Correspondence: Steven C. Ricke, Center for Food Safety, Department of Food Science, University of Arkansas, Fayetteville, AR 72704, USA. Tel.: +1 479 575 4678; fax: +1 479 575 6936; e-mail: email@example.com
Salmonella infections are reported as the second most common pathogen caused foodborne disease in the United States, and several Salmonella serovars can colonize in the intestinal tracts of poultry. Reducing Salmonella in poultry is crucial to decrease the incidence of salmonellosis in humans. In this study, we evaluated the immune response of chicken macrophage cells (HD-11) and effects of bacteriophage P22 against the extra- and intracellular S. Typhimurium LT2. Four treatments, (1) HD-11 cells as control, (2) HD-11 cells with LT2, (3) HD-11 cells with LT2 and P22, and (4) HD-11 cells with P22, were administered, and IL-8 responses of HD-11 cells were measured using an ELISA. Also, four cytokine (IL-4, IL-8, IL-10, and IFN-γ) gene expression levels in the presence of LT2 and/or P22 were quantified by qRT-PCR. We found that P22 lysed the extra- and intracellular LT2, which adhered and were taken up by the HD-11 cells. The ELISA indicated that HD-11 cells produced significantly higher IL-8 cytokine levels in the supernatant during the intracellular lyses of LT2 by P22 (P < 0.05). The IL-8 expression levels measured by qRT-PCR also exhibited similar results with the IL-8 production based on ELISA measurements.
Salmonella has the potential to cause fatal bacterial infections in infants and individuals with a suppressed immune system (Scallan et al., 2011; Finstad et al., 2012). The majority of the foodborne Salmonella serovars can colonize in the intestinal tracts of humans, and one of the major routes of human salmonellosis is believed to be consumption of contaminated poultry and poultry products (Finstad et al., 2012). Thus, strategies for the control and prevention of poultry colonization are needed to further reduce the incidence of salmonellosis in humans. Currently, there are several preventative measures for limiting Salmonella establishments in poultry flocks including dietary alterations, prebiotics, probiotics, antimicrobials such as organic acids, and the administration of vaccine strains (Ricke, 2003a, b; Vandeplas et al., 2010). However, there are very limited options for reducing already established Salmonella in the avian gastrointestinal tract (Toro et al., 2005; Atterbury et al., 2007).
In many developed countries, bacteriophage therapy was abandoned in favor of the development and widespread production of antibiotics (Stone, 2002). However, interest in phage therapy has gained momentum in animal production systems over the past few years as antibiotics have fallen out of favor (Nakai & Park, 2002; Joerger, 2003; Levin & Bull, 2004; Atterbury et al., 2007; Ricke et al., 2012). The bacteriophage P22 used in this study is able to bind specific somatic antigen structures of lipopolysaccharide (LPS) present in Salmonella serogroup A, B, and D1 including S. Typhimurium via tailspike proteins (TSP) (Marietto-Gonçalves et al., 2011). P22 utilize TSP enzymes to penetrate the outer membrane of S. Typhimurium and allow P22 to inject genetic materials into host cells. The virions produced by replication of P22 DNA inside the host cells assemble to form mature P22 and subsequently lyse the respective host cells.
Because cytokines are considered as crucial regulators or mediators against antigens in the host immune system, the change of cytokine expression levels in the presence of Salmonella and bacteriophage is important for understanding roles in inflammation and apoptosis during pathogen infections (Liu et al., 2010). Research on avian cytokines has expanded due to the increased interest in avian immune responses against pathogens and advanced techniques that are now available for studying these responses in detail (Giansanti et al., 2006).
In this study, we hypothesized that P22 could enhance reduction in intracellular S. Typhimurium if the phage was allowed to come in contact with the host cell. To test this hypothesis, we used an in vitro host chicken cell model (HD-11 macrophage) to differentiate the cellular immune response against S. Typhimurium with or without P22. As part of this study, HD-11 cytokine expression levels were assessed using an enzyme-linked immunosorbent assay (ELISA) and quantitative reverse-transcriptase polymerase chain reaction assays (qRT-PCR).
