PDF
Citation
Print
INTRODUCTION
Salmonella is a major cause of foodborne diseases and clinical infections, causing approximately 155,000 deaths worldwide annually and the third-leading cause of diarrhea-related mortality. The genus Salmonella includes two species: Salmonella enterica and Salmonella bongori. S. enterica has six subspecies (I–VI), and subspecies I (enterica) is further categorized into 1,586 serotypes (e.g., Typhi, Typhimurium, and Enteritidis), representing unique antigenic formulas of the O and H antigens [1]. For simplicity, S. enterica serotypes are categorized as typhoidal, paratyphoidal, and nontyphoidal. While Salmonella Typhi and Salmonella Paratyphi cause severe illness, nontyphoidal Salmonella (NTS) usually causes self-limiting illness. However, up to 5% of infections caused by NTS progress into extra-intestinal invasive diseases such as bacteremia that, when left untreated, have case fatality ratios of 20%–25% [2, 3].
In a 2017 review, the reported occurrence of invasive NTS (iNTS) infections across Africa was 1.4 per 100,000 population per year (all ages, South Africa, 2003–2004), and the prevalence of NTS-related community-acquired bacteremia ranged from 8% in Nigeria and South Africa to 45% in the Central African Republic [4]. The epidemiology of iNTS in Asia, including China, Hong Kong, Vietnam, and Korea, has also been reported [5–9]; however, significant gaps remain in the understanding of the genetic factors associated with extra-intestinal invasion.
Numerous Salmonella virulence genes have been identified, including those involved in the effector delivery system, adhesion, immune modulation, motility, nutrition/metabolism, regulation, biofilm formation, stress survival, exotoxin, antimicrobial activity/competitive advantages, invasion, exoenzyme, and post-translational modification, and uncategorized genes [10]. These genes enable bacteria to adapt to the host environment and to resist and overcome the host’s defensive immunity. Hence, virulence gene profiling is useful to assess the potential pathogenicity of bacteria. Despite these insights, data on the genetic diversity and specific virulence factors associated with invasive NTS infections in Korean pediatric populations are limited. Therefore, we investigated the genetic diversity and explored the genomic correlates of extra-intestinal invasion in Korean children by performing whole-genome sequencing (WGS) on 58 NTS isolates collected from Korean children.
MATERIALS AND METHODS
Study population
This was a retrospective observational study of Salmonella species isolated from stool or blood samples of patients <18 yrs of age diagnosed with moderate-to-severe salmonellosis that were prospectively collected from a tertiary referral university hospital in South Korea between January 2019 and December 2021. The inclusion criteria were as follows: 1) patients with Salmonella species isolated from stool or blood cultures that were collected and stored; 2) age <18 yrs; 3) having moderate-to-severe salmonellosis, defined as requiring admission and/or intravenous fluid therapy. Cases that did not meet the inclusion criteria were excluded. Cases of extra-intestinal invasive disease included 1) sepsis, defined as fever and Salmonella isolated from blood cultures, and 2) appendicitis with peritonitis and/or intestinal perforation associated with Salmonella infection, as confirmed by clinical symptoms and abdominal computed tomography. This study was approved by the Institutional Review Board of Seoul St. Mary’s Hospital, Seoul, Korea (approval No.: KC22 RISI0669). The requirement for informed consent was waived because of the retrospective study design.
Identification and antimicrobial susceptibility testing
Salmonella isolates were identified to the genus level using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry, and the O antigen serogroup was determined using antisera (Joongkyeom Co., Ltd., Goyang-si, Korea). Detailed information is provided in the Supplemental Data. H2S production was examined using triple sugar iron slants (Asan Pharmaceutical, Seoul, Korea) to differentiate Salmonella Typhi. Antimicrobial susceptibility was tested for ampicillin, trimethoprim-sulfamethoxazole (TMP/SMX), ciprofloxacin, and cefotaxime using the disk diffusion method. Antibiotic susceptibility cutoffs were determined following CLSI guidelines [11]. Multidrug resistance (MDR) was defined as resistance to ampicillin, chloramphenicol, and TMP/SMX [12].
