Food-to-Humans Bacterial Transmission

ABSTRACT

Microorganisms vehiculated by food might benefit health, cause minimal change within the equilibrium of the host microbial community or be associated with foodborne diseases. In this chapter we will focus on human pathogenic bacteria for which food is conclusively demonstrated as their transmission mode to human. We will describe the impact of foodborne diseases in public health, the reservoirs of foodborne pathogens (the environment, human and animals), the main bacterial pathogens and food vehicles causing human diseases, and the drivers for the transmission of foodborne diseases related to the food-chain, host or bacteria features. The implication of food-chain (foodborne pathogens and commensals) in the transmission of resistance to antibiotics relevant to the treatment of human infections is also evidenced. The multiplicity and interplay of drivers related to intensification, diversification and globalization of food production, consumer health status, preferences, lifestyles or behaviors, and bacteria adaptation to different challenges (stress tolerance and antimicrobial resistance) from farm to human, make the prevention of bacteria-food-human transmission a modern and continuous challenge. A global One Health approach is mandatory to better understand and minimize the transmission pathways of human pathogens, including multidrug-resistant pathogens and commensals, through food-chain.

INTRODUCTION

Food is considered one of the main environmental drivers shaping the human microbiota across the life span. Microorganisms vehiculated by food can be related to a variety of scenarios, including those benefiting health (e.g., stimulation of host antibodies, release of chemicals to stimulate the health of the overall system, or inhibition of pathogen development), those causing minimal change within the equilibrium of the host microbial community, and those that are pathogenic or have been associated with gut-host dysbiosis (1–3). Recently there has been an increase in knowledge on gut bacterial genera and species commonly affected by diet, as well as evidence suggesting that the intestinal microbiome plays an important role in modulating the risk of several chronic diseases (e.g., inflammatory bowel disease, obesity, type 2 diabetes, cardiovascular disease, and cancer) (1–3). Nevertheless, comprehensive information about the types of diet that transmit bacteria implicated in those diseases, as well as environmental and host factors favoring their colonization, remains scarce. Notwithstanding, food as a transmission mode for microorganisms reaching humans is extensively characterized for different pathogenic bacteria, the environment, animals, and humans being their main reservoirs (Fig. 1) and the fecal-oral route their main transmission route (4–6). A triad including a contaminated food item, a susceptible human host, and bacterial pathogens able to survive and multiply in specific environmental conditions must be present for the occurrence of a foodborne disease. Nevertheless, transmission of typical foodborne pathogens can also occur more rarely by alternative transmission modes, as by direct contact of humans with infected animals or between humans (7).

FIGURE 1.

Reservoirs of the main pathogenic and potentially beneficial bacteria transmitted from food to humans.

In general, bacterial pathogens cause foodborne diseases by three mechanisms: ingestion of preformed toxins in foods (intoxication; e.g., Staphylococcus aureus), production of toxins within the gastrointestinal tract following ingestion of pathogens (food toxicoinfection; e.g., Clostridium perfringens), or invasion of the intestinal epithelial cells (infection; e.g., Salmonella) (6). Most foodborne bacterial pathogens are often associated with a self-limiting gastroenteritis syndrome with nausea, vomiting, diarrhea, abdominal pain, and sometimes fever. Nevertheless, such bacteria might also cause severe illness with extraintestinal infections, postinfection sequelae, and even death, especially in individuals in high-risk groups (infants, young children, the elderly, and immunocompromised patients). Few bacterial pathogens (e.g., Clostridium botulinum and Salmonella enterica serovar Typhi) are associated with systemic clinical symptoms and more-severe clinical outcomes, even in individuals without risk factors (6).

In this review, we will focus on pathogenic bacteria for which food is conclusively demonstrated as their transmission mode to humans. The impact of foodborne diseases on public health, different scenarios evidencing relevant drivers for bacterial pathogen transmission, and the implication of the food chain in the widespread resistance to antibiotics that is critical to the treatment of human infections will be presented.

FOODBORNE DISEASES AND PUBLIC HEALTH

The first global estimates of foodborne diseases by the World Health Organization (WHO) indicate that about 1 in 10 people (600 million) around the world is sickened after eating contaminated food each year, with 420,000 deaths reported (8). The global burden of foodborne diseases has been measured in disability-adjusted life years (DALYs), which more fully capture disease symptoms and severity. In 2010, DALYs data showed the loss of 33 million healthy life years, most among children <5 years of age and in low-income countries (African and Southeast Asian regions) (8). The most frequent causes of foodborne disease worldwide are bacterial pathogens, the most important being the zoonotic Campylobacter and nontyphoidal Salmonella (NTS) (8). Certain diseases, such as those caused by NTS, are a public health concern across all regions of the world, in high- and low-income countries alike. Other diseases, such as typhoid fever, foodborne cholera, and those caused by pathogenic Escherichia coli, are much more common in low-income countries, while Campylobacter is an important pathogen in high-income countries (8). Besides the direct impact of foodborne diseases in human health, they also have health care, economic, and welfare costs, as in the case of the United Kingdom and the United States, where available data estimated a cost of £1.5 billion and of $14 billion, respectively (9, 10).

Currently, most human infections have a zoonotic origin, and ∼75% of new and emerging pathogens originated from animals, including the foodborne ones (11). Transmission starts from the original habitat (the animal reservoir) to the food, along the complex farm-to-table pathway. Maintenance of a reservoir of zoonotic foodborne pathogens is favored by their transmission within the food animals and on-farm production environment, through diverse routes (e.g., contaminated feed and water, humans, rodents, and flies) (12).

Almost all type of foods (vehicles of transmission) can be contaminated by foodborne pathogens. Nevertheless, certain vehicles pose a greater threat than others, being classified as high-risk foods. They include raw or undercooked foods of animal origin (meat, poultry, eggs, fish, and milk) and, more recently, fresh produce (including leafy vegetables and sprouts) (13, 14). Also, food constituted with multiple ingredients and with a high level of handler manipulation has a greater potential to be associated with foodborne diseases (13). Some food types are considered unexpected vehicles of pathogens by the nature of the food matrices preventing microorganism multiplication, such as peanut butter, caramel apples, peppers, and chocolate. However, such food vehicles have been recently associated with large outbreaks, sometimes in wide geographic regions (15–19).

In recent decades, the epidemiology of foodborne outbreaks changed from acute and local to diffuse and widespread (e.g., geographically dispersed in many places at once), mostly due to production intensification and wide distribution of food (20–22). This challenge spurred the development of rapid and more sensitive molecular methods for foodborne pathogen surveillance, such as whole-genome sequencing (WGS). It became crucial not only to rapidly identify the responsible pathogen (classical and newly emergent ones) and source of origin (including the contexts that led to food contamination) but also to identify new or unsuspected transmission routes and to support public health interventions in global food markets (23, 24). For example, WGS has been applied to source attribution of campylobacteriosis (25, 26), outbreak investigations of salmonellosis (10, 18, 21, 27) and listeriosis (19, 28), as well as to differentiating persistent contamination by Listeria monocytogenes from reintroduction in food-associated environments (29). Beyond high-resolution subtyping, WGS could also allow characterization of foodborne pathogen virulence markers (linked to greater pathogenicity) and stress tolerance as well as antimicrobial resistance prediction, both features important to microbial risk assessment (23, 24). For example, in 2011, Shiga toxin-producing E. coli (STEC) O104:H4 caused one of the largest foodborne outbreaks of recent history, with >3,000 cases of infection and >40 deaths reported, in multiple European countries and North America. Within a week, WGS revealed that the unusually virulent outbreak strain belonged to a distinct group of enteroaggregative E. coli that had acquired the combination of genes coding for Shiga toxin 2, antimicrobial resistance, and other virulence factors (30). Despite the promising application of WGS to improve foodborne pathogen surveillance, the use of standardized methods that can be practiced among routine laboratories remains a challenge, namely, the selection of standardized analytical tools as well as epidemiologic interpretation of WGS data (23, 24).

