Genetic Characterization of Intimin Gene (eae) in Clinical Shiga Toxin-Producing Escherichia coli Strains from Pediatric Patients in Finland

Shiga toxin (Stx)-producing Escherichia coli (STEC) infections cause outbreaks of severe disease in children ranging from bloody diarrhea to hemolytic uremic syndrome (HUS). The adherent factor intimin, encoded by eae, can facilitate the colonization process of strains and is frequently associated with severe disease. The purpose of this study was to examine and analyze the prevalence and polymorphisms of eae in clinical STEC strains from pediatric patients under 17 years old with and without HUS, and to assess the pathogenic risk of different eae subtypes. We studied 240 STEC strains isolated from pediatric patients in Finland with whole genome sequencing. The gene eae was present in 209 (87.1%) strains, among which 49 (23.4%) were from patients with HUS, and 160 (76.6%) were from patients without HUS. O157:H7 (126, 60.3%) was the most predominant serotype among eae-positive STEC strains. Twenty-three different eae genotypes were identified, which were categorized into five eae subtypes, i.e., γ1, β3, ε1, θ and ζ3. The subtype eae-γ1 was significantly overrepresented in strains from patients aged 5–17 years, while β3 and ε1 were more commonly found in strains from patients under 5 years. All O157:H7 strains carried eae-γ1; among non-O157 strains, strains of each serotype harbored one eae subtype. No association was observed between the presence of eae/its subtypes and HUS. However, the combination of eae-γ1+stx2a was significantly associated with HUS. In conclusion, this study demonstrated a high occurrence and genetic variety of eae in clinical STEC from pediatric patients under 17 years old in Finland, and that eae is not essential for STEC-associated HUS. However, the combination of certain eae subtypes with stx subtypes, i.e., eae-γ1+stx2a, may be used as risk predictors for the development of severe disease in children.


Introduction
Shiga toxin (Stx)-producing Escherichia coli (STEC) is a major cause of human gastrointestinal diseases ranging from mild, watery to bloody diarrhea.The severe complications associated with toxin production and release can cause the life-threatening hemolytic uremic syndrome (HUS), leading to kidney failure and neurological episodes [1].It has been reported that 6-25% of patients infected with STEC develop HUS, and the rate is higher in children [2][3][4].More than 1000 different serotypes of STEC have been defined in humans, animals and the environment [5][6][7].O157:H7 is the predominant serotype causing more severe disease.However, non-O157 serogroups such as O26, O45, O103, O111, O121, and O145 (referred to as the 'top six' non-O157 STEC) have been increasingly reported in outbreaks and sporadic cases in recent years in different countries like Europe and North America [8][9][10].
Studies on the molecular characteristics of eae in clinical STEC strains are limited, especially in pediatric patients, and the relationship between eae subtypes and clinical symptoms remains to be addressed.In this study, we investigated the subtypes and polymorphisms of eae among clinical STEC strains from pediatric patients under 17 years old in Finland, and assessed the association of eae subtypes with clinical symptoms (HUS and non-HUS).

Comparison with eae-Positive STEC Strains from Swedish Pediatric Patients
In a previous study in Sweden [21], eae-γ1 was found to be statistically overrepresented in clinical eae-positive STEC strains from HUS patients, while β3 was significantly higher in non-HUS STEC strains from both adults and children.In addition, eae was found to be more prevalent in strains from children; however, the relationship between eae subtypes and clinical outcomes in strains from children was unclear in Sweden.We were interested to know if the correlation between eae subtype and clinical outcome was age-dependent.We then extracted 110 eae-positive strains from Swedish patients under 17 years old, and performed the same statistical analysis on the Swedish strains, as well as all 319 strains, including 110 from Swedish patients and 209 from Finnish patients within the same age group (under 17 years).The results showed that the presence of eae and the eae-γ1 subtype was significantly higher in strains from patients with HUS compared to strains from patients without HUS under 17 years old in Sweden (p = 0.0005 and p < 0.0001) (Table 4).The same correlation was observed in merged strains from Sweden and Finland under 17 years old (Table S5).eae-β3 was significantly higher in strains from non-HUS-STEC patients in Sweden, and in merged non-HUS-STEC strains from the two countries compared to HUS-STEC strains (p = 0.0001 and p = 0.0033) (Tables 4 and S5).

