Evaluation of the Safety and Protection Efficacy of spiC and nmpC or rfaL Deletion Mutants of Salmonella Enteritidis as Live Vaccine Candidates for Poultry Non-Typhoidal Salmonellosis

Salmonella enterica serovar Enteritidis (S. Enteritidis) is a host-ranged pathogen that can infect both animals and humans. Poultry and poultry products are the main carriers of S. Enteritidis, which can be transmitted to humans through the food chain. To eradicate the prevalence of S. Enteritidis in poultry farms, it is necessary to develop novel vaccines against the pathogen. In this study, we constructed two vaccine candidates, CZ14-1∆spiC∆nmpC and CZ14-1∆spiC∆rfaL, and evaluated their protective efficacy. Both mutant strains were much less virulent than the parental strain, as determined by the 50% lethal dose (LD50) for three-day-old specific-pathogen free (SPF) White Leghorns and Hyline White chickens. Immunization with the mutant candidates induced highly specific humoral immune responses and expression of cytokines IFN-γ, IL-1β, and IL-6. In addition, the mutant strains were found to be persistent for almost three weeks post-infection. The survival percentages of chickens immunized with CZ14-1∆spiC∆nmpC and CZ14-1∆spiC∆rfaL reached 80% and 75%, respectively, after challenge with the parental strain. Overall, these results demonstrate that the two mutant strains can be developed as live attenuated vaccines.


Introduction
The Salmonella enterica serovar Enteritidis (S. Enteritidis) is one of the most common Salmonella serotypes causing human infections worldwide [1,2]. Human salmonellosis is frequently caused by consumption of Salmonella-contaminated foods. From 2015 to 2017, the proportion of human salmonellosis caused by S. Enteritidis increased from 54.7% to 61.2% in the EU (EFSA, 2018). In the USA, S. Enteritidis was the predominant serotype in reported human Salmonella infections from 2011 to 2016, and it was also the most common serotype in clinical chicken isolates [1]. In China, our surveillance results over the last 10 years showed that about 30% of Salmonella isolates from poultry meat were

Chickens
The specific-pathogen free (SPF) White Leghorns and Hyline White chickens were purchased from Jinan Sipaffrey Poultry Co., Ltd. (Jinan, Shangdong, China). The 3-day old chickens were confirmed as free from Salmonella infection by both bacteriological examination and serum detection. Each group of chickens was housed in separate rearing isolators and supplied with commercial feed and drinking water. The food and the water for chickens were tested to be Salmonella negative. All of the animal experiments and management procedures were undertaken with the permission of the Animal Welfare and Ethics Committees of Yangzhou University (Approval No. SYXK(Su)2016-0020) and complied with the guidelines of the institutional administrative committee and the ethics committee for laboratory animals.

Virulence Assessment
To evaluate the virulence of the two mutants (CZ14-1∆spiC∆nmpC and CZ14-1∆spiC∆rfaL) in chickens, the 50% lethal doses (LD 50 ) were determined using 50 chickens for each strain randomly assigned into five groups. Each group was inoculated intramuscularly with 10-fold dilutions of the mutants from 1 × 10 5 to 1 × 10 9 CFU. Another group of 50 was inoculated intramuscularly with the parental strain CZ14-1 from 1 × 10 4 to 1 × 10 8 CFU. In addition, ten chickens were inoculated with 200 µL of phosphate-buffered saline (PBS) via the same route. Birds that died were recorded daily for three weeks post-infection, and the LD 50 was calculated using the Reed-Muench method [25].

Biochemical Test and Growth Ability In Vitro
To investigate the effect of deleted genes on the biological properties of S. Enteritidis mutants, a biochemical test was performed using an API20E plate (BioMérieux, France), and the results compared to those for the parental strain CZ14-1. In addition, the growth ability was determined by measuring the optical density (OD 600 ) of each strain cultured in 15 mL of LB broth at 37 • C with shaking at 180 rpm. The OD 600 was monitored every hour for 18 h as previously described [26].