DRIVERS OF FOOD-TO-HUMAN BACTERIAL PATHOGENIC TRANSMISSION

Several drivers have been identified to promote the increase of food-to-human bacterial pathogenic transmission and foodborne diseases worldwide by leading to the introduction and amplification of pathogens along the complex farm-to-human pathway (31–37). This section will address the driving factors promoting such events associated with the features of bacterial pathogens, the food chain, and the human host (Fig. 2). Diverse examples of transmission scenarios involving those drivers will be given.

FIGURE 2.

Factors that drive transmission of pathogenic bacteria from food to humans.

Drivers Related to the Food Chain

Changes in the food chain, involving intensive and large-scale production and distribution, climate change, and globalization of food and live animal suppliers, are considered particularly significant drivers for the emergence of foodborne diseases (34, 36).

Transmission Scenarios Related to Intensive and Large-Scale Food Production and Distribution

Shifts in food production, processing, and distribution during the last several decades have contributed to the increased risk and complexity of foodborne diseases, resulting in new opportunities for pathogen transmission to humans, particularly of zoonotic agents (e.g., Salmonella and Campylobacter). The increased demand for food due to worldwide population growth contributed to the growth of industrial-scale production systems, including intensification of animal production plus agriculture and large-scale food processing and distribution (31, 33).

Intensive livestock practices with high animal densities promote persistence and dissemination of zoonotic pathogens in food animals and farm production environments, leading to the maintenance of a reservoir of foodborne pathogens directly transmitted to foods of animal origin and indirectly to fresh produce (12). This is illustrated by pandemic S. enterica serovar Enteritidis in the 1980s and ’90s, mostly attributed to intensive production of eggs/broilers, which was later overcome with successful control programs, including adequate biosecurity measures (38). More recently, in the European Union, contamination in the large turkey production chain was associated with a large outbreak of S. enterica serovar Stanley (39). Also, in the case of Campylobacter it was shown that the major contamination source of chicken meat was at the large production rearing farm (40). More recently, changing farming practices for free-range and organic animal production (allowing outdoor access for farm animals) have raised new questions about food safety, as this type of production is also associated with several hazards, including Campylobacter and Salmonella (41).

The intensification of vegetable and fruit production has been also associated with several risks related to the use of unsafe irrigation water and manure, poor worker hygiene, inadequate pest control, and improper cleaning of sanitizing or harvest equipment and utensils, among others (42). It has been shown that organic agriculture practices related to seed production (e.g., contaminated irrigation water and/or manure) have contributed to identification of sprouted seeds as a high-risk food for STEC and Salmonella (43). In the United States, unsanitary conditions identified in one farm and processing operations environment (failure in sanitizing and fruit precooling; ineffective cleaning due to inadequate facilities and equipment design) resulted in contamination of cantaloupes with diverse L. monocytogenes strains, leading to a large and fatal outbreak (147 cases, 33 deaths, and 1 miscarriage) (22, 44). Other foodborne outbreaks associated with fruits and vegetables through contact with surface or irrigation water have been reported and recently revised (45).

Besides production practices, climate change also can impact crops, food animals, and the growth and survival of pathogens, as well as exposure and transmission pathways of foodborne diseases (46, 47). Climate changes potentially enhance the ability of pathogens to survive and persist in soil, crops, and water, increasing the probability of their transmission thorough water, seafood, or vegetables (46, 48–50).

The complexity of the food chain due to industrial-scale production systems can also increase the risk of foodborne outbreaks, with some occurring during long periods of time (51–54). A simple failure in a single industrial-scale production resulting from contamination with pathogens during food processing (e.g., from the environment, surface, or food ingredients) can result in distribution of contaminated food batches to millions of consumers, affecting their health in different locations at the same time. In 2000, a massive outbreak of food intoxication from milk products in Japan resulted in >10,000 cases, the contamination point being a production line valve of a major dairy-products processing plant in Osaka that became contaminated with staphylococcal enterotoxin type A (51). Recently, a 2017–2018 outbreak in the European Union due to contaminated infant formula from a French dairy company with S. enterica serovar Agona led to the recall of >10 million boxes from multiple countries (52). A 2-year (2015–2016) contaminated food-processing environment resulted in a deadly L. monocytogenes outbreak linked to bagged salads (33 cases and 4 deaths in the United States and Canada) (53). One example of a contaminated food ingredient affecting multiple products was the large outbreak (714 cases and 9 deaths in 46 states in the United States) of S. enterica serovar Typhimurium linked to peanut products from a plant evidencing diverse possibilities of pathogen transmission (e.g., rodent-accessible entryways, an unsealed air-handling system, and rain leakage into peanut storage areas) to food (17). There are also comprehensive published examples of food safety failures (e.g., ignored positive tests of microbiological hazards, falsification of food safety documents, inadequate controls to prevent contamination and kill bacteria, and insufficient cleaning and sanitation) that led to large outbreaks and deaths from foodborne diseases, demonstrating that maintaining a food safety culture within the entire organization is crucial (55).

Transmission Scenarios Related to Globalization of the Food Supply

A growing international food trade resulting in the export and import of contaminated farm animals, food products (e.g., vegetables, fruit, seeds, meat and poultry, and ethnic foods), and single meals containing ingredients from multiple regions is contributing to changing trends in foodborne diseases. Most of the food products consumed in developed countries are produced on industrial scale, being also increasingly imported (not grown or produced locally) due to changing consumer demand for a wide diversity of foods and fresh produce year-round (33, 56). However, import from other countries includes those with lower food safety standards (e.g., lacking meat processing plant inspections) or without appropriate food production and processing safety practices (e.g., untreated human/animal manure, contaminated water, or untreated wastewater sewage) (33, 42). In the European Union, the most-imported food categories in 2016 were fruits and nuts among vegetable products and fish, crustaceans, and aquatic invertebrates among animal products (57), as occurs in the United States, where most seafood and a moiety of fruits are imported (56). Both types of foodstuffs have been reported in developed regions as presenting an increased association with foodborne disease outbreaks in recent years (13, 14, 27, 56, 58).

In addition, the trade of live animals and food-animal products has also been implicated in the transmission of foodborne pathogens to humans. In Europe, this is exemplified by contaminated products resulting in multicountry foodborne outbreaks, as the 2017–2018 S. Enteritidis outbreak linked to eggs (table eggs and consumed raw in some recipes) from Poland, with its origin in egg-packing centers and laying-hen farms (59). Recently, poultry products imported from Brazil, one of the biggest exporters, have been linked to the spread of epidemic clones of an uncommon and multidrug-resistant (MDR) S. enterica serotype, S. Heidelberg, into European countries (60, 61). Aquaculture practices applied in seafood and fish production in many exporting countries, mostly in Asia (e.g., integrated fish-pig farms in Southeast Asia), are also related to food contamination by pathogens (33, 62). For example, using a novel combination of WGS analysis and geographic metadata to create a transmission network, it was possible to trace the origin of a U.S. outbreak of S. enterica serovar Bareilly to scraped tuna imported from a fishery facility in India (10).

Recent studies showed that illegal importation of food-animal products (including bush meat and livestock meat) from Africa to Europe through airport passengers’ luggage was commonly verified, some of those illegal food items being contaminated with foodborne pathogens such as NTS, suggesting a potential public health risk by a neglected route of transmission (63, 64). Also, uncontrolled entry of foodstuffs into the European Union (including from the black market at the EU border) have been recently reported as a neglected route of potential methicillin-resistant S. aureus (MRSA) transmission (65).

Drivers Related to the Host

Changes in consumer demands for foods, lifestyles, and behaviors related to food safety as well as the increasing proportion of vulnerable human host groups can also increase exposure to bacterial pathogens and susceptibility to infectious diseases (33, 34, 36).

