constant exposure to “low-level” or sub-inhibitory concentrations of antibiotics may similarly play important roles in driving antibiotic resistance in environmental microbial communities such as freshwater biofilms and agriculture-impacted environments (Rodriguez-Mozaz et al., 2015; Abe et al., 2020).

Agricultural practices such as irrigation, soil fertilization and animal production are major sources of antibiotic resistance transmission to soil and water resistomes (Manaia et al., 2020) since intensive agriculture operations are more likely to carry ARGs, antibiotic-resistant bacteria (ARB) and run-off chemicals. Even though antibiotics from agriculture run-off can be found at sub-inhibitory concentrations in receiving water bodies or downstream systems, constant exposure over long time periods will contribute to selective pressure for antibiotic resistance in environmental microbial communities such as freshwater biofilms (Andersson and Hughes, 2012; Manaia et al., 2020; Sasikaladevi et al., 2020).

One often overlooked dimension of biofilms is their virome, or the complete assemblage of viral particles that exist within a particular biofilm consortium. The ecological roles of viruses in the spread of antibiotic resistance are not yet fully understood (Bekliz et al., 2019; Palermo et al., 2019). Though it is known that bacteriophage can spread antibiotic resistance or virulence genes via transduction, the role of bacterial viruses as potential vectors for environmental antibiotic resistance dissemination is still an ongoing debate (Parsley et al., 2010; Enault et al., 2017; Petrovich et al., 2020). Overall, more studies are required to resolve the complexity of environmental resistomes in pursuit of minimizing the spread of antibiotic resistance (Quinlan et al., 2011; Kraemer et al., 2019; White and Hughes, 2019).

Most studies analyzing the prevalence of ARGs in hotspots of antimicrobial resistance focus on a single species or clinically relevant strains tested in vitro (Høiby et al., 2010; Partridge et al., 2018). While the information extracted from these in vitro experiments is relevant, its noteworthy to remember environmental biofilms are typically formed by myriad species, and the effect after exposure to antibiotic residues or antibiotic-resistant bacteria would likely differ significantly from that of mono-species biofilms. Interactions between the biofilm community and other organisms can range from the biofilm acting as a barrier to allochthonous, incoming organisms, to biofilms serving as a source of diversity which expands membership and adds heterogeneity to key microbial components of the aquatic food chain (Balcázar et al., 2015).

Resistance assays also often simply demonstrate growth in the presence (or absence) of an antibiotic using traditional approaches (e.g., MIC). The ability of sub-MIC levels to act selectively on a population of organisms (i.e., biofilm) is currently not well understood in the context of innate vs. acquired resistance. Innate resistance refers to wildtype genes encoding antibiotic resistance or innate properties of the cell, whereas acquired resistance is conferred following chromosomal mutations or the acquisition of resistance genes through HGT (Fernández and Hancock, 2012; Hall and Mah, 2017; Schroeder et al., 2017). Both mechanisms are known features of biofilm communities, and as such contribute toward the development, maintenance and transmission of ARGs and general resistance (Hall and Mah, 2017). It is clear that more community-based exposure experiments conducted under controlled conditions are needed to provide meaningful data to broaden our understanding of how complex communities respond to stresses induced by very low concentrations of antibiotics.

In addition to the release of antibiotic residues via the discharge of effluent from WWTPs, agriculture waste products and associated field run-off following application of animal waste as source of fertilizer represents a second major contributor to environmental antibiotic resistance (Topp et al., 2018; Goulas et al., 2020; Henriot et al., 2021). The demand for animal protein has been increasing over the last decade and global trends indicate intensive animal production will continue to increase the occurrence of ARB (Tiseo et al., 2020; Wang et al., 2021). It is estimated that 73% of all antimicrobials sold globally are used to raise animals for human consumption (Van Boeckel et al., 2019). While many countries have legislated tighter controls over antibiotic use (Zhao et al., 2020), others continue to permit prophylactic antibiotic usage in animal operations (Finley et al., 2013; Kraemer et al., 2019).