WGS and molecular typing
Isolates were cultured on 5% sheep blood agar plates (Bandio Bio Science, Pocheon, Korea). Genomic DNA was extracted using the QIAamp DNA Mini Kit (Qiagen, Hilden, Germany). Sequencing data acquisition and processing were performed as previously described [13–15]. Briefly, next-generation sequencing libraries were prepared using the TruSeq Nano DNA sample preparation kit (Illumina, San Diego, CA, USA) and sequenced on an Illumina NovaSeq instrument. Sequencing reads were assembled into contigs using SPAdes [16], and the contigs were annotated using Prokka [17]. The raw sequencing reads have been deposited in the Sequence Read Archive under accession No. PRJNA1065147. Detailed information regarding the WGS analysis is provided in the Supplemental Data.
Salmonella pathogenicity islands (SPIs) were identified using SPIFinder [18], and serotypes were determined using SeqSero2 [19]. The acquisition of antimicrobial resistance genes, resistance-conferring mutations, and virulence factors were determined using ABRicate, employing the Comprehensive Antibiotic Resistance Database [20] and Virulence Factor Database [21]. Virulence genes were classified into 14 categories according to previous reports [10, 22]. Details of genome-based molecular typing are provided in Supplemental Data Table S1.
Statistical analyses
Clinical variables were compared between groups using Fisher’s exact test for categorical variables (sex, immunocompromised, fever, diarrhea, hematochezia, vomiting, sepsis, and appendicitis with peritonitis) and using the Mann–Whitney U-test for continuous variables (age, fever duration, diarrhea duration, white blood cell, hemoglobin, erythrocyte sedimentation rate, C-reactive protein, aspartate transaminase, alanine transaminase, blood urea nitrogen, creatinine, sodium, potassium, and chloride). Univariate analysis for genetic factors was performed using Fisher’s exact test with risk estimation function. Stepwise multiple logistic regression (backward likelihood ratio method) was used for multivariate analysis. Variables with P<0.05 in univariate analysis (SPI-1, SPI-2, SPI-5, SPI-13, STM0271, STM0275, cytolethal distending toxin [CDT]-related genes, fepC, and tcpC) were included as covariates. All tests were two-tailed, and statistical significance was set at P<0.05. All statistical analyses were performed using SPSS version 26 (IBM Corp., Armonk, NY).
RESULTS
Demographics and clinical characteristics
Over the 3-yr study period, 61 children were confirmed to be infected with Salmonella species based on stool cultures. All isolates were identified as NTS. Three isolates failed to grow in subculture; therefore, 58 cases were included in further analyses. The mean patient age was 6.0 yrs (95% confidence interval [CI], 5.1–7.0), and 43.1% (N=25) were male (Table 1). Among the patients, 10.3% (N=6) had diseases causing immunosuppression, such as ALL (N=3), AML (N=3), and cyclic neutropenia (N=1). Fever was present in 94.8% (N=55) of the patients, with a mean duration of 4.8 days (95% CI, 4.1–5.5). In addition, 87.9% (N=51) of the patients experienced diarrhea, with a mean duration of 5.3 days (95% CI, 4.5–6.1). Vomiting was observed in 46.6% (N=27) of patients, and hematochezia in 32.8% (N=19) (Table 1). Of the patients without diarrhea, all had acute gastrointestinal symptoms of fever and abdominal pain (N=5/7) or vomiting (N=2/7) and stool cultures positive for NTS.
Extra-intestinal invasion was observed in 20.7% (N=12) of patients, with 83.3% (10/12) and 16.7% (2/12) presenting with sepsis and appendicitis with perforation, respectively. Comparison of the clinicopathological characteristics between patients with extra-intestinal invasion and those with non-invasive disease revealed no significant differences in the proportions of patients with immunocompromised underlying diseases or in initial clinical symptoms, such as fever duration, diarrhea duration, and hematochezia. Regarding the laboratory findings, patients with extra-intestinal invasion exhibited lower levels of C-reactive protein (P=0.021) and AST (P=0.002) and higher levels of ALT (P=0.003) than those with non-invasive disease (Table 1).
