Salmonella is one of the most common foodborne pathogens worldwide (Tabu et al., 2012). Gastroenteritis caused in both human and animals by non-typhoid Salmonella (NTS) has become a global public health concern. To date, more than 2,500 Salmonella serotypes have been identified (Yang et al., 2010), and S. Enteritidis is the predominant serotype in many countries (Galanis et al., 2006; Matheson et al., 2010; Zhang et al., 2014a). According to a report by Chinese Centers for Disease Control (CDC), around 792 people died as a result of S. Enteritidis infection annually (Ren et al., 2016).

In general, compared to other NTS serotypes, such as S. Typhimurium or S. Derby, S. Enteritidis is usually self-limiting and shows susceptibility to most clinical first-line medicines (Ke et al., 2014). But antibiotics are needed for complicated cases and are used if the infection spreads or is highly likely to spread from the intestines to the bloodstream and other organs. Third generation cephalosporins have excellent curative effects, but with the abuse and inappropriate use of these antibiotics, an increasing number of S. Enteritidis strains are nowadays resistant, mainly caused by production of extended-spectrum beta-lactamases (ESBLs) (Tenover, 2006). This phenomenon represents an ever-increasing threat to public health and a challenge that must be met by the development of alternative therapeutic strategies. Cefepime, which is the first fourth generation cephalosporin to be approved for use in China, has been reported to show low toxicity and high activity against Enterobacteriaceae that resistant to third-generation cephalosporins (Chapman and Perry, 2003; Yahav et al., 2007). There are some reports stated that cefepime should be used with caution for serious ESBL infections based on its failure on some clinical experiments (Labombardi et al., 2006; Wang et al., 2016; Tamma and Rodriguez-Bano, 2017). But it can be considered for treatment if the MIC ≤ 1 μg/mL and meanwhile taken the patient’s health conditions into consideration, such as the severity of infection, resistance to antibiotics (Patel et al., 2019).

Despite the important role of cefepime in treating ESBL-producing S. Enteritidis infections, the epidemiological data in this field are scarce. Therefore, in this study, we conducted a systematic characterization of the epidemiology of S. Enteritidis infections in Shanghai, China from 2005 to 2014 to clarify the mechanisms underlying the development of cefepime-resistant S. Enteritidis.

In the investigation period from 2005 to 2014 in Shanghai, China, S. Enteritidis strains showed only a slight variation in resistance to commonly used drugs, including third generation cephalosporins (Supplementary Table S4). This finding is in contrast to previous reports that other Salmonella serotypes (e.g., Typhimurium), exhibited an increasing tendency of resistance to at least 10 types of antibiotics during the same period (Ni et al., 2013). However, it is noteworthy that CRSE strains emerged in 2010 and their frequency is rising annually.

In recent years, cefepime ranked the leading anti-infective medicine in many hospitals in China. In order to prevent the spread of antibiotic resistance, cefepime is rarely used for animals or in breeding industry. However, our study shows that cefepime-resistant S. Enteritidis has also been isolated from animal-derived foodborne sources since 2013.

In this study, we analyzed 38 CRSE isolates, all of which exhibited MDR and the frequency of isolates resistant to more than five antibiotics was very high (100%). The emergence of severe drug-resistant status in China is largely related to the abuse of antibiotics as feed additives in veterinary clinical practice and the transfer of movable components carrying drug-resistant genes between isolates, which has become a common problem in both human and veterinary clinical practice (Rahmani et al., 2013). However, a relatively low resistance to amoxicillin-clavulanic acid (15.79%), gentamicin (13.16%), and amikacin (13.16%) was detected in the 38 cefepime-resistant S. Enteritidis isolates. They all showed intermediate susceptibility to ciprofloxacin, ofloxacin, polymyxin B and imipenem; therefore, these drugs should be used with caution. The fact also emphasizes the importance of standardization of clinical drug application and to prolong the effectiveness of new drugs. In addition, although the isolates were sensitive to aminoglycosides and fluoroquinolones, these agents should be used with caution due to the ototoxicity and adverse effects of cartilage on children.