Transmission Scenarios Related to Changes in Consumer Lifestyles and Risk Behaviors

In recent decades, consumer demand for more-diverse (e.g., tropical products, fruits and vegetables year-round), convenient ready-to-eat (RTE) (e.g., pre-prepared meals, ready-washed fruit and vegetables, and quick-cook sauces), and healthy foods (e.g., those with less salt), known to be associated with higher foodborne pathogen contamination, constitutes a driving factor in the emergence of foodborne diseases. One of the major examples is the increasing consumption of fresh produce due to health concerns, which have been frequently associated with foodborne outbreaks linked to leafy green vegetables (e.g., spinach and lettuce), sprouts, and fruits (e.g., melons and papayas) in developed countries (56, 66).

The increasing trend of consumption of uncooked or undercooked food of animal origin (e.g., meat and poultry, fish, seafood-and-meat recipes, and raw milk) and convenience foods like RTE also potentiate transmission of foodborne pathogenic bacteria (34). For example, consumption of raw milk, in which interest is growing in the European Union, was identified as clearly associated with transmission of human disease caused by Campylobacter, Salmonella, and STEC (67). Moreover, the increasing consumption of minimally processed foods with extended shelf lives, such as refrigerated or frozen RTE products, was proposed as a possible factor in the promotion of human listeriosis in Europe (68). Also, several cases and outbreaks of Vibrio parahaemolyticus infections have been reported in European countries, associated with higher consumption of raw and undercooked fish and shellfish (including imported) in recent years coupled with an increase in the number of susceptible individuals consuming seafood products and with warming of marine waters as a result of global climatic change (50, 69, 70).

Besides consumer demand for specific food products, the new generation of active consumers frequently eat out of the home, being more exposed to transmission of foodborne pathogens by contaminated food prepared in public places. According to the European Food Safety Authority (EFSA), in the European Union almost half of foodborne outbreaks occurred in food service establishments (13). Eating out exposes consumers to food-handler behaviors, whose food safety practices (e.g., personal hygiene to avoid microbial spread, proper food handling to avoid cross-contamination, and efficient cooking and storage temperatures to avoid microbial growth) are crucial to prevent transmission of foodborne pathogenic bacteria to humans. A recent report studying EU salmonellosis cases associated with catering facilities over a 15-year period identified food-handler behaviors (e.g., cross-contamination between heat-treated foods and raw materials or improperly cleaned food-contact surfaces) as potential risk factors (71). Also, it was shown that food risk behaviors (e.g., serving chicken at barbecues when unsure it was fully cooked—a Campylobacter risk factor) in U.K. kitchens were widely prevalent among chefs and catering students and the public (72).

Consumers have more opportunities to travel today, positioning travel and tourism as driving forces contributing to exposure to contaminated food (73, 74). For example, an international outbreak of S. Heidelberg associated with in-flight catering meals affected 25 people of 5 nationalities (74). In addition, food tourism creates more opportunities for consumers to be exposed to pathogenic bacteria by consumption of traditional food recipes or exotic cuisine (75). This was illustrated by the association of Peruvian goat cheese (often made with unpasteurized milk) with brucellosis (11) and reptile meal with salmonellosis (75). Migration may also change food preparation behaviors and consumption habits (34). This was the case with an outbreak of listeriosis among Mexican immigrants (mostly affecting pregnant women) in the United States as the result of consumption of illicit, noncommercial, homemade, Mexican-style cheese produced from contaminated raw milk sold to unlicensed cheese makers by a local dairy (76).

However, it is of note that a large number of foodborne outbreaks occur in the home/domestic kitchen (13). Such events are mostly associated with poor food-handling behaviors, such as inadequate time-temperature control, cross-contamination, poor personal hygiene, insufficient cleaning, and using food from unsafe sources (6, 77). It was noted that unsafe practices (including those related to storage time and temperatures) are not uncommon with elderly people (>10% of the persons studied), one of the risk group consumers, having a potential impact on the human listeriosis risk (68). In addition, it was recently reported that the temperatures in EU domestic refrigerators are high (ranging from −8 to −4 to 11 to 21°C), which could potentiate L. monocytogenes growth (68). Also questionable is the proper use of home cooking or reheating methods like microwave and other emerging cooking machines (“cooking robots”). For example, the use of ready-to-cook products for microwave preparation before consumption could be a risk if the temperatures are not sufficient to kill pathogens like Salmonella (33, 34). This is illustrated by an outbreak of Salmonella possibly related to failures in the microwave cooking of frozen, not-ready-to-eat, microwaveable foods due to misinterpretation of the product’s cooking label instructions and/or unfamiliarity with the oven’s wattage (78). Also, during summer activities (picnics, barbecues, and summer festivals), the increased risk of cross-contamination and lack of personal hygiene together with the difficulty of keeping food at safe temperatures coincides with peaks of some foodborne infections (47).

Transmission Scenarios Related to Host Risk Population

In recent decades, the worldwide increase of vulnerable people, particularly the elderly and the immunocompromised (e.g., due to cancer, transplant interventions, AIDS, diabetes, liver or kidney disease, and malnutrition), contributed to more-effective transmission of particular pathogens to a wide number of susceptible hosts. Around 15 to 20% of people of developed countries, such as the United Kingdom and United States, belong to this group, with increased vulnerability resulting in development of foodborne disease by a lower infectious dose than typically expected or in diseases with increased severity (79). L. monocytogenes is recognized as one of the main foodborne pathogens with great impact in vulnerable groups. In a recent EFSA report, the only clearly identified risk factor for the increasing trend in cases of listeriosis in the European Union was the increase in the elderly population (68). In France, patients with chronic lymphocytic leukemia had a listeriosis incidence >1,000 times greater than that of the population with no risk factors (80), and low infectious doses of Listeria associated with milkshakes affected only hospitalized patients in the United States (81). The elderly was the group most likely to die after infection with STEC, and children <5 years old develop more-severe infections (82). Infant botulism is associated with honey consumption and with the growth of Clostridium botulinum and neurotoxin production in the gut of infants <1 year old (83). It was also demonstrated that in patients with chronic liver diseases (cirrhosis), elevated serum iron levels, or immunodeficiency, Vibrio vulnificus causes frequently rapid fatal septicemia, mostly associated with previous consumption of raw oysters (84, 85). The wide use of antacids (particularly proton pump inhibitors) or constipating drugs (e.g., antipsychotics) by people whether or not they are in the vulnerable groups might also increase susceptibility to foodborne diseases, at least during a transitional period (79, 86). There is evidence that patients with hypochlorhydria or achlorhydria or who have been treated with proton pump inhibitors or H₂ receptor antagonists are more susceptible to Campylobacter, E. coli O157:H7, L. monocytogenes, Salmonella, Shigella, and Vibrio cholerae than healthy persons (79, 87–89). The increasing numbers of vulnerable populations and their risk of easily acquiring a foodborne disease impose the need for food safety management systems among food suppliers of hospitals, nursing homes, elderly-care homes, schools, and day care centers for children (79).

Drivers Related to Foodborne Bacterial Pathogens

Foodborne bacterial pathogens face a variety of adverse effects or stresses during transmission events from their reservoirs to the human host. Stressful conditions occur from the agriculture level to food processing, storage, distribution, cooking, and within the human host (12, 90–93). Such stresses often co- and/or sequentially occur in the same ecosystem (e.g., application of hurdle technology in food preservation), during ecosystem transition (e.g., from acid food to acid pH in stomach and bile salts in animal and human gut), or during infection (e.g., high temperature or acidic pH within macrophages), imposing on bacteria the necessity for ceaseless adaptation (90, 91, 93–97).