Genomic characteristics of NTS isolates
We analyzed 58 isolates in total, of which 11 isolates were isolated in 2019 (19.0%), 18 in 2020 (31.0%), and 29 in 2021 (50.0%). The isolates were categorized into 20 Salmonella serotypes, with Enteritidis being the most prevalent (N=21; 36.2%), followed by Infantis (N=6; 10.3%), Bareilly (N=5; 8.6%), and I 4,[5],12:i:- (N=5; 8.6%) (Fig. 1 and Supplemental Data Table S2). These four serotypes constituted 63.8% of all isolates collected, representing the predominant salmonellosis population in children in South Korea, regardless of invasiveness. Notably, Enteritidis consistently dominated in each year, comprising 27.3% in 2019, 44.4% in 2020, and 34.5% in 2021, indicating its continuous circulation in South Korea (Supplemental Data Table S1). Excluding Enteritidis, no evidence of an outbreak caused by a single circulating serotype was observed.
The isolates were classified into 24 distinct sequence types (STs), the most prevalent being ST11 (N=21; 36.2%), followed by ST32 (N=6; 10.3%), and ST203 (N=5; 8.6%). Upon closer examination of serotypes and STs, all isolates of Enteritidis, Infantis, and Bareilly were ST11, ST32, and ST203, respectively (Fig. 1). In contrast, I 4,[5],12:i:- isolates comprised four STs: ST34 (N=2), ST19 (N=1), ST36 (N=1), and ST2379 (N=1), indicating a heterogenous genomic background. Analysis of the evolutionary relationships among these STs revealed that ST19 and ST2379 were single-locus variants of ST34, whereas ST36 was a quadruple-locus variant of ST34.
Salmonella Enteritidis-ST11 was the predominant circulating population among children in South Korea; however, it was less common in patients with extra-intestinal invasion (1/12) than in those without (20/46) (P=0.004; odds ratio [OR]=0.118; 95% CI, 0.014–0.993; Fig. 2). Only one isolate of Salmonella Infantis (ST32; 1/6) was found in patients with extra-intestinal invasion. In contrast, all isolates of Salmonella Oranienburg (ST23; 2/2), Give (ST516; 1/1), Javiana (ST1547; 1/1), Paratyphi B var. L(+) tartrate+ (ST43; 1/1), Schwarzengrund (ST96; 1/1), Singapore (ST501; 1/1), and half of Montevideo (ST4; 1/2), Saintpaul (ST50; 1/2), and I 4:b: (ST2814; 1/1) isolates were detected in patients with extra-intestinal invasion, although statistical significance was not reached because of the small number of cases.
Genetic factors associated with an increased risk of extra-intestinal invasion
We identified 175 virulence genes in the 58 NTS isolates, with an average of 142.2 (range, 123–166) per isolate. Among them, 110 genes were ubiquitous, whereas the remaining 65 genes were detected at variable rates: 12 genes were found in 80%–99% of isolates, nine in 60%–80%, 21 in 40%–60%, 15 in 20%–40%, and eight in 0%–20% (Fig. 2A). Virulence gene profiles showed identical or similar patterns within each Salmonella serotype, with some virulence genes being almost exclusively present in certain serotypes (Fig. 2B). When we compared the distribution of virulence genes, no significant difference was observed between iNTS and non-iNTS groups regarding the numbers of total virulence genes and genes belonging to major virulence categories. However, at the individual gene level, several genes, including CDT-related genes (cdtB, pltA, and pltB), fepC, tcpC, STM0271, and STM0275, were more prevalent in patients with extra-intestinal invasion than in those without: the P-value was extremely low for CDT-related genes (P=1.4×10–4, OR= 45.000, 95% CI 4.59–440.81 for CDT-related genes; P=0.021, OR=9.240, 95% CI 1.10–77.58 for fepC; and P=0.040, OR= infinity for tcpC) (Fig. 3 and Table 2). In addition, the five genetic factors (CDT-related genes, fepC, tcpC, STM0271, and STM0275) were significantly more frequent in invasive isolates (mean, 3.08, 95% CI 2.17–4.00) than in non-invasive isolates (mean, 1.37, 95% CI 0.98–1.76; P=5.4×10–4) (Supplemental Data Table S3).
We identified 13 SPIs in the 58 NTS isolates, with each isolate containing at least one SPI. These included SPI-1 (eight islands), SPI-2 (14 islands), SPI-3 (five islands), SPI-4 (three islands), SPI-5 (four islands), SPI-8 (one island), SPI-9 (one island), SPI-10 (one island), SPI-13 (three islands), SPI-14 (two islands), centisome 63 pathogenicity island (C63PI), CS54 island, and SGI1 (Supplemental Data Table S1). Notably, SPI-1 (P=0.039; OR= 4.762; 95% CI 1.136–19.957), SPI-2 (P=0.020; OR=5.625; 95% CI 1.332–23.757), SPI-5 (P=0.014; OR=7.500; 95% CI 1.609–34.954), and SPI-13 (P=0.010; OR=11.000; 95% CI 1.311–92.299) were significantly more frequent in isolates causing extra-intestinal invasion than in non-invasive isolates (Fig. 3 and Table 2).