Discussion
Over 90% of pediatric HUS cases are caused by gastrointestinal infection with STEC [2,27,28].The reasons why children are more commonly affected than adults are unknown, but potential explanations may relate to the ability of specific strains to establish disease in specific populations, the age-specific expression on cells of receptors for toxins, or diameters of renal blood vessels.The presence of eae has been reported to be significantly associated with disease severity, and has previously been identified as a risk factor for the development of HUS.However, information about the characteristics and genetic diversity of eae and their contribution and correlation to disease severity is currently limited, especially in pediatric patients.In this study, eae was present in 87.1% of clinical pediatric strains, in accordance with previous studies, revealing that most clinical STEC strains possess eae [29,30].The majority of eae-positive strains were of serotype O157:H7 (60.3%).It is worth noting that 72.8% of non-O157 strains in this study were eae positive, which was higher than previous findings reported in England (52.5%) [31] and Sweden (62.1%) [21].However, the prevalence was reported in strains from children and adult patients together.This may indicate that eae is more prevalent in STEC strains from children than adults.The difference may be also due to the variations between geographical regions, sex, etc.
Various eae subtypes may confer distinct colonization patterns within the human intestine, thus leading to distinct pathogenic capability.STEC strains with eae-β, ε, γ1, θ, and ζ subtypes have been reported to be more virulent, and thus pose a higher pathogenic risk [21,25].Previously, eae-γ1-positive STEC strains were isolated from children with HUS in Uruguay, highlighting the clinical significance of eae-γ1 [32].In this study, five eae subtypes were identified (eae-γ1, β3, ε1, θ, and ζ3), of which eae-γ1 was the most prevalent, followed by β3 and ε1.In addition, eae-γ1 was more prevalent in children aged 5-17 years compared to children below 5 years, while β3 and ε1 were more prevalent in strains from children under 5 years, indicating that various eae subtypes tended to be present in different age groups.eae-γ1 was significantly associated with O157:H7 strains.Consistent with the previous findings, we found that eae-γ1 was carried by serogroup O157 and O145 strains; eae-β3 was carried by O26 strains; eae-ε1 was carried by O103 and O121 strains; and eae-θ was carried by O111 strains [25].Certain eae subtypes may play roles in the pathogenicity of these clinical high-risk strains, leading to different clinical manifestations.No association was found between the presence of eae/its subtypes and HUS status in Finland.Interestingly, the presence of eae and eae-γ1 was significantly associated with HUS in patients within the same age in Sweden.
STEC strains producing Stx2a or Stx2c subtypes are more associated with HUS in humans, especially with the presence of eae [33].We found that the presence of stx2a+eae-γ1 was strongly associated with HUS, while stx1a+eae-β3 and stx1a+stx2c+eae-γ1 were associ-ated with non-HUS.This supports that the combination of stx2a+eae (particularly eae-γ1) can be used in risk assessments for the development of HUS and severe disease outcomes.
To conclude, our study showed high prevalence (87.1%) and genetic diversity of eae in clinical pediatric STEC strains.No correlation was observed between the presence of the eae gene/its subtypes and HUS status in pediatric STEC strains in Finland, while the combination of stx2a+eae-γ1 was significantly higher in STEC strains from patients with HUS.

Collection of eae-Positive STEC Isolates and Metadata
STEC strains were collected from STEC-infected pediatric patients under 17 years old in Finland between 2000 and 2016, and the clinical information was retrieved from the medical records as previously described [34].Patients were classified into HUS and non-HUS groups according to clinical manifestation.Whole genome sequencing (WGS), assembly, and annotation were carried out as previously reported [35].Genetic characterization of STEC strains, including stx subtyping, determination of serotypes, the presence of intimin encoding gene eae, and the clade-8 specific SNP were carried out as described previously [35].Metadata of all STEC isolates, including strain ID, serotype, clade 8, presence of eae gene, eae subtype and genotype, stx subtype, and accession numbers of genomes, as well as clinical information of patients, including HUS status, age, and sex, are listed in Supplementary Table S1.
To examine the relationship between eae subtypes and HUS status in pediatric patients in different countries, STEC strains from STEC-infected patients under 17 years old in Sweden reported previously [21] were extracted and analyzed.

eae Subtyping
According to the genome annotation, complete eae sequences were extracted from the genome assemblies of 209 eae-positive STEC isolates, and then aligned with reference nucleotide sequences of eae subtypes downloaded from GeneBank.MEGA 11 software (Center for Evolutionary Medicine and Informatics, Tempe, AZ, USA) was used to compute the genetic distances of the eae subtypes with the maximum composite likelihood method, and a neighbor-joining phylogenetic tree was generated using 1000 bootstrap resampling.The phylogenetic tree structure and genetic distance were employed to determine the eae subtype.The diversity within each eae subtype was determined by eae GTs based on eae sequence polymorphism, as previously described [23].

Statistical Analyses
A statistical association between the existence of eae/its subtypes and strain characteristics (serotypes, stx subtypes) or clinical outcomes (HUS and non-HUS) was assessed with Fisher's exact test using R software version 4.3.1 (https://www.r-project.org)(accessed on 20 August 2023), and a p-value below 0.05 was considered statistically significant.

Institutional Review Board Statement:
The use of patients' information and the study protocol were approved by the Ethics Committee of the University of Helsinki (HUS/1274/2017) and date of approval was 20 April 2017.
Informed Consent Statement: Patient consent was waived due to retrospective review of the patients' medical records.Patient data were anonymous, and no consent was required to work with the bacterial strains.

Figure 1 .
Figure 1.Phylogenetic relationships of 23 unique eae sequences identified in this study (indicated in bold and different colors) and 30 reference sequences representing different eae subtypes based on the neighbor-joining method.For the branches that represent the identified 23 eae subtypes in this study, the tips are labeled in the order of eae subtype/genotype, serotype, stx subtype, and HUS status.The number of the strains are shown in brackets.Scale bar indicates genetic distance.
The association was analyzed between eae gene and serotypes (O157 and non-O157), clinical symptoms (HUS and non-HUS), age groups (<5 years and 5-17 years; the age was grouped according to the median age of 5 years), sex (male and female).The number represents the number of eae-positive strains, and percentage was shown in brackets.* Statistically significant difference. #

Table 2 .
Association between eae subtypes and serotypes or age or stx subtypes # .

Table 3 .
Association between stx+eae and HUS status or age or serotypes # .
#The association was analyzed between eae+stx subtype and clinical symptoms (HUS and non-HUS), age groups, and serotypes.The number represents the number of strains carrying specific stx+eae subtypes, and percentage is shown in brackes.* Statistically significant difference.pos: number of positive strains.

Table 4 .
Association between eae subtypes and HUS status in pediatric patients under 17 years old in Sweden # .The association was analyzed between eae/its subtype and clinical symptoms (HUS and non-HUS) in strains from pediatric patients in Sweden.* Statistically significant difference.pos: number of positive strains. #