Diarrheagenic Escherichia coli Phylogroups Are Associated with Antibiotic Resistance and Duration of Diarrheal Episode

Conventionally, in Escherichia coli, phylogenetic groups A and B1 are associated with commensal strains while B2 and D are associated with extraintestinal strains. The aim of this study was to evaluate diarrheagenic (DEC) and commensal E. coli phylogeny and its association with antibiotic resistance and clinical characteristics of the diarrheal episode. Phylogenetic groups and antibiotic resistance of 369 E. coli strains (commensal strains and DEC from children with or without diarrhea) isolated from Peruvian children <1 year of age were determined by a Clermont triplex PCR and Kirby-Bauer method, respectively. The distribution of the 369 E. coli strains among the 4 phylogenetic groups was A (40%), D (31%), B1 (21%), and B2 (8%). DEC-control strains were more associated with group A while DEC-diarrhea strains were more associated with group D (P < 0.05). There was a tendency (P = 0.06) for higher proportion of persistent diarrhea (≥14 days) among severe groups (B2 and D) in comparison with nonsevere groups (A and B1). Strains belonging to group D presented significantly higher percentages of multidrug resistance than the rest of the groups (P > 0.01). In summary, DEC-diarrhea strains were more associated with group D than strains from healthy controls.


Introduction
Conventionally, Escherichia coli, a common isolate in clinical laboratories, is classified into two major groups: commensal and pathogenic. Additionally pathogenic isolates may produce different diseases, being then subdivided in diarrheogenic and extraintestinal E. coli. Human infections caused by extraintestinal E. coli include meningitis, urinary tract infections, sepsis, pneumonia, surgical site infections, and infections in other extraintestinal locations [1]. However, when classified into subtypes, E. coli mainly fall into four phylogenetic groups: A, B1, B2, and D [1]. Previous studies have shown that commensal E. coli strains tend to be associated within phylogenetic groups A and B1 [1,2], whereas the extraintestinal pathotypes fall within phylogenetic groups B2 and D [3,4]. Regarding uropathogenic E. coli (UPEC) strains determinants including phylogroups markers are well established. However, for diarrheagenic E. coli (DEC), the scenario remains unclear. There is no information about the association between the phylogenetic group and the clinical data of the diarrheal episode [5].
Previous reports describe the emerging antibiotic resistance in commensal and diarrheagenic E. coli in Peru [6][7][8][9][10]. However, there is no sufficient data in the correlation of phylogeny and antibiotic resistance [11]. Therefore we conducted this study to determine the association between the phylogenetic group and antibiotic resistance in a large number of E. coli (commensal and diarrheagenic) strains from Peruvian infants.

Samples.
Commensal and diarrheagenic E. coli strains were isolated during a previous passive surveillance diarrhea 2 The Scientific World Journal study [12]. In this study 1032 children were followed from 2 to 12 months of life, obtaining a total of 1079 E. coli strains that were analyzed by a real time multiplex PCR to determine DEC pathotypes [13]. A total of 369 isolates from this study were randomly selected and analyzed, including 74 commensal E. coli, 94 DEC from asymptomatic children (DEC-control), and 201 DEC isolated from children with diarrhea (DEC-diarrhea). DEC pathotypes included in this study were enteropathogenic E. coli (EPEC), enterotoxigenic E. coli (ETEC), enteroaggregative E. coli (EAEC), and diffusely adherent E. coli pathotypes (DAEC); Shiga toxin-producing E. coli (STEC) and enteroinvasive E. coli (EIEC) strains were not included due to their low prevalence [12].

Phylogenetic Group Determination.
The phylogenetic groups were determined as previously described [14]. In all cases the bacteria DNA was extracted boiling.

Clinical Data of Diarrheal
Episodes. Variables such as episode duration (days), maximum number of stools per day, total number of stools per episode, and a modified Vesikari modified score [15] were analyzed and associated with phylogenetic groups in those cases in which no coinfections were previously reported [12].

Statistical
Analysis. Vesikari severity score was expressed by mean ± standard deviation and median (range) values were given for duration of the episode and number of stools. The comparisons between groups were made using chisquared or Fisher's exact test. Student's -test was used for the comparison of Vesikari severity scores between groups.

Clinical Data of Diarrheal Episodes.
From the 201 DECdiarrhea isolates analyzed in the study, 127 strains were isolated from diarrhea episodes in which no other pathogen was detected. In these 127 patients, no significant differences were found for the studied variables among the four phylogenetic groups. In general, the episode duration was 5 days (1-25), the maximum number of stools/day was 5 (3)(4)(5)(6)(7)(8)(9)(10)(11), the total number of stools/episode was 20 , and the Vesikari score was 6 ± 2.6. We found a higher proportion ( = 0.06) of persistent diarrhea (14 or more days) among B2-D groups (23.9%) compared to among A-B1 groups (9.88%). No differences were found for either acute or prolonged diarrhea (7-14 days) between severe (B2-D) and nonsevere (A-B1) groups.

