Campylobacter jejuni is a causative agent of bacterial enteritis in humans with serious consequences for children, the elderly and the immunocompromised ^[1][2][3][1,2,3]. One of the rare sequelae associated with C. jejuni is Guillain–Barré syndrome (GBS), manifested as demyelination of the peripheral nerves due to autoimmune reaction and paralysis ^[4][5][4,5]. About 34%–49% (≈42%) of GBS cases are likely to be associated with prior Campylobacter infection [6]. Further post-campylobacteriosis sequelae are reactive arthritis, Erythema nodosum, and inflammatory bowel disease ^[6][7][6,7].

The avian reservoir is the predominant source of C. jejuni infections, which have been strongly linked to contaminated retail chicken by numerous studies ^[8][9][8,9]. As a commensal organism with high prevalence (10⁹ CFU/g) in chicken ceca, C. jejuni easily contaminates carcasses of slaughtered birds [10]. Davis and colleagues demonstrated experimentally that Campylobacter survives well on both chicken skin and meat at refrigerated temperatures [11].

Cases of C. jejuni-associated food poisoning have been rising worldwide. Studies from Brazil, Great Britain, Canada, Italy, Korea, Australia and Japan report contamination rates ranging from 20% to 80% ^{[12][13][14][15]}[12,13,14,15]. According to the European Food Safety Authority (EFSA), Campylobacter continues to be the most commonly reported human gastrointestinal bacterial pathogen in the European Union (EU) since 2005 [16]. C. jejuni infections occurring each year cause close to 350,000 cases of disability and cost 2.4 billion Euros yearly [17]. There were 246,571 confirmed cases of Campylobacteriosis reported in the EU in 2018 [18].

Currently, there is no single effective intervention method for reduction of bird colonization by C. jejuni. Biosecurity measures only achieve partial clearance [19]. This was effectively demonstrated by a study conducted in central Italy where Campylobacter isolates were obtained, both by cloacal swabs and at the post-harvest stage, from farm chicken flocks kept under good sanitary practices [20]. Colonization of broiler chicks by C. jejuni takes place within the first three weeks after hatching, and, since Campylobacter colonization is not associated with signs of disease in chickens, horizontal spread of the pathogen usually remains unnoticed ^[9][21][9,21]. Combination of strict biosecurity, good manufacturing practice (GMP), hazard analysis and critical control points (HACCP) attempts can alleviate contamination levels but do not lead to a complete elimination of the bacteria. It is difficult to quantify the effectiveness of individual measures because they are dependent on very many interrelated local factors [17]. Additionally, there is no vaccination available against C. jejuni due to serological diversity of the pathogen and the short lifespan of broilers [19]. The estimated public benefits in terms of a higher reduction of the disease burden of campylobacteriosis and its sequelae are greater if efforts are made towards controlling C. jejuni during the primary production stage [17].

To control C. jejuni, scientists have proposed several methods based on either supplementing chicken feed with probiotic bacteria capable of inhibiting C. jejuni and/or administering Campylobacter-specific phages to chickens ^[22][23][24][22,23,24]. Campylobacter phages are natural “predators” that could potentially be able to bring C. jejuni under control. Phages are already being used to control other food-borne pathogens, such as Salmonella spec. and Listeria monocytogenes ^[25][26][25,26]. However, despite their historical use in typing Campylobacter spec., there are limited number of reports about application of Campylobacter phages for therapeutic purposes ^[27][28][27,28].

2. Classification of C. jejuni Phages and Receptor Specificity

Phage typing has been used extensively for epidemiology studies of C. jejuni and Campylobacter coli [29]. The number of C. jejuni bacteriophages reported to date exceeds 170 and the majority of these phages exhibit a narrow host range [13]. Most Campylobacter phages are lytic, i.e., virulent or able to lyse the host bacteria upon infection and are thus preferred due to their immediate effect [30]. Lysogenic or temperate phages are generally avoided for their ability to integrate into a bacterial genome and transduce virulence genes from strain to strain ^{[31][32][33][34]}[31,32,33,34]. There are several ways to classify phages. Generally, phage classification schemes are based on their morphology, genetic makeup (e.g., DNA vs. RNA) or genome size (Table 1) [35]. Morphologically, the majority of C. jejuni phages are categorized into the family Myoviridae (Bradley’s morphotype A1, contractile tail), while some have been designated as Siphoviridae (Bradley’s morphotype B1, non-contractile tail) [36]. Lytic Campylobacter phages are divided into three groups according to the size of their genome: group I contains relatively rare phages with large genomes (approx. 320–425 kbp), while groups II and III include most Campylobacter phages isolated and characterized to date [37]. These are all distantly related to T4 bacteriophages and have smaller genomes: 175 to 183 kbp (group II) and 131 to 135 kbp (group III). Group II and III Campylophages exhibit resistance to many restriction endonucleases. These phages are characterized with great similarities within each group but differences between the groups ^[37][38][37,38]. Group II Campylophages are often specific to both C. jejuni and C. coli, while group III phages demonstrate specificity exclusively to C. jejuni and have stronger lytic activity [37]. The latest phage classification scheme, which is based on whole genome sequencing and protein homology, puts group II and III phages into the Eucampyvirinae sub-family of myoviridae as “CP220-like viruses” and “CP8-unalike viruses”/“CP8 virus” ^[37][38][39][37,38,39]. Sørensen and colleagues showed, with some exceptions, the connection between receptor dependency and grouping of C. jejuni phages. For example, the majority of group III phages use capsular polysaccharide (CPS) receptors, while group II phages target mostly flagellar receptors ^[27][40][27,40].