Bacteria stress responses can be stable or transient. Stable phenotypes, for example, associated with genetic mutation, can be positively selected and fixed in the bacterial population under stressful conditions. A transient stress response can be associated with differential expression of particular genes and generally occurring while the stress is present. These types of responses contribute to protect vital processes, to restore cellular homeostasis, to repair the damage, to counteract or eliminate the stress agent, and/or to increase the cellular resistance against subsequent stress challenges (96). The ability to overcome sublethal or normally lethal stressful conditions was described for such diverse foodborne pathogens as Campylobacter jejuni (e.g., to oxidative stress, osmotic stress, or antibiotics) (98–100), Salmonella (e.g., to acid, low water activity, thermal treatment, high hydrostatic pressure, metals, or antibiotics) (101–106), STEC (e.g., to acid, osmotic stress, oxidative stress, or heat) (99, 107, 108), L. monocytogenes (e.g., to acid, osmotic stress, or disinfectants) (99, 109–111), V. parahaemolyticus (e.g. to salt, acid, or cold) (112), Bacillus cereus (e.g. to acid, cold, or reducing atmospheres) (96, 113), and Cronobacter sakazakii (e.g., to heat, osmotic stress, or desiccation) (99, 114), among others, stress responses being strain dependent, i.e., variable among members of the same species (90). Moreover, the presence of a particular stress can modulate bacterial response to that stress as well as to other ones, the latter phenomenon denominated stress cross-protection (90, 91). In fact, preexposure to sublethal levels of a given stress protects bacteria during subsequent exposure to normally lethal conditions, as was described for Salmonella adaptation to bile or to acid (115, 116). Diverse Salmonella serotypes (S. Typhimurium, S. Enteritidis, S. Newport, and S. Infantis) were cross-protected against UV irradiation, various sanitizing agents, dry heat, and bile salts when under desiccation stress (117). Induced response after contact with sublethal stress or cross-protection events can be found in the literature for other foodborne pathogens, including V. parahaemolyticus, L. monocytogenes, and STEC (112, 118, 119).

Against all the above, it is evident that bacterial stress response can be a complex scenario, dependent on the interplay of microbial-specific features and intrinsic (e.g., food matrix), extrinsic (e.g., temperature), or food-processing factors (e.g., hurdle technology). According to the nature of such interactions (e.g., bacterial species, strain type, stress concentrations, one or multiple stresses, food components, and competing microbiota in the host), different outcomes in bacterial survival, multiplication, and persistence in the food chain and human host can occur, with potential implications for foodborne pathogen transmission dynamics. In fact, the ability of bacteria to survive and multiply under the stressful conditions occurring in the food chain might increase human exposure to a greater number of pathogenic bacterial cells in food (e.g., through higher pathogen shedding from animal reservoirs, survival of food processing, or persistence in food-processing facilities) as well as to induced stress strains producing greater amounts of toxins responsible for food poisoning symptoms or requiring lower infectious doses than typically expected (120–125).

Bacterial Strain Diversity Favoring Pathogen-Food-Human Transmission Scenarios

It is well known that bacteria adapt to environmental conditions by the acquisition of specific genetic clusters through horizontal transfer, genetic recombination, or mutations as well as by differential expression of regulatory pathways (126). With the advances of genomic, transcriptomic, and single-cell studies applied to foodborne pathogens, it is becoming evident that the dynamic response of microorganisms to changing environmental conditions depends on the behavior of individual cells within the bacterial population (127, 128). Thus, it is critical to understand the genetic context, epigenetic mechanisms, and occurrence of heterogeneous behaviors of individual bacteria, as they could have an impact on the success of food-processing and -preservation strategies, on bacterial persistence within food facilities, and on the food-to-human transmission of particular strains (e.g., serotypes or clonal lineages). Some examples illustrate variable features among bacteria that potentially contribute to pathogen-food-human transmission scenarios.

At the farm level, emergent clinically relevant S. enterica serotype Rissen/ST469 and the European clone of S. Typhimurium/ST34 or S. enterica serotype 4,[5],12:i:-/ST34 of pig origin are more tolerant to toxic concentrations of copper under anaerobiosis, due to the genetic horizontal acquisition of the sil gene cluster, than other Salmonella serotypes from the same origin that are less associated with human infections (102). Such data suggest that the widespread use of copper as a feed additive in pigs might contribute to the selection of these fitter clones, and consequently to pig meat contamination and their transmission to humans.

At food-processing facilities, L. monocytogenes acquiring the qacH and bcrABC genes had a survival advantage when sublethal concentrations of quaternary ammonium compounds (QACs) often used in disinfection remained on surfaces due to insufficient rinsing methods (110). QAC tolerance genes were often found in isolates belonging to clonal lineage II-serotype 1/2a-ST121, the dominant persistent ST type worldwide associated with food-processing plants and occasionally isolated from human infections (129–132), as well as in lineage I-serotype IVb-ST6 isolates associated with human meningitis (111). Using a WGS approach combined with assays evaluating the ability of L. monocytogenes to grow in cold (4°C), salt (6% NaCl, 25°C), and acid (pH 5, 25°C), it was found that the bacterial stress response seems to be related to serotype and clonal complex, among other factors (133).

At the food cooking level, it was shown that E. coli AW1.7 isolated from commercial beef slaughter plants had a D60°C of 71 min, with cell counts reducing by only <5 log₁₀ CFU g⁻¹ in ground-beef patties cooked to an internal 71°C, the temperature considered effective for bacterial elimination in beef (134). This thermal resistance was linked to the acquisition of an island termed locus of heat resistance (LHR) present in only 2% of E. coli genomes, including clinical ones (107, 135). These data suggest the potential inefficacy of standard thermal conditions applied in food safety when particular strains are present.

Uncommon bacterial genotypes occurring through DNA acquisition and recombination events allied to transmission-favorable contexts might also contribute to the emergence of unusual problems. In 2011 in the European Union, thousands of people were infected by MDR and highly virulent STEC O104:H4, linked to sprout consumption, a serotype that rarely caused human infection in the past. WGS revealed an unexpected genetic context of a hybrid enteroaggregative (EAEC)/enterohemorrhagic (EHEC) E. coli strain, potentially supporting the high pathogenicity. The evolution analysis of STEC O104:H4 strains, by the description of the genome and population structure of the outbreak and non-outbreak isolates obtained from sporadic infections, suggested that outbreak STEC O104:H4 might have evolved to public health importance from EAEC by exploiting a specific cocktail of mobile genetic elements (e.g., prophage with stx2, plasmids carrying CTX-M15 or aggregative adherence fimbriae, high-pathogenicity island [HPI], among others), underlining the possibility of further outbreaks if strains achieve novel combinations of mobile genetic elements (136–139).

Finally, the heterogeneous behavior of individual cells within the bacterial population can differentially evolve to cope with the same stress at different levels, or with changing stresses (104, 115, 140). For example, a heterogeneous and multifactorial response to high pressure in S. enterica was described with respect to both degree of inactivation and mechanisms (related to the cell membrane and RpoS regulon) used to overcome it. Parker et al. (140) showed that outbreak E. coli O157:H7 strains isolated from patients and the spinach production environment (all indistinguishable by pulsed-field gel electrophoresis [PFGE]) differ in their stress responses, and that marketed spinach bags carried a mix of both subpopulations. Clinical strains carrying the mutated rpoS gene were more susceptible to acid, osmotic shock, or oxidation than environmental ones with wild-type rpoS. However, clinical strains had a greater nutrient-scavenging ability, potentially favoring survival and rapid adaptation of E. coli O157:H7 at critical points of its transmission cycle from “field to fork.”