It is abundantly clear that exposure of microorganisms to veterinary/agricultural antibiotic residues can contribute to the development of ARGs and ARB (Manaia et al., 2020; Wang et al., 2021); however, little is known about the selective pressure that sub-MIC concentrations of antibiotics play on initial microbial colonization processes as well as biofilm development in receiving environments along with the environmental resistome (Yang et al., 2020). The environmental resistome influenced by the agriculture sources is characterized by two processes: extended periods of selective pressure imposed by antibiotic residues, and the subsequent dispersion of ARGs or ARB (Zhao et al., 2020). Dispersion is enhanced by the incomplete biodegradation of antibiotics, as the half-life of antibiotics in manure are estimated to be between 2 and 100 days. Released antibiotics are chemically diverse and degrade at different rates in the environment via sorption, photodegradation oxidation, and biodegradation (Grenni et al., 2018), making it difficult to predict how quickly they degrade and the extent to which they might continue to exert selective pressure for ARB development (O’Neill, 2016; Manyi-Loh et al., 2018).

It is noteworthy that unimpacted environmental resistomes deemed to be “pristine” with respect to anthropogenic activities, are increasingly difficult to find (Hooban et al., 2020). Widespread environmental pollution from antibiotic production facilities or agricultural operations has necessitated the call to monitor wildlife, agriculture, aquaculture systems, along with supposedly unimpacted environmental sites (Khan et al., 2013; Wellington et al., 2013; Thakur and Gray, 2019). In agricultural systems, veterinary drugs reach the surrounding environments, often in a metabolically-active form, through distribution and disposal of livestock manure and urine (Zhao et al., 2020). Manure is a common fertilizer used in farmland soils and crop production, and due to its potential to contribute to ARG dissemination to soils, aquatic systems, including plant microbiomes, it is a practice of ongoing concern (Zhao et al., 2020).

With regard to agriculture and aquaculture practices, 51 antibiotics have been reported as being in use according to the World Health Organization (WHO, 2017), including the six most-common classes of antibiotics: tetracyclines, quinolones, aminoglycosides, polymyxins, macrolides, penicillins, and sulfonamides (Chekabab et al., 2021). It is important to note that antibiotic usage for humans and animals often occurs within similar drug classes (e.g., gentamicin/aminoglycoside, ampicillin/ β-lactams, ciprofloxacin/fluoroquinolones, trimethoprim/antifolate, and tetracycline) (Finley et al., 2013; EUCAST, 2019); thus, the potential for cross-resistance to those drugs used in either setting exists. For instance, in soil microbiomes exposed to manure (Han et al., 2018), 163 ARGs were identified in untreated soils whereas 245, 230, and 245 ARGs were detected in soils treated with manure from poultry, swine and cattle, respectively. Of the total ARGs, the highest abundance was found to encode β-lactamases (19.6%), multi-drug resistance (18.8%) and macrolides (15.4%).

Tetracyclines and macrolides have historically seen extensive use in agriculture (Ebmeyer et al., 2021) and ARGs conferring resistance to these antibiotics are among the most prevalent (Chekabab et al., 2020, 2021). In China, most of the veterinary antibiotics consist of tetracyclines and sulfonamides (Zhou et al., 2020). Manure, soil and water samples of sixteen animal farms in Southeastern China were sampled to determine the abundance of ARGs (Wang et al., 2016); 22 ARGs were found to confer resistance to five major classes of antibiotics including tetracyclines, sulfonamides, quinolones, aminoglycosides, and macrolides. The author’s results showed that the spread of the agriculture-impacted resistome was dominated by sul genes, which were the most extensive, followed by tet and erm genes.

Metals can also act as co-selective agents driving the enrichment of genetic elements which contain resistance genes to both metals and ARGs (Baker-Austin et al., 2006; Zhao et al., 2020). The co-selection of metals and ARGs has been recorded in agriculture-impacted resistomes (Zhao et al., 2020). In some agricultural settings, Cu, As, and Zn continues to be used as supplementary feed additives for animal growth promotion and disease prevention (Poole, 2017).

Opportunistic bacterial pathogens such as Aeromonas spp., Citrobacter spp., Enterobacter spp., Mycobacterium spp., Legionella spp., Staphylococcus spp., Enterococcus spp., P. aeruginosa, and K. pneumoniae naturally occur in aquatic and soil environments and can persist and grow in biofilms of drinking water systems (Schwartz et al., 2003; Obst et al., 2006; Wingender and Flemming, 2011; Ahmad et al., 2021). Agricultural practices, such as irrigation, as well as drinking water distribution systems in which biofilms are known to form, have also been suggested to act as reservoirs for pathogenic microorganisms (Wang et al., 2021). As biofilms develop, resistant bacteria can be released to continue their life cycle as planktonic cells, and subsequently colonize, or disperse to, other ecosystems carrying their ARG compliment. Additionally, pathogenic bacteria such as P. aeruginosa and Legionella pneumophila that are resistant to an antibiotic can survive sewage treatments and migrate downstream to eventually be included in a biofilm (Wingender and Flemming, 2011). Most bacterial populations inhabiting freshwater biofilms are non-pathogenic. However, non-pathogenic bacteria have a higher chance of interacting with pathogenic bacteria in high-density systems such as biofilms (Wingender and Flemming, 2011).