Bacterial Vaccination and Changes in Body Weight in Chickens
A total of 60 7-day old Hyline White chickens were randomly divided into three groups (n = 20 for each group): CZ14-1∆spiC∆nmpC-vaccinated group, CZ14-1∆spiC∆rfaL-vaccinated group, and non-vaccinated control group. The vaccinated groups were immunized intramuscularly with 1 × 10 5 CFU of one of the mutant strains in 100 µL PBS, while the control group was only inoculated with PBS. At 14 days old, the chickens were again immunized via the same route and the same dose of bacteria. The body weights of these chickens were then measured at 7, 14, and 21 days after the second vaccination.

Persistence and Clearance of Bacteria in Chicken Tissues and Organs
A total of 90 3-day old Hyline White chickens were randomly assigned into three groups (n = 30 for each group). Each group was immunized intramuscularly with 1 × 10 5 CFU of CZ14-1, CZ14-1∆spiC∆nmpC, or CZ14-1∆spiC∆rfaL. Liver, spleen, ileum, and cecum samples from five chickens from each group were aseptically collected at 1, 3, 7, 14, and 21 days post immunization. The samples were weighed and suspended in 1 mL of PBS for homogenization. The homogenates were diluted serially and subsequently inoculated on Brilliant blue agar plates for bacterial recovery at 37 • C. After overnight cultivation, the bacterial number was counted and reported as log 10 CFU/g.

Immune Response
The humoral immune response was evaluated through measurement of IgG titers specifically against S. Enteritidis by indirect enzyme-linked immunosorbent assay (ELISA) using the heat-killed CZ14-1 strain as the coating antigen. Serum samples were collected at 7, 14, and 21 days post immunization and then serially diluted to be considered as the primary antibody. Horseradish peroxidase (HRP)-conjugated goat anti-chicken IgG (1:10,000 dilution, Sigma-Aldrich) was used as the secondary antibody. The HRP activity was determined using 3,3 ,5,5 -tetramethylbenzidine (TMB, Sigma-Aldrich), and the OD 450 value was determined with an ELISA reader (BioTek, USA).
To evaluate the T-cell immune response and inflammatory response induced by the vaccine strains, the expression of cytokines IFN-γ, IL-1β, IL-6, and CXCLi1 was measured by qRT-PCR analysis with Taqman TM probes as previously described [27]. Total RNA was extracted from spleen samples collected at 7, 14, and 21 days post-vaccination using an RNA plus Mini kit (Qiagen, Germany) according to the manufacturer's instructions. The RNA was reverse transcribed into cDNA and subjected to qRT-PCR analysis, which was performed in triplicate and repeated three times.

Immune Protection Assessment
To evaluate the protective efficacy of the two vaccine candidates, 60 7-day old Hyline White chickens were randomly divided into three groups of 20 chickens (groups A, B, and C). Groups A and B were immunized intramuscularly with 1 × 10 5 CFU of the mutants CZ14-1∆spiC∆nmpC and CZ14-1∆spiC∆rfaL, respectively. Group C received 100 µL of PBS as the negative control. One week post-infection, groups A and B were immunized again with the same mutants using the same route and dose. After one week, all of the groups were challenged with 2 × 10 9 CFU of CZ14-1. The number of surviving chickens was recorded, and clinical symptoms were observed daily for three weeks. In addition, to examine the persistence and the clearance of the wild-type strain in the vaccinated chickens, 90 7-day old chickens were divided into three groups of 30 chickens and subjected to the same immunization and challenge process as described above. After challenging, five chickens from each group were killed at 3, 7, 14, 21, 28, and 35 days post-challenge, liver, spleen, ileum, and cecum samples were collected, and the bacterial number was counted on Brilliant green agar plates, as mentioned above.

Statistical Analysis
All data are described as mean ± standard error of the mean (SEM) unless otherwise specified and were analyzed using GraphPad Prism version 7.0. A p value of less than 0.05 was considered statistically significant.

Biochemical Property and Growth Ability of Mutant Strains
The biochemical test performed using the API20E plate reflected that the mutated genes had no significant effect on bacterial metabolism (unpublished). Mutation of rfaL in CZ14-1∆spiC∆rfaL caused a deficiency in S. Enteritidis in utilizing mannitol. Moreover, the bacterial growth data showed that mutation of rfaL decreased growth under laboratory conditions. However, the growth curve of CZ14-1∆spiC∆nmpC was similar to that of the parental strain, which showed that deletion of nmpC did not perturb the bacterial growth in vitro Figure 1a.
curve of CZ14-1∆spiC∆nmpC was similar to that of the parental strain, which showed that deletion of nmpC did not perturb the bacterial growth in vitro Figure 1a.