Antimicrobial susceptibilities and resistomes
Overall, 31.0% (18/58), 6.9% (4/58), 1.7% (1/58), and 6.9% (4/58) of isolates exhibited resistance to ampicillin, chloramphenicol, cefotaxime, and TMP/SMX, respectively. 36.2% (21/58) of Salmonella isolates were not susceptible to ciprofloxacin, but none were resistant (Supplemental Data Table S1). Notably, non-susceptibility to ciprofloxacin (P=3.0×10–6; OR= 20.48; 95% CI 5.17–81.18) was significantly enriched in Salmonella Enteritidis-ST11, whereas resistance to TMP/SMX and chloramphenicol was more prevalent in I 4,[5],12:i:- than in other serotypes (P=0.001; OR=78.00; 95% CI 5.41–1,123.71) (Fig. 3). A total of 6.9% (4/58) NTS isolates were MDR, of which Salmonella I 4,[5],12:i:- (ST34, ST36, and ST2379) accounted for 75.0% (3/4; P=0.001; OR=78.00; 95% CI 5.41–1,123.71), and Stanley-ST29 for 25.0% (1/4). One Salmonella Enteritidis-ST11 isolate was resistant to cefotaxime. We found no significant association between antimicrobial resistance and extra-intestinal invasion.
Among the isolates phenotypically resistant to ampicillin, 83.3% (15/18) harbored blaTEM-1, and one cefotaxime-resistant isolate harbored blaCTX-M-15. For chloramphenicol-resistant isolates, 75.0% (3/4) carried floR, and 25.0% (1/4) carried both floR and cmlA1. Regarding ciprofloxacin non-susceptible isolates, 81.0% (17/21), 9.5% (2/21), 4.8% (1/21), and 4.8% (1/21) harbored gyrA mutations, qnrS1 mutations, parC mutations, and mutations in both qnrS1 and parC, respectively. Among TMP/SMX-resistant isolates, sul2 was the most prevalent resistance gene (75.0%; 3/4), and all (100.0%; 4/4) isolates harbored more than one TMP/SMX resistance-conferring gene (Fig. 3).
DISCUSSION
NTS typically causes self-limiting acute gastroenteritis in children; however, virulent isolates may exhibit extra-intestinal invasion via the vascular or lymphatic system or direct invasion [23]. To the best of our knowledge, this study represents the first endeavor to estimate the incidence of iNTS infections in children in Korea, revealing a rate of 20.7% (N=12) in children with moderate-to-severe salmonellosis. Comparing invasive and non-invasive infections, no significant differences in initial clinical symptoms, laboratory findings, or presence of immunocompromised conditions were observed between patients with extra-intestinal invasion and those with non-invasive disease. In addition, we found no associations between any particular serotype and extra-intestinal invasion, which may be because the genomes within each Salmonella serotype were highly homogeneous. However, nine virulence factors, including SPI-1, 2, 5, and 13, were more frequent in isolates causing extra-intestinal invasion than in those causing non-invasive disease, with the CDT-related genes (cdtB, pltA, and pltB) being the most significant. The serotypes of the 12 invasive NTS isolates were as follows: two isolates of Oranienburg and one isolate each of Javiana, Paratyphi B var. L(+) tartrate+, Give, Saintpaul, Enteritidis, I 4:b:-, Infantis, Montevideo, Singapore, and Schwarzengrund (Supplemental Data Table S3).
iNTS infections are a major global health issue, particularly in regions such as sub-Saharan Africa and Asia, where case fatality rates are the highest [8]. The serotypes Typhimurium and Enteritidis are the most frequently implicated in iNTS infections in humans. Less common serotypes, such as Choleraesuis, Newport, Brancaster, Freetown, and Infantis, have been identified in smaller numbers in Africa and Asia [24]. The United States has witnessed a substantial increase in the incidence of Javiana-related illnesses since 1996, making it the fourth most common NTS serotype in this region [25, 26]. Salmonella Javiana is considered a highly virulent serotype because it can produce CDT [27]. This is consistent with our finding that CDT genes were the most frequently detected in invasive isolates and may explain why Enteritidis and Typhimurium, both lacking CDT genes, were less frequently associated with invasive disease in our study.