Interplay between Bacterial Strains and Food Matrix Favoring Pathogen-Food-Human Transmission Scenarios

Despite the fact that bacterial strain variability could determine its transmission success to the human host, the interplay of strains with supporting food matrices seems to be critical for such an event. Food composition might determine the natural colonization of food with particular strains (e.g., due to specificities of nutrient availability), protect foodborne pathogens from being eliminated during food processing or during their passage into the host gastrointestinal tract (e.g., bacteria can use food molecules to overcome stress), and can induce the expression of virulence features or of cross-protection regulatory pathways dealing with multiple stresses in particular strains (96). Thus, bacterial stress response stimulated by the food matrix could have multiple food safety implications, namely, in the efficacy of hurdle technology used in bacterial growth control (e.g., combination and sequence of stresses applied) (90). Moreover, more opportunities for pathogen-food-human transmission might arise through the consumption of new food vehicles associated with market availability of a variety of food products with specific intrinsic (e.g., different types of lettuce species) or processing factors (e.g., sublethal salt concentrations or low acid content), possibly allowing survival and multiplication of specific strains. Some examples illustrate the interplay between bacterial strains and food matrices that potentially contribute to pathogen-food-human transmission scenarios.

Vegetables constitute variable challenges (based, e.g., on nutrient availability, solar irradiation, and microbiota competition) to foodborne pathogens, which could fit better or worse accordingly with the interplay of strain type, plant species, or even plant age (141). For example, with the use of whole-transcriptome analysis, Crozier et al. (124) observed plant species-specific metabolic responses (e.g., to acid and nutrient stress) when E. coli O157:H7 (Sakai strain associated with a sprout outbreak in Japan) was exposed to lettuce or spinach extracts as well as to different parts of the same plant (leafs or roots). Brandl et al. (142) demonstrated that population sizes of E. coli O157:H7 (lettuce-associated outbreak H1827 clinical strain) and S. enterica serovar Thompson (cilantro-associated outbreak RM1987 clinical strain) were higher in young romaine lettuce leaves (enriched in total nitrogen and carbon) than in middle leaves harvested from mature lettuce heads, positioning younger leaves as vehicles of greater risk to transmit human pathogens. Moreover, several studies found that some foodborne pathogens (e.g., Salmonella, E. coli, and L. monocytogenes) have the ability to internalize in vegetables’ edible parts (e.g., tomatoes and lettuce), impairing their removal by standard disinfection methods before human consumption (143–146). Thus, knowing the interaction of specific strains with particular vegetables might contribute to predictive modeling or risk-based analysis of the potential for microbial contamination, colonization, and persistence in different horticultural crops (each day more often associated with outbreaks) (147) and to better understanding of the transmission routes and vehicles of particular strains.

Food matrices could have an impact on the global stress response of particular strains due to the expression of stress cross-protection mechanisms, often described in foodborne pathogens (91, 103, 117, 148). Poimenidou et al. (149) found that tomato-habituated L. monocytogenes under cold temperatures was more tolerant to acid or osmotic stress than that habituated on lettuce, and lettuce-habituated L. monocytogenes was more tolerant during heat challenge at 60°C compared to tomato-habituated cells. Also, Salmonella resistant to low water activity found in some food types (e.g. flour, chocolate, cocoa and hazelnut shells, and dried milk) was more thermotolerant (99).

It has been suggested that food molecules (e.g., specific amino acids, carbohydrates, urea, and fatty acids) could protect bacteria from stressful conditions (96). Using a simulated gastric passage model (pH 2.5; pepsin; bile salt and enzymes), Aviles et al. (150) showed that a peanut butter outbreak-associated S. enterica serovar Tennessee strain was better able to survive the acid environment when vehiculated in peanut butter matrix with low water activity and high fat content than in low-fat formulations. Also, Birk et al. (151) showed that S. enterica serovar Dublin was better protected when some types of proteins (pepsin, ovalbumin, and turkey meat) were included in a simulated gastric digestion compared with bovine serum albumin, indicating that protection could be protein specific. In a different context, Liu et al. (152) showed that E. coli LTH5807 differed in thermal resistance when heated to 60°C on mung bean, radish, or alfalfa seeds (153).

Bacterial stress protection by the food matrix seems to be also related to low infectious doses, as in the case of salmonellosis epidemiological linked to values of the order of 10 to 100 CFU vehiculated by food with low water activity, compared to >10⁵ CFU in other food types (103). Another explanation for such low infectious doses is that bacteria could be missed by standard microbiological food methods, as dehydration could induce the bacterial filament form (underestimate the true number of cells) or the viable but nonculturable (VBNC) state (impairing bacterial growth). Filamentous bacteria seem to be more acid or bile resistant and retained their virulence in vitro and in vivo in a mouse model (154), and cells in the VBNC state were described in some cases to maintain their pathogenic capacity after resuscitation (103, 155). For example, it was demonstrated that VBNC cells of C. jejuni maintained the ability to adhere to intestinal cells (156) and VBNC cells of E. coli O157:H7 expressed stx1 and stx2 toxin-coding genes and produced the toxin (157, 158). The VBNC state was also described for other foodborne pathogens such as Shigella spp., Yersinia enterocolitica, Aeromonas spp., Brucella spp., and Vibrio spp. (155), being induced by several stresses occurring in the food matrices and during food processing or storage (155, 159). Considering the large number of foodborne outbreaks described without the identification of the etiological agent, implicated food vehicle, or reservoir (13), more knowledge is needed to clarify how often VBNC bacteria could be involved in undetected pathogen-food-human transmission events. For example, the STEC O104:H4 of the sprout-associated European outbreak in 2011 was never isolated from the suspected origin (fenugreek seeds), and it was shown to be able to enter the VNBC state and resuscitate (160).

Finally, the components of food matrices or the processing factors applied to them might have an impact on the expression of foodborne pathogen virulence factors. This is the case for enterotoxins produced by Staphylococcus spp., B. cereus, or STEC, in which toxin expression seem to be dependent on the interplay between toxin type, strain, thermal treatment, and food matrix, although different studies reported disparate data (95, 121, 123,161–167). Thus, the certainty of assessing Staphylococcus or B. cereus food poisoning risk through colony-forming units present in food products has been disproven, with the scientific community and food business operators aware of the threat arising from unforeseeable enterotoxin production under stress conditions.

Interplay between Bacterial Strains and Host Favoring Pathogen-Food-Human Transmission Scenarios

A successful host colonization occurs by the ability of a pathogen to outcompete and outgrow native host microbiota, mostly in the intestinal tract (97, 168). For that, foodborne pathogens must survive specific stressful conditions (e.g., stomach acid, macrophages, and oxidative stress) as well as modulate host microbiota (e.g., changing the equilibrium of particular species), the immune system (e.g., inducing inflammatory reactions), and, indirectly, the gut physicochemical environment (e.g., available nutrients and oxygen tension) (95, 97, 168, 169). A previous contact with stressful conditions in the food chain might allow a preadaptation of the pathogen to those similarly occurring in the host, or even increase its virulence due to coregulation of stress and virulence genes (90, 105, 109). After the colonization step, foodborne pathogens shedding from their hosts vary in the amount of bacteria expelled and in the duration that shedding occurs, resulting in diverse levels of environmental, animal carcass, or vegetable contamination, with impact on pathogen transmission dynamics (168, 170). On the other hand, the interplay between bacterial genetic evolution and environmental challenges imposed by the specific physiology of each host (e.g., related to the variable gut environment in different animal reservoirs, changing diets, or antimicrobials used in intensive animal production) might increase bacterial shedding from particular individuals and drive strains’ evolution pathways toward specialist (e.g., one host) or generalist (multiple hosts) lifestyles (5, 168, 171). A selection of generalist strains conducts to greater transmission opportunities among a variety of reservoirs and hosts (172). In contrast, specialist strains have fewer opportunities to be transmitted but usually cause more severe, chronic and asymptomatic infections, with the possibility of long periods of host shedding, sometimes undetected (5). Some examples illustrate the interplay of strain-host scenarios potentially contributing to foodborne pathogen-food-human transmission.