Changes in Body Weight Following Immunization
The mean body weights of the chickens in the two vaccinated groups and the PBS group at 7, 14, and 21 days post-vaccination are shown in Figure 1b. Chickens in both vaccinated groups showed increased body weight compared to that of the PBS group. Furthermore, the bodyweight of chickens in the CZ14-1∆spiC∆rfaL group was significantly higher than that of the PBS group. No significant difference was detected between the CZ14-1∆spiC∆nmpC-vaccinated group and the PBS group at 7 and 21 days post-vaccination.

Bacterial Colonization and Persistence in Chicken Tissues and Organs
All of the tissues and organs collected from chickens inoculated with PBS were negative for bacterial recovery. However, although the virulence of the two mutant strains was significantly decreased compared to the parental strain, the colonization ability of both mutant strains in all four tissue and organ types tested was significantly stronger than that of the wild-type strain at 3 days post-immunization ( Figure 2). Bacterial recovery was not detected in tissue and organ types tested from the wild-type-infected group at 1 and 3 days post-immunization. While ≥4/5 (four out of five chickens) of CZ14-1∆spiC∆nmpC and all five of the CZ14-1∆spiC∆rfaL-immunized chickens were found to be positive for Salmonella in the liver at 1 and 3 days post-immunization (Figure 2a), the CZ14-1∆spiC∆rfaL exhibited faster invasion into the spleen than the parental strain and CZ14-1∆spiC∆nmpC (Figure 2b). At seven and 14 days post-immunization, the bacterial counts in the spleen, the liver, and the ileum were similar among the three groups (Figure 2a-c), but there was a significant difference between cecum from the wild-type-infected group and that from chickens vaccinated with CZ14-1∆spiC∆rfaL (Figure 2d). At 21 days post-immunization, bacteria were almost eliminated in the liver and the ileum in all three groups. Only one chicken vaccinated with CZ14-1∆spiC∆nmpC was found to be Salmonella-positive in the spleen (Figure 2c). However, in the cecum, a low level of bacterial recovery could still be detected in all three groups (Figure 2d).

Immune Response
To evaluate the humoral immune response induced by the two vaccine candidates, serum IgG levels were determined using an indirect ELISA assay. In both of the vaccinated groups, the serum

Changes in Body Weight Following Immunization
The mean body weights of the chickens in the two vaccinated groups and the PBS group at 7, 14, and 21 days post-vaccination are shown in Figure 1b. Chickens in both vaccinated groups showed increased body weight compared to that of the PBS group. Furthermore, the bodyweight of chickens in the CZ14-1∆spiC∆rfaL group was significantly higher than that of the PBS group. No significant difference was detected between the CZ14-1∆spiC∆nmpC-vaccinated group and the PBS group at 7 and 21 days post-vaccination.

Bacterial Colonization and Persistence in Chicken Tissues and Organs
All of the tissues and organs collected from chickens inoculated with PBS were negative for bacterial recovery. However, although the virulence of the two mutant strains was significantly decreased compared to the parental strain, the colonization ability of both mutant strains in all four tissue and organ types tested was significantly stronger than that of the wild-type strain at 3 days post-immunization ( Figure 2). Bacterial recovery was not detected in tissue and organ types tested from the wild-type-infected group at 1 and 3 days post-immunization. While ≥4/5 (four out of five chickens) of CZ14-1∆spiC∆nmpC and all five of the CZ14-1∆spiC∆rfaL-immunized chickens were found to be positive for Salmonella in the liver at 1 and 3 days post-immunization (Figure 2a), the CZ14-1∆spiC∆rfaL exhibited faster invasion into the spleen than the parental strain and CZ14-1∆spiC∆nmpC (Figure 2b). At seven and 14 days post-immunization, the bacterial counts in the spleen, the liver, and the ileum were similar among the three groups (Figure 2a-c), but there was a significant difference between cecum from the wild-type-infected group and that from chickens vaccinated with CZ14-1∆spiC∆rfaL (Figure 2d). At 21 days post-immunization, bacteria were almost eliminated in the liver and the ileum in all three groups. Only one chicken vaccinated with CZ14-1∆spiC∆nmpC was found to be Salmonella-positive in the spleen (Figure 2c). However, in the cecum, a low level of bacterial recovery could still be detected in all three groups (Figure 2d).