Materials and methods
Bacterial strains and growth conditions
Salmonella Typhimurium LT2 (ATCC 19585) and ST55, a reduced motility mutant, were used in this study (Aswad & Koshland, 1975). One loop of each S. Typhimurium strain was taken from frozen stock and streaked onto Luria–Bertani (LB) (EMD Chemicals Inc, Gibbstown, NJ) plate. After incubation for 24 h at 37 °C, one colony was selected and grown in 5 mL of LB broth for 24 h at 37 °C under aerobic growth conditions.
Propagation and enumeration of bacteriophage P22
The S. Typhimurium LT2 strain was grown on LB plates overnight at 37 °C under aerobic incubation conditions, and cells were collected in phosphate-buffered saline (PBS) solution. The OD600 adjusted LT2 strain was subsequently added to 50 mL of LB broth in a conical flask, and triplicate cultures were grown to late log phase including approximately 108 colony-forming units (CFU) at which point P22 including 106 plaque-forming units (PFU) was added to an approximate multiplicity of infection (MOI) of 0.01. Incubation with shaking was continued overnight, and grown cultures were filtered to remove LT2 cells and bound phages. Unbound free phages were enumerated on host lawns of LT2 strain and stored at −20 °C. P22 stocks were evaluated for the contamination of phage-resistant bacteria prior to use.
Chicken macrophage (HD-11) cells were maintained in minimal essential medium (MEM) with 10% fetal bovine serum (FBS) and grown routinely in a 75-cm2 flask at 37 °C in a 5% CO2-humidified incubator. Confluent stock cultures were treated with trypsin to release the attached cells, and new stock cultures were seeded with 105 cells per mL. For the adherence and uptake assays, 24-well tissue culture plates (BD Biosciences, Franklin Lakes, NJ) were seeded with 105 cells per mL of HD-11 cells and incubated at 37 °C in a 5% CO2-humidified incubator for 18–20 h, and a semiconfluent monolayer was obtained. Prior to the experiment, the monolayer was washed and incubated in MEM containing 10% FBS without antibiotic.
Adherence and uptake assays
Adherence and uptake assays were performed using a modified procedure derived from Biswas et al. (2006). One loop of LT2 and ST55 grown overnight was collected from LB plates and suspended in MEM with 10% FBS. The OD of each strain suspension was subsequently adjusted to an absorbance value of 0.2 at 600 nm. A 100 μL of the suspension containing approximately 107 CFU (MOI of 100) was inoculated into duplicate wells of a 24-well tissue culture plate containing semiconfluent monolayers of HD-11 cells. The concentration of each strain was determined simultaneously on LB plates as described previously. Infected monolayers were incubated for 2 h at 37 °C under a 5% CO2-humidified atmosphere to allow LT2 and ST55 adherence and uptake by the cells. One plate was washed five times with PBS, and P22 at 107 PFU (MOI of 1) was added followed by incubation for 4, 8, and 16 h to lyse bacterial cells. After incubation, the HD-11 cells were lysed with 0.1% Triton X-100 (Sigma, St. Louis, MO) in PBS for 15 min to enumerate the number of extra- (adherence) and intracellular (uptake) LT2 and ST55. Other plates were re-incubated for another 2 h in fresh media containing 250 μg mL−1 of gentamicin to kill the extracellular bacteria. After incubation, the number of intracellular (uptake) LT2 and ST55 was evaluated using the same method described previously. Three wells with only HD-11 cells and the other three wells infected with only P22 were prepared as controls. Results were expressed as the average number of adhered and invaded cells by LT2 and ST55 in three to five independent assays.
Detection of IL-8 cytokine produced by HD-11 cells using ELISA
Total four treatments were utilized in this study, (1) HD-11 cells as control, (2) HD-11 cells with LT2, (3) HD-11 cells with LT2 and P22, and (4) HD-11 cells with P22, to evaluate IL-8 production levels along with a negative control added with PBS instead of HD-11 cells. The ELISA was performed following the protocol provided by BD Biosciences (Catalog No. 555244; BD Biosciences). Visible light absorbance readings of four different treatments were taken at a wavelength of 450 nm to quantify the level of IL-8 cytokine.