C. jejuni is a major foodborne human pathogen vehiculated mainly by contaminated poultry (13, 173). C. jejuni can highly colonize poultry (up to 10¹⁰ CFU/g feces), and it is suggested that the high body temperatures (41 to 45°C) of poultry species, the poultry gut environment, and the inefficiency of poultry immune system activation contribute to the persistent colonization of the avian gut with thermotolerant and microaerophilic C. jejuni and, consequently, to its continuous shedding and transmission to the environment and other poultry hosts (173, 174). Also, for STEC O157:H7, individual host features allied to strain and environmental factors could determine the level of host colonization and shedding. Cattle are pointed to as the main reservoir of STEC O157:H7, with some animals classified as “super-shedders” when they release ≥10⁴ CFU/g of feces (175). Strain specificities (e.g., ability to colonize the rectal-anal junction of cattle), animal features (e.g., composition of the cattle native microbiome), and animal age and diet seem to determine whether E. coli O157:H7 can proliferate sufficiently for the host to obtain super-shedder status (168, 175–177). The supper-shedding state or a prolonged release period of bacteria in the feces has also been associated with consumption of antibiotics with impact on the host microbiota. This was detected for S. Typhimurium in mouse infection models, releasing bacteria on the order of >10⁸ CFU/g, or in outbreaks of S. Typhimurium from roast pork, in which antibiotic consumption by patients increased the carrier state period (125, 178). Prolonged human symptomatic infections (from weeks to several years) with other NTS serotypes (S. enterica serovars Mbandaka, Bredeney, Infantis, and Virchow) have also been described, with antibiotic consumption also cited as one of the factors contributing to their persistence (170).

High host shedding of foodborne pathogen strains allied to favorable transmission patterns might increase opportunities to colonize multiple available hosts (e.g., several animal species), which depend on bacteria making rapid adjustments to each new host ecosystem. However, the high prevalence and wide distribution of the so-called “generalist clonal lineages” complicates source attribution, as in the case of C. jejuni ST21/ST45 clonal complexes isolated from poultry, cattle, and infected humans (172). In spite of this, Sheppard et al. (179) showed that within ST21/ST45 clonal complexes exist sublineages with host association, determined by the occurrence of the panBCD genes encoding the vitamin B5 biosynthesis pathway, more prevalent in cattle (diet composed of grass with low vitamin B5) isolates than in chicken (diet composed of cereal and grains abundant in vitamin B5) ones. The discovery by WGS of gene loci gained or lost makes genetic elements possible targets for source tracking, with the genetic makeup of a human isolate theoretically helping in the potential identification of the isolate source and of transmission food vehicles (180). Generalist and specialist lifestyles could also be related to Salmonella serotypes, with S. Typhi (specialist serotype) associated with human reservoirs and infections, and the main emerging S. Typhimurium and S. Enteritidis (generalist serotypes) associated with human infections and colonization of diverse animals (13, 181). It remains to clarify whether generalism is a stable strategy taking advantage of the large number of transmission opportunities in agriculture or whether it reflects insufficient time for well-adapted host specialist lineages to have evolved in the recent man-made environment. For example, studies on S. Typhimurium ST313, causing invasive infections in immunocompromised populations in Africa, point to the presence of degraded genome capacity in the form of pseudogenes and deletions, suggesting evolution from a generalist to a specialist lifestyle, with transmission among humans potentially exerting genomic selection pressure (181). Understanding the differences of foodborne pathogen lifestyles is important to establish effective control measures adapted to the context of specific hosts as well as to preview future transmission scenarios of specific foodborne strains through particular reservoirs and food types.

ANTIMICROBIAL-RESISTANT FOODBORNE BACTERIA TRANSMISSION ALONG THE FOOD CHAIN

Antimicrobial resistance is considered a major global public health issue for humans and animals, and a major priority in food safety by a number of entities (the European Commission, WHO, and Food and Agriculture Organization of the United Nations) (182–186). According to Centers for Diseases Control and Prevention (CDC) estimates, antibiotic-resistant infections caused 23,000 deaths per year in the United States, and about one of five resistant infections were caused by microorganisms from food and animals (185, 187–189). Nevertheless, the contribution of food to the transmission to humans of bacteria carrying clinically relevant resistance genes is still poorly established, including the real enrichment of the human microbiota with such bacteria and genes.

Although in recent years great effort has been made by different organizations (e.g., the WHO, EFSA, and CDC) to understand the magnitude and burden of antimicrobial resistance transmission through the food chain, it remains uncertain at the global level (190). It is particularly underestimated in low- and middle-income countries where drivers of antimicrobial resistance (e.g., unregulated farming and food-production practices concerning food safety and antibiotic use; poor hygienic conditions, namely, related to water and sewage; and human travel and migration) are significant, eventually affecting the spread of antimicrobial resistance by the food chain in a wide geographic area (189, 191). The regular use of antimicrobials in livestock (therapeutic, metaphylactic, prophylactic, or growth-promoting use) associated with modern intensive food-animal production has been considered the main driver of the selection of both antibiotic-resistant bacteria and genetic elements worldwide, with food products (e.g., meat, poultry, eggs, vegetables, and farmed fish) the most relevant transmission mode to humans (186, 188, 189, 192). In fact, new WHO guidelines on the use of medically important antimicrobials in food-producing animals stressed that antibiotic-resistant bacteria in this sector were reduced by up to 39% after interventions restricting antibiotic use at the farm level (186). However, other nonantibiotic compounds with antimicrobial activity and widely used in food-animal production (e.g., copper) may also contribute to the selection of antibiotic-resistant zoonotic bacteria and to less successful intervention scenarios related to reduction of antibiotic use (101, 102, 188, 193).

Antimicrobial-resistant bacteria with human relevance include pathogenic zoonotic organisms (e.g., Salmonella and Campylobacter) as well as likely commensal zoonotic bacteria (e.g., E. coli and Enterococcus). Both groups have been recognized as important or major contributors to the burden of antimicrobial resistance (188, 192, 194). Antimicrobial-resistant zoonotic pathogenic bacteria may be associated with more-prolonged infections; treatment failures as a result of invasive infections; or the use of last-line antimicrobials for therapy, which are more expensive and/or toxic. Commensal zoonotic bacteria could be reservoirs of clinically relevant antibiotic resistance genes mobilizable to pathogens and be potentially involved in extraintestinal opportunistic infections (e.g., urinary tract or bloodstream infections) after transmission by food and colonization of humans (Fig. 3) (190, 194). In general, these commensal strains with likely animal origin are widespread in multiple hosts and often participate in genetic exchange events with other microorganisms sharing the same ecosystems, making direction of transmission (e.g., food-human, environment-human, or human-human) hard to assess (189, 190).

FIGURE 3.

Implications of antibiotic-resistant bacteria transmitted from food to humans. AB^R, antibiotic resistance.

For several decades, the contribution of food-animal reservoirs and food vehicles in the transmission of antimicrobial-resistant bacteria relevant to human health has been controversial. However, accumulating evidence linking livestock production with antimicrobial resistance burden in humans has been reported (188–190, 192, 195), which we classify in three categories.