Veterinary antibiotic usage over the last several years have been reduced worldwide as result of greater awareness of the threat of antibiotic resistance (Kraemer et al., 2019); however, agricultural antibiotic use still remains a major source of environmental antibiotic pollution (Wang et al., 2021). Overall, the effects of sub-MIC antibiotics outside clinical settings have not yet been as fully explored nor monitored possibly due to detection difficulties pose by low effective concentrations. However, recent studies support the premise that acquisition of resistance can occur at sub-MIC antibiotic concentrations (Marti et al., 2014; Chow et al., 2015, 2021). Clearly, more experimental data is required to further comprehend the risk of exposed to low dosages of antimicrobials, as well as the response interactions from the entire microbial community and not just that of individual bacteria.

Over the past decade, studies of the ecology bacterial viruses (bacteriophage) have become more feasible thanks to the increased accessibility of high-throughput sequencing methods. However, a clear understanding of the ecological role of bacteriophage in aquatic environments, including freshwater biofilms, is still lacking (Battin et al., 2016; Bekliz et al., 2019; Palermo et al., 2019). This lack of knowledge of environmental phage is compounded by the inability to characterize the majority of bacterial viruses, with over 65% of viral sequences from surface waters being unrecognized when compared to existing databases (Palermo et al., 2019).

Flood and Ashbolt (2000) observed that wetland biofilms could entrap viral-sized particles and concentrate them over 100-fold, compared to suspended numbers in the surrounding water column. Further, in the different stages of WWTPs, phage have been observed at concentrations ranging from 10⁸ to 10¹⁰ ml^–1, which is 10–1,000 times higher than in natural (unimpacted) aquatic environments, indicating WWTPs as important reservoirs of bacteriophage (Wu and Liu, 2009). While bacteriophage do not represent a direct threat to human or domestic animals, they shape the diversity and genetic architecture of their hosts (Chopyk et al., 2020). Since there are records of phage invading biofilms by disrupting the EPS matrix or by killing embedded cells (Sutherland et al., 2004), it was originally believed that biofilm formation and phage transmission were mutually exclusive (Abe et al., 2020). However, more recent work has described phage as playing supporting beneficial roles in high density microbial colonies such as biofilms, as cell lysis mediated by phage infection often leads to production of e-DNA. The e-DNA released by phage may then be integrated by the biofilm’s matrix, thus increasing the opportunity for acquisition of resistance genes through HGT (Fernández et al., 2018).

Most studies of bacteriophage composition have focused on WWTP microbial communities, where the order Caudovirales comprises most of the viral sequences, with dominant viral families including Myoviridae, Siphoviridae, and Podoviridae (Parsley et al., 2010; Allen et al., 2011; Cantalupo et al., 2011; Tamaki et al., 2012; Chopyk et al., 2020; Petrovich et al., 2020). A similar virome composition was observed in benthic biofilms of three different freshwater streams, where Siphoviridae, Podoviridae and Myoviridae were the dominant viral taxa (Bekliz et al., 2019). Moreover, it was observed that ARGs such as sul1, bla_TEM and bla _CTX–M, were more persistent after WWTP treatments (chlorination, thermal treatment and natural inactivation) when more of these ARGs were found within the protein capsid of phage particles than found within bacteria (Calero-Cáceres and Muniesa, 2016). It is thought that the protein capsid could provide DNA fragments with enough protection to survive various environmental stressors, thereby serving as a significant reservoir. Furthermore, the possibility of ARGs being transferred to a suitable host would also increase under such a scenario (Calero-Cáceres and Muniesa, 2016).

One study that described both the bacterial and viral composition of the swine gut identified ARGs such as MDR efflux pumps [bcrA, macB, mef(A)] within the virome and that the abundance of phage integrase-encoding genes were significantly increased in the viromes of medicated swine relative to non-medicated swine, indicating induction of bacteriophage in the presence of antibiotics (Allen et al., 2011). However, the authors did not observe significant changes in bacteriophage abundance, suggesting that while they were induced in the swine gut, this did not result in the transfer of ARGs via prophage to bacteria in the swine gut. In contrast, a study by Zhang and LeJeune (2008) demonstrated that blaCMY-2, tetA and tetB were transferred via phage between Salmonella serovars of bovine origin, providing evidence of ARG propagation by phage-mediated transduction events.