CDTs are important virulence factors produced by gram-negative bacteria, including both extracellular and intracellular pathogens (e.g., Campylobacter spp., Escherichia coli, Helicobacter spp., Salmonella enterica, Shigella spp., and Yersinia spp.) [28]. In vivo, CDT-induced damage has been associated with enhanced host colonization and chronic infections [29]. We identified CDT-related genes in the serotypes Javiana, Oranienburg, Give, Montevideo, and Schwarzengrund, all of which harbor CDT-related genes [30]. Of the remaining six isolates that caused bacteremia, all but one isolate (Salmonella Enteritidis) harbored four or more genes that were significantly more frequent among bacteremic isolates than among non-bacteremic isolates. Salmonella uses CDTs to establish persistent infections by altering host immunity. Mezal, et al. [31] demonstrated that Salmonella Javiana, harboring cdtB, pltA, and pltB, induced cell-cycle arrest, cytoplasmic distension, and nuclear enlargement in host cells and concluded that CDT production may play an important role in pathogenesis. Similarly, in the present study, most children infected with NTS carrying cdtB, pltA, and pltB displayed extra-intestinal invasion, with 83.3% exhibiting sepsis and 16.7% presenting with appendicitis accompanied by intestinal perforation. We also found that all iNTS isolates carrying the cdtB, pltA, and pltB genes belonged to clade B, a subpopulation of Salmonella enterica [30, 32]. Therefore, early detection of cdtB in NTS clade B isolates and aggressive treatment may reduce progression to severe disease.
We identified Salmonella Enteritidis-ST11, a major contributor to foodborne illnesses worldwide and commonly associated with poultry products [33, 34], as a major circulating clone (36.2%) causing human salmonellosis in South Korea. Notably, this clone is the most frequently isolated from chickens [35]. In Europe, Salmonella Enteritidis-ST11 was identified as the cause of three clustered outbreaks of salmonellosis associated with chicken products [36]. The emergence of MDR ST11 isolates has been reported [6, 37], and our findings also revealed a high level of resistance to ampicillin (47.6%, 10/21) and non-susceptibility to ciprofloxacin. Three-quarters of the MDR isolates were identified as serotype I 4,[5],12:i:-, which has been reported as an emerging MDR NTS [38]. These results underscore the importance of close surveillance of increases in Salmonella Enteritidis-ST11 and I 4,[5],12:i:- in both foodborne and human-transmitted infections.
Extra-intestinal invasion by NTS can lead to life-threatening complications, including bacteremia, endocarditis, meningitis, and osteomyelitis. Infants, older adults, and immunocompromised patients are recognized as at-risk populations for these complications [8]. Nevertheless, depending on the virulence potential of NTS, extra-intestinal invasion may occur in patients with immunocompetence [39], as observed in our study. Notably, we observed similarities in the initial symptoms and laboratory findings between patients with extra-intestinal invasion and those with non-invasive disease. These results emphasize that relying solely on clinical suspicion may not be sufficient for early therapeutic intervention to prevent disease progression. Therefore, understanding the genetic makeup that contributes to virulence potential is crucial for early detection and differentiation.
The small number of cases included in this study constitutes a possible limitation. In addition, we did not perform any functional assays related to the virulence of NTS. Nevertheless, this study stands out as one of the most extensive investigations on virulence factors associated with extra-intestinal invasion in children to date and may provide a basis for understanding the genomic landscape of NTS isolates in South Korea. Future studies with larger sample sizes and functional validation will help elucidate the definitive mechanisms of NTS invasiveness.
In conclusion, Salmonella Enteritidis-ST11 stands out as the dominant circulating cause of NTS in children in South Korea and exhibits high levels of resistance to ampicillin and ciprofloxacin. Our findings also revealed several genetic factors associated with extra-intestinal invasion. Although genetic markers, including CDT-related genes, have been associated with extra-intestinal invasion, further studies are required to validate their potential for early diagnosis. Confirmation of these findings could support timely targeted treatment and enable closer monitoring of chronic infections.
XML Download