Immune Response
To evaluate the humoral immune response induced by the two vaccine candidates, serum IgG levels were determined using an indirect ELISA assay. In both of the vaccinated groups, the serum IgG against S. Enteritidis was detected at 7 days post immunization and increased dramatically at 14 and 21 days post immunization then remained stable from 21 to 35 days post immunization  Figure 3a, 3b). At 28 days post immunization, the serum IgG level was significantly higher in the CZ14-1∆spiC∆rfaL-immunized group than that in the CZ14-1∆spiC∆nmpC-immunized chickens (Figure 3b). IgG against S. Enteritidis was detected at 7 days post immunization and increased dramatically at 14 and 21 days post immunization then remained stable from 21 to 35 days post immunization ( Figure  3a, 3b). At 28 days post immunization, the serum IgG level was significantly higher in the CZ14-1∆spiC∆rfaL-immunized group than that in the CZ14-1∆spiC∆nmpC-immunized chickens ( Figure  3b). qRT-PCR analysis of cytokine expression levels in the spleen showed that both vaccinated groups displayed higher IFN-γ levels than the control group at 7, 14, and 21 days post-immunization, and no difference was detected between the two vaccinated groups (Figure 3c). The expression of IL-1β was higher in the two vaccinated groups than the PBS group at 7 and 14 days post immunization (Figure 3e), while IL-6 showed higher expression at 14 and 21 days post-immunization in the qRT-PCR analysis of cytokine expression levels in the spleen showed that both vaccinated groups displayed higher IFN-γ levels than the control group at 7, 14, and 21 days post-immunization, and no difference was detected between the two vaccinated groups (Figure 3c). The expression of IL-1β was higher in the two vaccinated groups than the PBS group at 7 and 14 days post immunization (Figure 3e), while IL-6 showed higher expression at 14 and 21 days post-immunization in the vaccinated groups than that in the PBS group (Figure 3d). Immunization with the two mutant strains had no effect on the expression of CXCLi1, which was similar in the vaccinated groups and the PBS group (Figure 3f).
Vaccines 2019, 7, x 8 of 13 vaccinated groups than that in the PBS group (Figure 3d). Immunization with the two mutant strains had no effect on the expression of CXCLi1, which was similar in the vaccinated groups and the PBS group (Figure 3f).

Immune Protection
The survival percentages for the chickens vaccinated intramuscularly with CZ14-1∆spiC∆rfaL and CZ14-1∆spiC∆nmpC followed by challenge with the wild-type strain CZ14-1 are shown in Table  3. Four and five chickens died in groups A and B, respectively, following the challenge with the parental strain. However, 19 out of 20 non-vaccinated chickens died in group C. The clinical symptoms in the vaccinated group were slight and temporary after the challenge compared with the severe diarrhea seen in the non-vaccinated chickens. Table 3. Protective efficiency of two vaccine candidates in Hyline White chickens.

Group
Vaccination No. Challenge

Immune Protection
The survival percentages for the chickens vaccinated intramuscularly with CZ14-1∆spiC∆rfaL and CZ14-1∆spiC∆nmpC followed by challenge with the wild-type strain CZ14-1 are shown in Table 3. Four and five chickens died in groups A and B, respectively, following the challenge with the parental strain. However, 19 out of 20 non-vaccinated chickens died in group C. The clinical symptoms in the vaccinated group were slight and temporary after the challenge compared with the severe diarrhea seen in the non-vaccinated chickens. To detect the bacteria eradication ability of the vaccinated groups, liver, spleen, ileum, and cecum were collected for bacterial recovery and counting. As shown in Figure 4, the bacterial number was high at 3 and 7 days post-challenge but decreased dramatically at 14 days post-challenge. The bacteria were eradicated at 21, 28, and 35 days post-challenge in the ileum, the liver, and the spleen, respectively (Figure 4b-d). In the cecum, only one chicken in the CZ14-1∆spiC∆nmpC-vaccinated group had almost 10 CFU/g of Salmonella at 35 days post-challenge (Figure 4e). Since most of the chickens in the non-vaccinated group were dead at 7 days post-challenge, no data were collected for the following days in this group.  To eradicate intracellular Salmonella, it is critical to stimulate cellular immune responses in the immunized animals [39,40]. At the acute infection stage, the Salmonella-infected host can produce both Th1 and Th2 immune responses to diminish the growth of the invading bacteria, but in the second stage of persistent infection, the reduced Th1 response disrupts the balance between Th1 and Th2, and a new equilibrium between the host and persistent Salmonella infection is developed and continues [41]. Both mutants in this study can induce increased IFN-γ and IL-6 secretion from 7 to 21