Bacteriophage are becoming increasingly recognized for their potential role in the dissemination of antimicrobial resistance, as well as other genes that shape bacterial community composition (i.e., virulence factors, auxiliary metabolic genes) (Colavecchio et al., 2017; Breitbart et al., 2018; Chopyk et al., 2020). Evidence suggests that dissemination of ARGs by bacteriophage may be possible between different, and even remote, environments not contaminated by antibiotics as a result of anthropogenic activities (Muniesa et al., 2013). Just as exposure to antibiotics leads to selective pressure on susceptible bacteria, the constant presence of phage and/or competing viruses may exert a selective pressure on microbial communities, influencing the prevalence and abundance of antibiotic resistant genes (Madsen et al., 2012; Beceiro et al., 2013). In studies that compared viral and bacterial taxonomic composition, dsDNA phage were the most abundant of the viral members observed, though it was noted that virome studies targeting bacteriophage, and the availability of sequenced viromes, is currently limited (Calero-Cáceres et al., 2019).

There is some evidence to suggest viral taxa may not have a great impact on ARG transfer via transduction (Colavecchio et al., 2017; Enault et al., 2017). However, transduction tends to be more common in sites with high viral and bacterial density (e.g., WWTPs) (Petrovich et al., 2020), suggesting a different pattern for biofilms where their high microbial diversity and density could possibly enable them to be more susceptible to viral-ARG associations. Nonetheless, there is an ongoing debate on whether viruses are associated with antimicrobial resistance spread and if they contribute significantly to HGT (Parsley et al., 2010; Enault et al., 2017; Petrovich et al., 2020). Notably, very few studies have compared the dynamics of bacterial and viral communities in biofilms. Moreover, few studies have attempted to correlate the effects of antiviral drugs in combination with antibiotics on microbial communities, and monitoring of antivirals in wastewaters, surface waters, groundwater, and drinking water is not currently conducted. However, some antivirals are highly bioactive and are proposed to negatively affect non-target organisms and persist in aquatic environments (Jain et al., 2013; Sims and Kasprzyk-Hordern, 2020). For this reason, it is important to investigate the simultaneous occurrence of antiviral drugs with antibiotics.

Further research is clearly needed to understand associations between bacterial and viral taxa, particularly in critical environmental niches (e.g., biofilms) to better address whether viruses contribute to the spread or transfer of ARG. More insight into the relations of bacteriophage and the prevalence of antimicrobial resistance would provide data for improved methods for effluent water treatment, better comprehension of the interactions between viromes and sub-MIC-exposed environmental resistomes, and offer a new framework for monitoring strategies to mitigate the spread of pathogenic viruses and antimicrobial resistance.

While many studies of ARG and ARB spread have focused on hotspots where high antibiotic concentrations are released (e.g., hospitals, WWTPs and large animal production facilities), the majority of downstream environments are routinely exposed to much lower concentrations of these metabolically-active agents. The extent to which these more prevalent antibiotic contaminants drive proliferation of antibiotic resistance in the environmental resistome remains to be fully elucidated. Interactions of antiviral and antibiotic drugs between the microbial community of pristine (unimpacted) environmental reservoirs are complex and also poorly understood, although seemingly offer myriad non-mutually exclusive events with potential to impact the persistence and spread of antibiotic resistance.

Most environmental resistomes are becoming increasingly influenced by the constant selective pressure from human activities, and in particular, from industrialized agriculture practices. In this review, we have focused on the effects of low antibiotic concentration exposure of biofilms colonizing freshwater environments and as a result of agricultural practices given the growing evidence suggesting sub-MIC selectively influences the prevalence of ARGs and ARB. Furthermore, the rich microbial diversity of biofilms commonly includes high concentrations of viruses, and in particular, bacteriophage, which are a potentially important and under-appreciated mechanism of antibiotic resistance propagation. The environmental virome as an integrated part of the environmental resistome, and specifically the bacteriophage’s role in the environmental transfer of antimicrobial resistance, requires more study to better understand the role of the environment in antibiotic resistance transmission.

GF-V: writing—original draft. JB: supervision and writing—review and editing. JRL: writing—review and editing. DRK: supervision, writing—review and editing, and funding acquisition. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.