Development Toward a Bovine Salmonella Model and a Bacteriophage Treatment Cocktail to Reduce Salmonella Peripheral Lymph Node Carriage in Calves

Abstract

Non-typhoidal salmonellosis is the leading bacterial cause of human foodborne disease in the United States and is responsible for numerous hospitalizations and deaths each year. Although several strategies have been successful in controlling other foodborne pathogens such as Escherichia coli O157:H7, the incidence of Salmonella foodborne disease has been persistently above the national Healthy People target for the past decade, emphasizing the importance of the need for increased efforts to control this pathogen. Cattle are known to harbor Salmonella and ground beef has been identified as the source of several foodborne Salmonella outbreaks, including a multistate outbreak of Salmonella Enteritidis (SE) in ground beef during the summer of 2012. Although cattle hides are considered to be the largest contributor to beef carcass contamination by enteric pathogens, a growing body of evidence suggests that peripheral lymph nodes (PLN) are important sources of Salmonella contamination of ground beef. Unlike mesenteric lymph nodes, PLNs are not routinely removed during carcass evisceration and are a potential source of ground beef contamination from infected cattle, underscoring the importance of pre-harvest interventions to reduce pathogens in these sites. Researchers have examined the use of vaccines and direct-fed microbials to reduce Salmonella in peripheral lymph nodes with limited success. Bacteriophage (phage) treatment also has been explored to reduce pathogens and previous studies have demonstrated the efficacy of bacteriophage (phage) treatment to control E. coli O157:H7 in pre-harvest cattle. However, to my knowledge, nothing has been published on the use of orally applied phage as a pre-harvest intervention strategy for, or to reduce peripheral lymph node carriage of, Salmonella in cattle. In response to the 2012 outbreak of SE in ground beef, I hypothesize that SE causes enteric disease in calves and disseminates from the bovine gut to the peripheral lymph nodes, which are not removed at slaughter, and thus contaminate ground beef following carcass processing. The specific aims of this project were to (1) develop an experimental model of SE infection and peripheral lymph node carriage in calves; (2) evaluate the potential for a treatment cocktail of seven lytic phages targeting SE to reduce fecal shedding, disease signs, and peripheral lymph node carriage in the SE calf model, and (3) characterize the seven phages and optimize the treatment by selecting the three phages best suited for pathogen reduction. To address specific aims one and two, I worked toward developing a model of SE infection and PLN carriage in calves and evaluated the potential for a seven-phage cocktail to reduce disease signs, fecal shedding, and PLN carriage in infected calves. Three pairs of 5-7 week-old calves (four control calves, and two treated calves) were orally challenged with 5.0 x 109 - 1.3 x 1010 CFUs of a bovine SE isolate. Following inoculation, daily fecal samples were enumerated for SE and rectal temperatures were recorded twice daily. Blood, subiliac, and superficial cervical lymph nodes were cultured post-mortem. In treated calves, a cocktail of seven lytic phages targeting SE was orally administered following SE challenge. Oral challenge with SE produced mixed results. Fever spikes were noted for days two or three post inoculation. Although each calf received a high dose of SE, fecal shedding of the organism varied among calves in control (C1, C2, C3, C4) and phage-treated (T1 and T2) groups. Calf T1 shed low amounts of SE (2‒3 log10CFU/g feces); calves C1, C3 and C4 shed moderate amounts of SE (4-6 log10CFU/g feces); and calves C2 and T2 shed high amounts of SE (6‒8 log10CFU/g feces). Bacteremia was noted for two amongst the three most severely affected calves and SE was recovered from the PLNs of the same three calves. Following treatment, phages were recovered from PLNs of calf T2. These findings demonstrate that SE causes enteric disease and invades PLNs in calves and that phage treatment may be effective in mitigating Salmonella carriage in PLNs. Also, the presence of SE in the PLNs of the three most severely affected calves and its presence in the blood of two of these three suggests that bacteremia may mediate translocation of Salmonella from the gut to the PLNs. Although findings from my calf model of SE suggested that an oral treatment cocktail of seven lytic phages targeting SE may control peripheral lymph node carriage, I sought to reduce the cocktail to the three most suitable phages due to the constraints involved in validating the safety and efficacy of seven phages in a calf model. Characterization experiments were performed for each of the seven phages in the cocktail in order to establish exclusion criteria for cocktail optimization. Electron micrographs were prepared by negatively staining concentrated phage lysates with 2% phosphotungstic acid and viewed with transmission electron microscopy. Qualitative lytic activity was assessed by performing Salmonella growth curves in the presence of phage (lysis curves) at varying multiplicities of infection (MOI). Additionally, Salmonella host range, efficiency of plating, adsorption rate constants, and ultra-violet (UV) inactivation constants were determined for each phage. Phages were classified into three families based upon morphology: Myoviridae (three phage), Siphoviridae (two phage), and Podoviridae (two phage). Each cocktail phage demonstrated strong lytic activity against SE and was able to lyse or form plaques on multiple Salmonella serovars. Except for phage in the Podoviridae family, similarities in the host ranges, efficiencies of plating, lysis curve patterns and adsorption rate constants were found among phages in the same family, suggesting redundancy among cocktail phages in the Myoviridae and Siphoviridae families. These findings were used to select three of the seven phages for future treatment experiments in my calf model.