Sample treatments for cytokine expression
Five treatments were prepared to evaluate the different cytokine expression levels of HD-11 cells in the presence of LT2 and P22. Treatment A consisted of only HD-11 cells as the control, treatment B was HD-11 cells with P22, treatment C was HD-11 cells with LT2, treatment D was HD-11 cells with P22 and LT2, and treatment E was the same as treatment D except for adding gentamicin. As gentamicin is unable to kill LT2 cells that have already entered into HD-11 cells, treatment E was used to compare cytokine production levels against intracellular LT2 (Durant et al., 2000). All five treatments were used for qRT-PCR analysis to verify the expressions of the four cytokines (IL-4, IL-8, IL-10, and IFN-γ).
Total RNA isolation
Total RNAs from five treatment samples were isolated using Trizol reagent (Sigma) to perform qRT-PCR. A 1mL aliquot of Trizol reagent was added to each sample and collected immediately using a scraper. Total RNA was extracted, and subsequently, DNase I (New England Biolabs, Ipswich, MA) treatment was performed for 1 h at 37 °C to remove possible contaminating genomic DNA. Total RNA was purified by Qiagen RNeasy mini kit (Qiagen, Valencia, CA) according to the manufacturer's instruction, and the concentration was measured by a NanoDrop ND-1000 (Thermo Scientific, Wilmington, DE).
The qRT-PCR assay was optimized using an Eppendorf Masterplex thermocycler ep (Eppendorf, Westbury, NY) Gradient Technology. Three primer pairs for IL-4, IL-10, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were adopted from previous work (Abdul-Careem et al., 2007; Lee et al., 2010), and the other two primer pairs for IL-8 and IFN-γ were synthesized in current study study (Table 1). The specificity of designed primer pairs was confirmed by a basic local alignment search tool (BLAST) program. The 20 μL of reaction mixture consisted of 10 μL of EXPRESS SYBR GreenER™qPCRSuperMix with Premixed ROX (Invitrogen, Carlsbad, CA), 0.5 μL of EXPRESS SuperScript Mix for One-Step SYBR GreenER (Invitrogen), 500 nM of each primer, 500 ng of total RNA template, and DEPC-treated water to volume. The qRT-PCR was optimized for the reaction conditions of 50 °C for 5 min for the synthesis of cDNA. This was followed by 40 cycles of 95 °C for 15 s, 57 °C for 15 s, and 68 °C for 20 s. Melting curves were subsequently created, which consisted of 95 °C for 15 s and 60 °C for 20 min increasing by 0.5 °C per minute to a final temperature of 95 °C. Each experiment was performed in triplicate.
The GADPH gene was used as an internal standard to normalize the qRT-PCR, and the Ct values were calculated with the Eppendorf realplex software (version 2.0). The relative gene expression changes in transcription levels of the four cytokines between the control and treatments were determined using the 2−ΔΔCt method (Livak & Schmittgen, 2001). The data were generated by three independent experiments, and each trial was carried out in triplicate. Statistical analysis was performed using JMP® Genomics 5.0 software (SAS Institute Inc., Cary, NC). The experimental data were analyzed using a t-test, and a calculated P value of < 0.05 was used to delineate significant differences.
Adherence and uptake of S. Typhimurium
In this series of experiments, we measured the adherence and uptake of LT2 and ST55 and demonstrated that our cell culture model could be used to screen for extra- and intracellular survival of both strains. Based on Fig. 1, it appears that both LT2 and ST55 adhered to HD-11 cells at 3.6 ± 0.2% and 2.5 ± 0.3% of their initial inoculation levels, respectively. Similarly, LT2 (0.39 ± 0.04%) were taken up by HD-11 cells more than ST55 (0.25 ± 0.02%). As LT2 exhibited a greater difference in adherence and uptake than the ST55 strain (P < 0.05), we continued the study with LT2.
Evaluation of extra- and intracellular killing of LT2 by P22
In a follow-up study utilizing the HD-11 cell culture model, we observed that P22 was capable of killing both extra- and intracellular LT2. Figure 2 showed relative recovery of LT2 strains at each time point. By the first 8 h after inoculation, no significant differences in intracellular recoveries of LT2 occurred between 4 and 8 h, while almost all intracellular cells of LT2 were eliminated at 16 h.