Association between the Use of Antimicrobial Agents and the Occurrence of Antimicrobial-Resistant Bacteria in Food-Producing Animals and/or Humans

The correlation between antibiotic use and the occurrence of antimicrobial resistance in bacteria isolated from the food chain and humans is evident in studies from the last 25 years, including those of zoonotic pathogens or commensal bacteria. Among the more illustrative cases was the consequences of licensing the fluoroquinolone enrofloxacin for animal use in the 1990s, especially in poultry. It led to increased rates of ciprofloxacin resistance in S. Typhimurium DT104 recovered from animals/food and humans in the United Kingdom (196) and of C. jejuni from humans in the United States (197) and from chickens and humans in The Netherlands (198). More recently, a voluntary withdrawal of ceftiofur by poultry producers in Canada was correlated with decreasing occurrence of ceftiofur-resistant S. Heidelberg (one of the most common serotypes associated with salmonellosis in this country) from both human infections and retail poultry. After reintroduction of ceftiofur 2 years later in poultry production, an increase in resistance levels in poultry and humans was once more observed (199). Another example was the ban of the glycopeptide avoparcin as a growth promoter in Europe in the 1990s, which led to the reduction of vancomycin-resistant Enterococcus in community settings (producing animals and healthy humans) (200–203). Evidence on the correlation between antibiotic consumption and resistance among foodborne pathogens from humans and food-producing animals was also corroborated by the last European Centre for Disease Prevention and Control (ECDC)/EFSA/European Medicines Agency (EMA) joint report involving univariate/multivariate analysis in almost 30 European countries. It demonstrated that resistance to fluoroquinolones in Salmonella and Campylobacter and resistance to macrolides in Campylobacter coli from humans were related to the consumption of fluoroquinolones or macrolides, respectively, in animals (204). Previous findings have also suggested a link between S. Enteritidis resistance to nitrofurans and their illegal use in food-producing poultry in Portugal (205).

Correlation between Rates of Antibiotic Resistance among Bacteria from Food-Producing Animals, Food, and Humans Obtained from Systematic Surveillance Data

The 2016 EFSA reported that resistance to widely used antimicrobials (including critical ones important for the treatment of severe human infections) was commonly detected in Campylobacter and Salmonella from humans and poultry. This is illustrated by the high level of resistance to ciprofloxacin in C. jejuni isolated from broilers (69.8%), broiler meat (65.7%), and humans (60.2%). In the case of Salmonella, it was reported that resistance to ciprofloxacin was markedly higher in some serotypes commonly associated with poultry (S. Enteritidis, S. Infantis, and S. enterica serovar Kentucky), suggesting a greater contribution of the poultry production chain to the human burden (183). Moreover, a high prevalence of MDR Salmonella in humans (26%) and poultry meat (broiler, 24,8%; turkey, 30.5%) (including the ampicillin/streptomycin/sulfonamides and tetracycline resistance [ASSuT] profile) was also described (183). Data from the National Antimicrobial Resistance Monitoring System (NARMS) reported an increase of S. 4,[5],12:i:- with the ASSuT-MDR phenotype in humans, from 43 to 60% between 2014 and 2015, and of 65% in swine from 2015 (206).

In low- and middle-income countries (e.g., in Asia, African, and Latin America), the rates of antibiotic resistance, including to clinically relevant antibiotics (ciprofloxacin, extended-spectrum cephalosporins, colistin, and carbapenems), are undoubtedly higher than in high-income countries, with food products presenting a potential role in its emergence (189). For example, high levels of Salmonella nonsusceptible to ciprofloxacin (15 to 48%) and cephalosporins (38% to ceftriaxone) were observed in humans from Asian countries (207, 208). These data are in agreement with several studies documenting a high prevalence of resistance to fluoroquinolones (>22.5% to ciprofloxacin) and cephalosporins (12.5 to >23.4% to ceftriaxone and 26.6% to ceftazidime) in Salmonella from poultry meat obtained in South Korea and China (209, 210), as well as in E. coli from farmed fish (36.7% reduced susceptibility to ciprofloxacin) in China (211). Also, a recent Chinese surveillance study revealed high rates of the new plasmid-mediated colistin resistance gene mcr-1 in diverse foodborne bacteria recovered from water (71% of samples), animal feces (51%), food products (36%, mostly meat), and humans (28% of human subjects surveyed), further supporting the role of the food chain in the transmission of colistin-resistant bacteria and the mcr-1 gene to humans (212). However, direct zoonotic transmission of mcr-1 could not be excluded, as stated by the association of colonization of farmers with mcr-1-carrying bacteria with exposure to mcr-1-positive chickens from small-scale poultry farms in Vietnam (213).

The role of low- and middle-income countries as largely exporters of food to different geographic regions of the world, their changing from extensive farming systems to large-scale ones with a higher use of antibiotics due to consumer demand for protein, and the largely unregulated use of antibiotics in animal production suggest that international trade of contaminated breeding animals (e.g., poultry production depends on a pyramid-like breeding system), feed, and meat products with antibiotic-resistant foodborne pathogens may contribute to the rapid worldwide spread of these bacteria, with impact on human health (214). The last DANMAP (Danish Integrated Antimicrobial Resistance Monitoring and Research Programme) report described higher levels of extended-spectrum β-lactamase (ESBL)/AmpC-producing E. coli (56% of samples, including ST131 clone) and ciprofloxacin resistance in C. jejuni (71% of isolates) from imported broiler meat compared with domestically produced broiler meat (23% of the samples with ESBL/AmpC-producing E. coli and 22% of C. jejuni isolates with ciprofloxacin resistance) (215). Recently, Brazilian imported poultry meat contaminated with extended-spectrum cephalosporin-resistant bla_CMY-2-producing S. Heidelberg (a poultry-related serotype uncommon in Europe) was reported in Europe and included strains with indistinguishable PFGE profiles from epidemic clones with invasive potential that caused outbreaks in the United States, alerting for potential risk to human health (60, 61). The spread of E. coli carrying bla_CMY-2 from flocks of imported broiler parents to broiler meat (including potentially human-pathogenic types), even in a country with no cephalosporin use (216), suggests that the globalization of trade of food animals and food products is an important driver of antibiotic resistance spread through the food chain.

Transmission of Clinically Relevant Antibiotic Resistance Genes on Mobile Genetic Elements and/or Antibiotic-Resistant Clones of Zoonotic Pathogens and Commensal Bacteria from Producing Animals and Food to Humans

Transmission of antibiotic-resistant clones and genetic elements between bacteria (pathogens and commensal) from the food chain to humans has been described in recent decades (189, 191, 195, 217, 218). With the evolution of molecular methods (WGS versus multilocus sequence typing [MLST] and PFGE), the accuracy of such linkages may be clearer in the near future as data accumulate, by facilitating the identification of new cases as well as confirming or contesting older ones. Table 1 shows examples illustrating linkage and transmission of Salmonella, E. coli, and Enterococcus spp. clones from the food chain to humans. It includes studies developed in diverse time frames that established such linkages. Other examples including only mobilization of genetic elements are discussed in the paragraphs below.

TABLE 1.

Antibiotic-resistant pathogens and commensal zoonotic bacteria recovered from food products and humans^a