Discussion
Until now, three major types of vaccines, namely, killed bacteria, subunit vaccines, and live attenuated vaccines, have been developed to control S. Enteritidis in the poultry industry [28][29][30][31][32]. The killed bacteria and subunit vaccines can induce strong antibodies to eradicate extracellular bacteria, but it is difficult for these antibodies to kill the intracellular S. Enteritidis, which could be achieved with the live attenuated strains of Salmonella [28,32,33]. A live attenuated Salmonella vaccine should be attenuated, immunogenic, protective, and allow differentiation of vaccinated chickens from naturally infected flocks [34,35]. In this study, we constructed two mutants with virulence factor (SPI-2 effector) spiC mutated combined with a gene involved in the synthesis of LPS (rfaL) or out-membrane protein (nmpC). Both mutants are highly attenuated, strongly immunogenic, and provide excellent protective efficacy against parental strain attack. In addition, ELISA or plate agglutination test (PAT) can be used to differentiate vaccinated from naturally infected chickens.
A live attenuated vaccine should be avirulent. SpiC is a SPI2-encoded effector involved in the survival of Salmonella in macrophages [36,37]. Deletion of spiC in S. Typhimurium and S. Pullorum caused significantly decreased virulence in mice and chickens, respectively [13,37]. Although deletion of spiC has been achieved in S. Enteritidis C50041, which was confirmed to be less virulent than the parental strain, a mouse-derived isolate [38], in this study, we constructed an spiC-deleted strain combined with rfaL or nmpC gene using chicken-borne isolate CZ14-1. Both double-gene mutants showed greatly decreased virulence compared to that of the parental strain. Furthermore, immunization with the mutants did not affect the chickens' body weights, leading to better growth than in the non-vaccinated group. No indications of negative effects on the health of the immunized birds were observed.
To eradicate intracellular Salmonella, it is critical to stimulate cellular immune responses in the immunized animals [39,40]. At the acute infection stage, the Salmonella-infected host can produce both Th1 and Th2 immune responses to diminish the growth of the invading bacteria, but in the second stage of persistent infection, the reduced Th1 response disrupts the balance between Th1 and Th2, and a new equilibrium between the host and persistent Salmonella infection is developed and continues [41]. Both mutants in this study can induce increased IFN-γ and IL-6 secretion from 7 to 21 days post-immunization, indicating that a prolonged Th1 immune response can be induced by these mutants. Furthermore, the antibody against S. Enteritidis was produced at 7 days post-immunization, after which production increased and then remained at a stable concentration to kill the extracellular bacteria. However, bacterial shedding is another key point for live attenuated vaccines [42]. In this study, both mutants were almost eliminated in the liver, the spleen, and the ileum at 21 days post-immunization, and the bacteria were not detected at 28 days post-immunization. These results demonstrate that both mutants can be eliminated from the chickens but induce strong cellular immune and humoral responses.

Conclusions
Our data demonstrated that the two mutants could significantly protect the chickens from virulent parental strain challenge. The lower virulence and the higher protective efficacy showed that CZ14-1∆spiC∆nmpC was better than CZ14-1∆spiC∆rfaL. However, for the eradication of the colonized parental strain in tissues, the group immunized with CZ14-1∆spiC∆rfaL was better than that immunized with CZ14-1∆spiC∆nmpC. Since the spiC deletion mutant of Salmonella has proven to be an appropriate vaccine candidate against S. Pullorum/S. Gallinarum infection [13,14], the present two double-mutants show potential for use as vaccines against S. Enteritidis infection in poultry.