We also observed that HD-11 cells produced significantly higher amounts of cytokine (IL-8) in the supernatant in the presence of LT2 with P22 (P < 0.05). This finding indicates that intracellular lysis of LT2 strains enhanced the cell-mediated immune response of chicken macrophages. In this study, we detected the IL-8 produced by LT2-infected HD-11 cells during killing of extra- and intracellular bacterial cells by P22. The expression patterns of IL-8 in the LT2-infected HD-11 cells are illustrated in Fig. 3. The IL-8 expression level in the HD-11 cell increased over two fold in the presence of LT2 compared with HD-11 cells alone, and P22 also stimulated the expression of IL-8. In addition, IL-8 expression level increased significantly in the presence of both LT2 and P22.
Cytokine expression levels by qRT-PCR
The primer pairs amplified PCR products with high specificity for each target gene with the respective melting curve. Electrophoresis of PCR products was conducted on an agarose gel to confirm the exact PCR result as well as each amplicon size as shown in Table 1. Infection of LT2 and P22 increased cytokine expression levels in HD-11 cells. The fold changes in cytokine expression levels due to the four different treatments compared with the naïve cells are presented in Table 2. The GAPDH gene served as the reference gene to normalize cytokine expression levels as fold changes. The IL-4 gene exhibited increases when LT2 and P22 were administered to the chicken cells, but adding gentamicin caused a decrease in IL-4 gene expression. For the IL-8 gene, all treatments significantly increased gene expression levels (P < 0.05), and treatment E in particular exhibited a more than two fold increase compared with other treatments. The expression of IL-10 gene was markedly increased for treatment B (P < 0.05) with no significant differences in any of the other treatments. Finally, the IFN-γ gene appeared to be highly up-regulated for treatment B.
Table 2. Cytokine gene expression in response to five treatments
Fold changes: values with different superscript capital letters (A to E) in columns and rows are significantly different (P < 0.05).
HD-11 + bacteriophage (P22)
HD-11 + S. Typhimurium LT2
HD-11 + S. Typhimurium LT2 + bacteriophage (P22)
HD-11 + S. Typhimurium LT2 + bacteriophage (P22) + gentamicin
The significance of this study was the detection of changes in immune responses to LT2 on chicken macrophage cells (HD-11) when combined with P22 as macrophages play important roles in the innate immune system. Cytokines produced by innate immune cells influence the adaptive immune response and cell signaling molecules in intracellular communication (Witheanage et al., 2004). The cytokines evaluated in this study were selected for the important roles they play in innate and adaptive immunity. Furthermore, they interact with a wide variety of cell products during the immune response (Schroder et al., 2004). In the present study, the effects of P22 on the host (LT2) and the production of four different cytokines in chicken macrophage cells were evaluated by adherence and uptake assays as well as ELISA and qRT-PCR. The specific ability of P22 for killing LT2 has been reported in several studies (Pope et al., 2004; Toro et al., 2005). The phage utilized in this study was able to initiate killing of extra- and intracellular S. Typhimurium within a few minutes after infection and completed bacterial lysis within 16 h.
Cytokines are generally divided into several categories by the activity and/or effects they produce (Giansanti et al., 2006). Pro-inflammatory cytokine IL-8 is produced by stimulation of macrophages, and IFN-γ can also be induced by natural killer (NK) cells as well as T cells, and both cytokines are associated with innate immune response (Giansanti et al., 2006; Apte et al., 2008). In contrast, IL-10 produced by mast cells inhibits both NK cell activity and pro-inflammatory cytokine synthesis (Pestka et al., 2004). The cytokine IL-4, which is a key regulator in humoral and adaptive immune response, decreases the production of macrophages and IFN-γ (Apte et al., 2008). There were no significant differences in IL-4 expression levels among any of the treatments. IL-4 is a key cytokine for humoral immunity and stimulates B- and T-cell proliferation and is characterized as a signal for decreasing production and deactivation of macrophages (Bogdan & Nathan, 1993; He et al., 2011). The expression levels of IL-4 on each treatment exhibited no significant changes statistically when compared with control. IL-10 is an anti-inflammatory cytokine and inhibits the ability of antigen-presenting cells (APCs). Therefore, the presence of LT2 in HD-11 cells (treatments C, D, and E) led to no significant differences in expression levels of IL-10. The IFN-γ is an important cytokine in host defense mechanism against viral and intracellular pathogens. It is stimulated by macrophages and induces antimicrobial as well as antiviral activities (Liu et al., 2010).