Bacteria	Clonality approach	Clinically relevant antibiotic resistance gene(s)	Genetic elementPL – Inc group	Food source (human source)	Country(ies)/year(s)	Reference(s)
Salmonella
S. 1,4,[5],12:i:-	PFGE	MCR-1	PL-X4, HI2	Pork products (clinical cases)	Portugal/2011–2015	222
S. Enteritidis	PFGE, phage typing	ESBL-CTX-M-15	PL – FII	Chicken meat, chicken feces, or diseased chicken (clinical cases)	Korea/2009	245
S. Heidelberg	WGS-based hqSNV	AmpC-CMY-2	PL – I1	Retail poultry, abattoir poultry (clinical cases)	Canada/2012	220
S. Heidelberg	PFGE, MLST	AmpC-CMY-2	PL – I1, A/C	Imported poultry meat (clinical cases)	Portugal/2014–2015; The Netherlands/1999–2013	60, 61
S. Heidelberg	PFGE	AmpC-CMY-2	PL – I1	Chicken abattoir, chicken retail, bovine and porcine (clinical cases)	Canada/2002–2004	246
S. Indiana	PFGE	ESBL-CTX-M-65, CTX-M-14; PMQR-AAC (6′)-IB-CR+OQXAB	PL-NR	Chicken (clinical cases)	China/2009–2013	224
S. Infantis	PFGE, MLST, WGS-based SNP	ESBL-CTX-M-1, CTX-M-65	PL-I1+P	Broiler meat, broiler chickens (clinical cases)	Italy/2011–2014	219
S. Infantis	PFGE	ESBL-TEM-52	PL – I1	Poultry (clinical cases)	Belgium/2001–2005	247
S. Kentucky (ST198-X1-SGI1)	PFGE, MLST	CipR-		Poultry and turkey meat, seafood, broilers (clinical cases)	Europe, Asia/2000–2013	221
S. Typhimurium	PFGE	ESBL-CTX-M-2	PL – NR	Poultry (clinical cases)	Brazil/2003–2004	248
	PFGE	ESBL-CTX-M-1	PL – N	Swine meat (human)	Germany/2007	249
S. Virchow	PFGE, MLST	ESBL-CTX-M-15	PL-HI2	Poultry meat, pig farms (clinical cases)	Korea/2012	98
	PFGE	ESBL-CTX-M-2	PL – NR	Poultry (clinical cases)	Belgium-France/2000–2003	250
	PFGE	ESBL-CTX-M-9	NR	Broilers (clinical cases)	Spain/2000–2004	251
Escherichia coli
A-ST-10, ST-117	MLST	ESBL-diverse		Chicken meat, other meats (healthy humans-fecal, and human blood)	The Netherlands	228
O25:H4-B2-ST131	MLST, PFGE	ESBL-CTX-M-9		Chicken meat and avian (human infections)	Spain/1993–2010	252
A-ST-10, ST-117	MLST	ESBL-CTX-M-1, TEM-52	PL-I1 (ST3, ST7, ST10)	Poultry (blood and urine isolates)	The Netherlands/2006–2010	230
A, B1, D	MLST, AFLP, PFGE	ESBL	PL	Chicken meat (human carrier or blood isolates)	The Netherlands/2008–2009	227
ST38	WGS-based SNP; MLST	AmpC-CMY-2	PL-K	Chicken meat (clinical isolates-urinary tract infections)	Norway/2012–2014	232
Enterococcus
Enterococcus faecium	PFGE	aac(6′)-Ie-aph(2″)-Ia	NR	Poultry carcasses (healthy human feces)	Portugal/2001	233
Enterococcus faecalis	PFGE	aac(6′)-Ie-aph(2″)-Ia	NR	Poultry carcasses (healthy human feces)	Portugal/2001	233
	PFGE	aac(6′)-Ie-aph(2″)-Ia	NR	Poultry carcasses (outpatients’ feces)	United States/1998–2000	234
	PFGE	aac(6′)-Ie-aph(2″)-Ia	NR	Ground pork (healthy human feces)	United States/1998–2000	234

^a Abbreviations: AFLP, amplified fragment length polymorphism; hqSNV, high-quality core genome single-nucleotide variant; NR, not reported; NT, plasmids that were not typeable with the scheme used; PL, plasmid.

In several European countries, multiple clones of pathogenic Salmonella serotypes ESBL producers (e.g., CTX-M-2, CTX-M-9, and TEM-52) shared between poultry and human cases have been reported (Table 1). A recent example describes an MDR and ESBL-producing (CTX-M-1) clone of S. Infantis causing human infections and isolated from poultry (including meat) in Italy during 2001 and 2014 (Table 1) (219). In Canada, the transmission of S. Heidelberg (one of the top MDR poultry-related serotypes in North America) between abattoir, retail poultry, and humans as well as the transmission of a common CMY-2 plasmid among those strains was suggested (220). Also of concern is the dissemination of the ciprofloxacin-resistant S. Kentucky ST198 strain (with a reservoir in northern Africa [Egypt]) in Europe and elsewhere since 2010, including that associated with an increasing number of contaminated sources, mainly poultry but also seafood (221). A possible transmission from pork products to humans of mcr-1 colistin resistance observed in an MDR and copper-tolerant clinically relevant S. 1,4,[5],12:i:- clonal lineage was suggested in a Portuguese study (222). Other studies also point to the food chain as an important reservoir of mcr-1-bearing plasmids transmitted among different bacterial clones and species (213, 223) and suggest that mcr-1 first arose in animals before being transmitted to humans, due to colistin use in food-animal production (152). Other descriptions of shared genetic elements include identical oqxAB-carrying plasmids carried by Salmonella and E. coli and identified in food products and humans, particularly in China (177), with a few reports of shared clones (224).

In relation to commensal zoonotic bacteria, evidence of antibiotic resistance transmission between animals/food and humans is harder to establish, with some studies reaching different conclusions (195, 225, 226). Those suggesting such transmission through the food chain were based on the detection of common clinically relevant genes, mobile genetic elements (e.g., plasmids), and/or clones between food and humans (including those associated with infections). For example, transmission from livestock and/or retail meat to humans of ESBL and AmpC β-lactamase genes on plasmids (including plasmids shared with pathogenic bacteria like Salmonella) and/or of E. coli clones has been suggested (195, 218, 226–232). The last DANMAP also reported the observation of phylogenetically related (WGS-based single-nucleotide polymorphism [SNP]) ST131 bla_CMY-2-producing E. coli from imported meat and human bloodstream infections in Denmark during 2015-2016, thus suggesting a potential zoonotic link between imported broiler meat and severe human infections (215). Among Enterococcus spp. the same clones of a gentamicin-resistant E. faecalis or E. faecium were found in the feces of healthy humans and in poultry carcasses for human consumption in Portugal (233). Clonal linkage between poultry or ground pork for human consumption and humans was also detected in the United States (234). Freitas et al. (235) detected the same MDR, vancomycin-resistant E. faecium clone (CC5-PFGE A) in a piggery environment, feces from live and slaughtered swine, healthy human feces, and hospitalized patients, strongly suggesting its transmission across the food chain in Europe and the United States between 1995 and 2008. Tn1546-vanA with the same structure and shared by humans and food animals was also described (236, 237).

More recently, new alerts of public health threats, but those with indefinable risks for human health, linked to food animals and food products have been observed. This is the case for carbapenem-resistant pathogenic or commensal MDR bacteria (e.g., VIM-1-producing Enterobacteriaceae and NDM-1- or OXA-23-producing Acinetobacter spp.) isolated from food-animal farms and food products of animal origin (217, 231, 238, 239) and from other potential new food vehicles like seafood and produce (239–242). Also, the potential risk of transmission of MRSA carrying different antimicrobial resistance and virulence genes through the food chain cannot be ignored (238). Evidence of food-to-human transmission of livestock-associated MRSA was recently described, with broiler chicken and turkey meat implicated as probable vehicles of the hybrid CC9/CC398 lineage (243, 244). However, more data will be critical to clarify the contribution of the food chain to the transmission of carbapenem-resistant bacteria and MRSA to humans.

CONCLUSIONS

Transmission of bacterial pathogens from food to humans has a relevant impact on public health, aggravated by the acquisition of antimicrobial resistance. Despite the improvements in food safety measures in different parts of the world, the modern era imposed a plethora of drivers for transmission of bacteria to humans through food consumption. The multiplicity and interplay of drivers related to the intensification, diversification, and globalization of food production; consumer health status, lifestyles, and behaviors; and bacterial adaptation to different challenges from farm to human make the prevention of bacteria-food-human transmission a modern and continuous challenge. To minimize the selection and spread of foodborne zoonotic pathogens and commensals, including MDR, it is critical to improve biosecurity measures from the farm (e.g., high hygiene standards and vaccination) and throughout the food chain (e.g., slaughtering and food processing, food handling, and education of consumers); to control antibiotic use, and the live-animal and food trade; as well as to explore bacterial adaptation mechanisms to food production ecosystems. A global One Health approach is mandatory to better understand and minimize the transmission pathways of human pathogens through the food chain (8, 189).