3. Phage Applications Targeting Food-Borne Pathogens

The Earth’s biosphere is estimated to contain approximately 10³² phage particles, exceeding the number of prokaryotic organisms by 10-fold ^[33][41][42][33,41,42]. Bacteriophages differ in size and shape, and are characterized with incredible genome mosaicity resulting from different environmental pressures, including the adaptation to continuous bacterial resistance [43]. At the time of writing this article, only 8852 complete genomic sequences have been listed in the NCBI database of phage genomes. Thus, numerous novel phage genes, potentially leading to development of new phage-based drugs, are yet to be discovered and studied. Bacteriophages were successfully used for human therapy almost immediately after their discovery in the beginning of the 20th century [41]. After the triumph of penicillin in 1940s, phage use gradually dwindled in the West, but continued in Georgia (the former USSR) and Poland [44]. The Eliava Institute of Bacteriophage, Microbiology and Virology founded by George Eliava and d’Herelle in Tbilisi in the 1930s still continues its work and has accumulated extensive experience in phage work and phage therapy over the decades. The advent of the era of multi-drug resistant bacteria has rekindled the interest in bacteriophages due to several major advantages of phages over antibiotics. First, bacteriophages are effective regardless of the susceptibility of their host bacteria to antibiotics, i.e., antibiotic-resistant bacterial strains are affected in the same way and with the same effectiveness as non-resistant strains. Second, phages are ubiquitous and relatively easy to isolate. Third, their narrow specificity is instrumental in avoiding the disruption of the entire microbiome in the treatment subject [45]. The fourth advantage is that new, mutant, phages may be developed in a matter of days [46]. Phage-based applications can be used as an alternative to antibiotics in live animals, in situ on processed meat and other foods, and for decontamination of surfaces and equipment in food processing to ensure that no multi-drug resistant bacterial strains selected as a result of antibiotic-based evolutionary pressure are released into the wastewater and thus into the environment [25]. Thus, bacteriophage application is manifold and covers different stages of primary (live animals) and secondary (harvested meats) production in animal farm settings. Having no associated toxicity, phage treatments can be complexed with other methods, such as probiotic supplements, aimed at reduction of C. jejuni ^[1][4][45][1,4,45]. Humans have been continuously exposed to phages via drinking water and fresh foods without any adverse reaction ever recorded [24]. In fact, phages have been isolated from human saliva and intestines [47]. A gram of human feces contains approximately 10⁸–10⁹ virus-like particles comprised mostly of DNA phages ^[48][49][48,49]. Based on the evaluation of experts qualified by scientific training and experience, the United States Food and Drug Administration (FDA) recognized phages and phage derivatives as generally regarded as safe (GRAS) through the 1958 Food Additives Amendment to the Federal Food, Drug, and Cosmetic Act ^[50][51][50,51]. A phage manufacturer, however, has to provide proof to the FDA in the form of a GRAS notification indicating the intended use [26]. For example, the GRAS notification for ShigaShield^TM—a phage product developed by Intralytix against Shigella—specifically states that the GRAS status has been determined “through scientific procedures” [52]. Although most phages are harmless to human health and environment, not all phages are safe, as some lysogenic phages can actually carry and transduce virulence genes. For example, E. coli O157:H7 and Streptococcus pyogenes owe their virulence to the acquisition of phage virulence genes into their genomes [33]. To avoid “phage lysogenic conversion”, i.e., changing the properties of specific bacteria, lytic phages must be used, as they immediately lyse the host and do not integrate into the bacterial host genome. For assuring the safety of phage preparations used in animal and human therapy, the latest achievements in sequencing technology are utilized. Bioinformatics methods can rapidly predict undesirable phage properties. The GRAS designation and the FDA approval of phages P100 and LMP-102 were specifically based on this approach following phage sequencing. Advances in phage research have increased the interest in their possible use as biocontrol agents to preserve or decontaminate food products and, thus, prolong their shelf life while preventing frequent disease outbreaks. According to the US CDC, 841 foodborne disease outbreaks were reported by 50 states, Washington, D.C., and Puerto Rico, resulting in 14,481 illnesses, 827 hospitalizations, 20 deaths, and 14 food recalls in 2017 [53]. The major foodborne bacterial pathogens most frequently associated with lethal outcomes are L. monocytogenes, S. enterica and E. coli (e.g., E. coli