In addition, the expression levels of IL-8 have been shown to be greatly increased by S. Typhimurium infection at multiple organs in chicken such as liver, cecal tonsil, and jejuna (Witheanage et al., 2004). As IL-8 is an important chemokine in immune system against bacterial and viral infections, the expression level of IL-8 was investigated by both indirect ELISA and qRT-PCR. Although individual infection of LT2 and P22 stimulated IL-8 expression levels, the presence of both increased IL-8 production significantly more than either individual treatment. These two results were supported by both the ELISA and qRT-PCR. In addition, the drastic over 40-fold increase in IL-8 production compared with control in all treatments implied that the infection of LT2 and P22 could stimulate IL-8 production in HD-11 cells. IL-8 mRNA expression levels in treatment B were higher than that in treatment C (Table 2); however, both treatments showed reverse results in protein expression levels (Fig. 3). The difference in correlation between mRNA and protein expression levels may be due to several factors such as various and complicated post-transcriptional mechanisms in translation from mRNA to protein, different half lives of protein as well as both mRNA and protein experimental limitations (Greenbaum et al., 2003). IL-8 is one of the CXC chemokines produced by macrophages and is an important mediator for initiation of innate immune response in the infected cells.
The utilization of bacteriophages to modulate pathogen load in complex ecosystems such as the intestine represents additional logistical challenges (Barrow, 2001; Ricke et al., 2012). Nonetheless, in the late 1980s, Smith et al. (1987) successfully used bacteriophages to control E. coli diarrhea in calves. Their study demonstrated the potential effectiveness of bacteriophage use to treat intestinal bacterial infections even in the complex milieu of the gastrointestinal system. In previous reports, several studies have shown that bacteriophages may be useful in reducing the number of bacterial foodborne pathogens including Escherichia coli O157 : H7 (Sheng et al., 2006), Campylobacter jejuni (Goode et al., 2003), Listeria monocytogenes (Leverentz et al., 2003), and Salmonella serovars (Higgins et al., 2005; Wall et al., 2010) contaminating the surface of food, poultry, and swine. In addition, bacteriophages have been investigated for their ability to reduce Salmonella already established in the poultry intestine; however, this application has resulted in only modest success (Toro et al., 2005; Atterbury et al., 2007; Higgins et al., 2007). Huff et al. (2010) examined the immune response of chicken against bacteriophage SPR02 by IgGlevel titers in serum. Prior exposure to the same bacteriophage increased IgG levels in the chicken such that the therapeutic effectiveness of bacteriophage was believed to be decreased by the avian immune response.
Bacterial infections in animal hosts theoretically can be controlled by bacteriophage treatment through two mechanisms: direct bacteriophage lysis or immune response via bacterial lysate produced by bacteriophages (Merril et al., 1996; Borysowski & Górski, 2008). Thus, bacterial infection in a host can be directly eliminated by adding bacteriophages. However, bacteriophage itself can increase specific IgG serum levels in the animal host by intramuscular injection because outer protein structures of bacteriophage are recognized as antigens in the host cells; thus, they can be neutralized by antibodies (Huff et al., 2010; Ricke et al., 2012). As a result, bacteriophage specific antibodies decreased the antibacterial phage activities and increased the mortalities of animal host (Huff et al., 2010).
In this study, we evaluated several chicken macrophage cell (HD-11) cytokine responses to the presence of either bacteriophage or S. Typhimurium and were able to detect differential immune responses by the host cells. However, as this was an in vitro model system, this does not ensure that such direct cell to phage interactions would occur in vivo. To assess such interactions when using bacteriophage for systemic treatments in food animals such as chickens, it will be essential in future studies to investigate the overall animal host immune response against the bacteriophage activities as well as responses at the cellular levels of the host.
We thank Dr. Gary Acuff, Department of Animal Sciences, Texas A&M University, College Station, Texas, USA, for providing bacterial strains used in this study. This study was supported by National Integrated Food Safety Initiative (NIFSI) grant # 2008-51110-04339 and a USDA Food Safety Consortium Grant at the University of Arkansas, Division of Agriculture.