Antibiotic Resistance.
Resistance to trimethoprim-sulfamethoxazol, tetracycline, chloramphenicol, and nalidixic acid and multiresistance were significantly different among the four phylogenetic groups ( < 0.05) (Figure 2). In general, B2 and D groups presented higher percentage of antibiotic resistance than A and B1 groups. In the case of multiresistance D group presented significantly higher percentages than the rest of the groups ( < 0.01) (Figure 2). Table 1 Phylogroup

Discussion
The E. coli phylogroups differ in their ecological niches, lifehistory characteristics, and propensity to cause disease. In this manner B2 and D groups are less frequently isolated from the environment [17]. Regarding human illness, E. coli isolates recovered from extraintestinal body sites are more likely to belong to B2 or D phylotypes than to A or B1 [18,19]. However, differences in the prevalence of the different phylogenetic groups among virulent extraintestinal E. coli have been observed in previous studies [20]. Despite the fact that the two phylogenetic groups most frequently related with virulence in the extraintestinal E. coli are the aforementioned B2 and D, some reports have shown a high frequency of group A (46%) among E. coli causing urinary infections [20]. Alternatively, some reports showed that gut commensal E. coli were mostly related with group A [14]. In addition, in a previous study in children in Costa Rica, commensal E. coli were related to phylogenetic groups A and D (36%), while the studied DEC belong to B1 (35%), A (29%), B2 (23%), and D (14%) [19]. Another study analyzing DEC (EAEC, EPEC, and STEC) isolated from children in Rumania showed that 51% of the strains belonged to group A, followed by 23% of the strains that belonged to group B2 [21]. In the present study, commensal strains were more associated with A and D groups as has been previously reported [19]. Phylogenetic groups B2 and D have been related with virulence factors that cause infections at an extraintestinal level. Additionally, the B2 strains have been shown to persist for longer periods in infants than other E. coli strains [22].
Previous studies have tried to relate the clinical data of infection with the phylogenetic groups [23]. However, these types of studies have not addressed intestinal pathogenic strains that cause diarrhea. In this report, no significant differences were found for the clinical variables analyzed among the four phylogenetic groups. However, we found a greater tendency of persistent diarrhea in phylogenetic groups B2 and D (24%) than in groups A and B1 (10%). Regarding isolates belonging to the B2 group, this long persistence has been previously observed [22], while no data has been found regarding other phylogroups. In this line an association between the diarrheagenic pathotype DAEC and persistent diarrhea was also observed, in accordance with what has been previously reported [24]. Interestingly, DAEC isolates belong largely to the D phylogroup.
When we analyzed the relation between phylogenetic group and antibiotic resistance we found that the isolates belonging to the group D were more related with multiresistance than those belonging to other groups. Although previous reports in extraintestinal strains showed that virulencerelated phylogenetic groups, especially B2, were associated with low levels of antibiotic resistance [11] our data showed a different scenario, in which B2 and D isolates were those exhibiting high levels of antibiotic resistance. A possibility to take into account is the possibility that as phylogroup D results in more prolonged diarrhea, the use of antibacterial agents may be needed, and then these isolates may be under a more intense antibiotic pressure which may facilitate the acquisition of antibiotic resistance mechanisms.
DEC accounts for more than 120,000 deaths/year, being involved, together with rotavirus in around 40% of all diarrhea related children deaths [25]. Recent studies have showed that both EPEC and ETEC isolates are related with an increased mortality [26]; in fact, it is considered that ETEC isolates account for more than 40,000 deaths each year. In this sense it is of interest to note that in the present study most of the ETEC or EPEC isolates belong to the phylogroup A, classically considered as a low-virulent phylogroup. Studies on extraintestinal E. coli have showed that the development of antibiotic resistance may be correlated with a decreased virulence [27]. Thus, although no clear reason may be stated to explain the high percentage of DEC isolates belonging to low-virulent phylogroups, the high levels of antibiotic resistance present in the area and these described inverse relations between virulence and antibiotic resistance may be suggested as a potential explanation.
One study limitation is the fact that specific virulence factors encoding genes were not studied; therefore, a direct relationship between virulence and resistance was not possible to evaluate. Furthermore, multilocus sequence typing (MLST) method is more specific than the Clermont triplex PCR method used in this study [28]. However, the Clermont triplex PCR method is a rapid and cost-effective method that has been used extensively. Moreover, this method is able to be implemented in resource constraining sites. A recent study indicates that strains belonging to cryptic lineages of Escherichia are the more related to failure by this triplex method; however, in human fecal samples these lineages are unlikely to be found (2-3% frequency) [29].

Conclusion
The present data show the relationships between E. coli phylogenetic lineages, diarrheagenic character, and severity The Scientific World Journal 5 of illness. Moreover, an association between phylogroup D and prolonged disease duration has been found. Interestingly, the isolates belonging to this phylogroup also presented the higher levels of multiresistance, showing that those isolates able to cause a more prolonged disease also possess higher levels of antibiotic resistance, probably because they usually required antibiotic treatment and then are under a more intense antibiotic pressure.
Further investigations to elucidate the relationship between phylogeny, specific virulence factors, and mechanisms of resistance are needed in order to better understand DEC and commensal E. coli strains.