O157:H7), followed by C. jejuni ^[26][54][26,54]. Naturally, these food-borne pathogens became the focus of the first phage-based applications developed. In dairy production, for example, S. enterica contamination can occur at virtually any stage and phages can be used to reduce the shedding of Salmonella in farm animals, as a food additive in post-harvest food products, and as biocontrol agents in food processing to control the pathogen [25]. Phage therapy against Salmonella has been successful in broiler production and significantly reduced the pathogen in the ceca of broilers [25]. The general recognition of phage applications as GRAS also prepared the ground for regulatory clearance of several phage-based biocontrol products targeting L. monocytogenes, S. enterica and E. coli. The first such product, ListShield^TM (LMP-102) from Intralytix targeting L. monocytogenes “ready to eat” (RTE) foods, was approved by the US FDA in 2006. A similar biocontrol agent against L. monocytogenes, Listex P100, got the FDA approval for the use in meat, cheese and other foods, including fish, shellfish, fresh fruits and vegetables. Thereupon ListShield^TM was approved in Canada and Israel, while Listex has been approved in Switzerland for use in cheesemaking. Approvals for further phage applications followed: in 2007, the FDA approved a product from Omnilytics for decontamination of live animals from E. coli and Salmonella. Within the same year, the FDA followed up with the approval of new phage applications, including the “Finalyze” spray against E. coli O157:H7 in cattle (Elanco Food Solutions) and “Armament” against Salmonella in poultry. Another phage application from Intralytix-EcoShield, 95–100% effective against E. coli O157:H7, received the FDA’s regulatory approval in 2011 for use on red meats prior to grinding [30]. Up to the present time, the FDA’s GRAS inventory includes phage applications against Shiga-like toxin-producing E. coli, L. monocytogenes, Salmonella and Shigella. None of the listed products target Campylobacter spp., although Intralytix and Micreos have publicly expressed their interest in pursuing the development of such applications [24]. To date, there are only very few patents on phage products relating to their use for Campylobacter germ load reduction in poultry flocks or on processed meat. Instead of a complete bacteriophage, Fischetti and colleagues patented a process that uses a bacteriophage-derived lysis enzyme for bacterial decontamination of food products [55]. This bacteriophage-derived lysis enzyme can be increased in its effectiveness of bacteriolysis by modification, e.g., by construction of chimeric lytic enzymes, shuffled lytic enzymes or by additional holin proteins. This patent also includes Campylobacter spp. in the list of potential target organisms. However, a specific bacteriophage-derived lysis enzyme was not named in the 2002 patent [56]. In the 2004 version, the bacteriophage-derived lysis enzyme PaI with activity against Streptococcus pneumoniae and other streptococci of the Viridans group were also listed [55]. Based on their in vivo study, Connerton patented the application of the bacteriophages CP8 and CP34 [2] to reduce Campylobacter spp. in the intestine of birds [57] (for further details see Section 4.1, study II and III, as well as Table 2). Burnett and coworkers patented a procedure to reduce or to prevent bacterial contamination of any type of food product using bacteriophages [58]. The bacteriophages are to be administered in a specific embodiment, which may contain a buffer, surfactant, adjuvants and enhancers to prevent degradation of the bacteriophage and even to enhance its performance as an antibacterial agent. Application of the bacteriophages can also be done on non-food surfaces or water systems. Besides the shown example, L. monocytogenes, Campylobacter spp. and other bacteria are listed as possible target organisms for this procedure. The patent of Ter Haar and Hanna complements the application of bacteriophages to food by a rubbing process using a vibratory conveyor to distribute the bacteriophage more efficiently on the food surface [59]. The European Food Safety Agency (EFSA) approached the usage of phages as biocontrol agents rather carefully and issued three scientific opinions in 2009, 2012 and 2016. While the 2009 opinion focused on the nature of bacteriophages and their potential use for decontamination of foods, the 2012 opinion was a response to Micreos’ application for the approval of LISTEX^TM P100 to reduce L. monocytogenes [60]. This statement addressed safety and efficacy issues of the phage product, but concluded that the product was safe, although the agency had several concerns, specifically on (i) the absence of industrial scale studies, (ii) the limited number of phages used, and (iii) the lack of evidence of significant pathogen reduction. The 2016 opinion finally recognized the safety and efficacy of LISTEX^TM P100 for the use on meat, poultry, fish, shellfish and dairy products at up to 1 × 10⁹ PFU [60].