In this study, nine kinds of β-lactamase resistance genes were detected, among them, subtype bla_CTX–M–55 was predominant. Based on amino acid sequence differences, CTX-M-type ESBLs can be divided into more than five groups and hundreds of variants. CTX-M-55 belongs to the CTX-M-1 group, which was first detected in E. coli in Thailand in 2005 (Kiratisin et al., 2007) and first detected in Salmonella in 2011 (Yu et al., 2011). CTX-M-55 has only one amino acid site difference (valine to alanine) compared with CTX-M-15 (Kiratisin et al., 2007), and exhibits high hydrolytic activity to cefotaxime and ceftazidime (Dallenne et al., 2010). The prevalence of CTX-M-55-producing bacteria in China has grown significantly in recent years in both animal and human populations (Bevan et al., 2017). Data showed that CTX-M-55 is widely distributed in foodborne E. coli in Asia regions (Zhang et al., 2014b), which was the second most common CTX-M genotype (26.1%, 29/111) following CTX-M-14 (Zheng et al., 2012). Regarding hospital-associated Enterobacteriaceae infections, CTX-M-55 was found to be more common than CTX-M-15 in China (Xia et al., 2014). Also, CTX-M-55 was the most prevalent ESBL type observed in Salmonella isolates from livestock animals in China (Zhang et al., 2019). It is known that CTX-M-55 exists in bacteria that show resistance to second or third generation cephalosporins, and in this study, we found that CTX-M-55-producing S. Enteritidis was also with resistance to fourth-generation cephalosporin, cefepime. Therefore, CTX-M-55 should be considered as an important surveillance target around the world.

The bla_CTX–M genes of 35 of the 38 isolates were transferred to the recipient by conjugation. Under the influence of bla_CTX–M, the transconjugants showed resistance to cephalosporins but the MICs varied. They did not reach the same MIC values as the donor bacterium, a difference that may be related to the complex mechanisms of antibiotic resistance such as efflux pumps and penicillin-binding protein in the donor bacterium.

In this study, transconjugants carrying bla_CTX–M shared different types of replicons, mainly IncI1 and InFrepB, with sizes ranging from 76.8 to 138.9 kb. The genetic context of bla_CTX–M was mainly “ISEcp1-F-bla_CTX–M–55-ORF477.” ISEcp1, like IS26 and ISCR1, is a member of the IS1380 family. This component shares a similar structure with a promoter, which positively regulates the high-level expression of bla_CTX–M and facilitates its horizontal transmission among various species (Dhanji et al., 2011; Singh et al., 2018).

PFGE analysis showed that most CRSE displayed similar PFGE patterns (>82.90%). Notably, the pattern of one foodborne isolate (SH14SF008) was identical to that of another four isolates collected from clinical diarrhea human samples and all had the same bla_CTX–M–55 gene, which suggested that we should pay attention to food consumption. Furthermore, S. Enteritidis isolates from different years showed 100% homology, indicating possible long-term outbreaks caused by clonal transfer of S. Enteritidis strains within the hospital or recurring introduction and spread of Salmonella in the hospital by colonized humans or contaminated food (Blahova et al., 1998). CRSE long term presence in animal breeding and especially fattening lots has been described (Ramchandani et al., 2005; Parveen et al., 2007). On the other hand, the limited discriminatory power of PFGE for several Salmonella serovars is known; this might falsely indicate clonal transfer. Therefore, WGS-based phylogenetic analyses are needed in future to assess the genetic relationship of the CRSE isolates in more detail.

WGS analysis of the plasmid 44C, which was from CRSE isolate of a diarrhea patient, showed that it had genes associated with plasmid replication, conjugation transfer, inhibition, intracellular multiplication, segregation and stability, functional genes that confer the ability to transfer bla_CTX–M–55 and integrate into the genome of other bacteria. Moreover, compared with other bla_CTX–M–55-harboring plasmids, we found that plasmid 44C showed high sequence similarity with plasmid p628-CTX-M isolated from a K. pneumoniae strain in October 2010 (Wang et al., 2015). In particular, both plasmids shared an ISEcp1-related component in the genetic context around bla_CTX–M–55, indicating that such structures may transfer from K. pneumoniae to Salmonella. Thus, it is crucial to implement precautionary measures to prevent the spread of bacteria with these mobile resistance genes in both clinical settings and breeding industry by strengthening the detection and analysis of this mobile genetic element.

In conclusion, increase of cefepime resistance in S. Enteritidis is a serious public health concern. Our data demonstrate that the bla_CTX–M–55 gene mobilized by the ISEcp1-bla_CTX–M–55-ORF477 transposition unit is the main feature shared by CRSE and plays an important role in the transmission among different species.

The data sets generated for this study can be found in the NCBI Sequence Read Archive (SRA) under accession number SRR9734360 (https://www.ncbi.nlm.nih.gov/search/all/?term=SRR9734360).

Ethical approval for this study was provided by Shanghai Municipal Center for Disease Control and Prevention (Shanghai, China). This study is a retrospective study and individual patient identification is not accessed and informed consent is not required.

YF, JZ, and ML conceived and designed the experiment. YF and YM carried out the experiment. YF, LZ, ZX, YM, YW, ZC, and JB contributed to sample preparation and collected the data. YF, LZ, ZX, and YM performed the analysis. YF wrote the manuscript. JZ and ML revised the manuscript. XX provided the samples. ML and JZ fund the project. All authors discussed the results and commented on the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.