Two or more enteropathogens are associated with diarrhoea in Mexican children

Background Diarrhoeal diseases constitute a major public health problem, particularly in the developing world, where the rate of mortality and morbidity is very high. The purpose of this study was to conduct a 2 years and 3 months study in order to determine the prevalence of five enteropathogen diarrheogenic agents in Mexico City. Methods Faecal samples were obtained from 300 Mexican children diagnosed as positive for diarrhoea, aged > 2 to < 12 years old, and from 80 children matched for age but with no symptoms of the disease (control group). Two multiplex PCR were used to detect Escherichia coli, Salmonella spp., and Shigella spp. In addition, the two protozoan parasites Entamoeba histolytica/Entamoeba dispar and Giardia intestinalis were detected by conventional methods. Results All diarrhoeal samples were positive for one or more enteropathogens. The most common enteropathogens in diarrhoeal samples were E. histolytica/E. dispar (70.3%), Salmonella (ohio 28.3%; typhimurium 16.3%; infantis 8%; anatum 0.6%; Newport 0.3%), G. intestinalis (33%), E. coli (ETEC 13.3%; EPEC 9.3%; VTEC 8.6%; EIEC 1%) and Shigella spp. (flexneri 1.6%, sonnei 1%). Infections by two (24%) three (16%) and four (12%) pathogens were observed. Conclusion This study revealed that 52% of the patients were infected by more than one enteropathogen, notably E. histolitica/E. dispar and Salmonella ohio. These results are useful for clinicians to improve the empiric treatment used in such cases.

ical agents of diarrhoea described in epidemiological studies are transmitted as waterborne and foodborne. Some foodborne pathogens have been recently considered as emerging diseases [4], despite the fact they have been known since a long time ago. For example, outbreaks of salmonellosis have been described for many decades, and yet their incidence have increased over the last 25 years. Diarrhoeal infections can be caused by many etiological agents, but mainly by enterobacteria such as Escherichia coli, Salmonella spp., Shigella spp., Campylobacter jejuni and Vibrio cholerae; as well as parasites such as Entamoeba histolytica and Giardia intestinalis, and some rotaviruses are also important agents [5].
Escherichia coli is considered as the etiological agent for many diseases including some affecting the urinary tract and intestine. The classification of diarrhoegenic E. coli strains is based on their virulence properties, and comprises six groups: Enterotoxigenic E. coli (ETEC), Enteropathogenic E. coli (EPEC), Enteroinvasive E. coli (EIEC), Enterohaemorragic E. coli (EHEC), Enteroaggregative E. coli (EAggEC) and Diffuse Adhering E. coli (DAEC) [6]. Salmonella spp. is a facultative, gram negative, flagellated member of the Enterobacteriaceae family. The most extensive accepted classification of Salmonella strains is based on the diversity of two differentially expressed H flagellar antigens: flagellin phase I and phase II antigens (codified by fliC and fljB genes), and the O antigens of the bacterial lipopolysaccharide, both determined by serotyping [7]. Until now, 2501 serotypes have been described [8]; which turns Salmonella classification into a complex and laborious process in the clinical laboratory; therefore, several PCR based methods have recently been developed, and were reported to be a simple, highly sensitive, fast and reliable alternative when compared to traditional clinical laboratory methods [9,10].
Shigella is a Gram negative, non-motile, rod-shaped bacteria, closely related to E. coli and Salmonella, and it is the etiological agent of human shigellosis and dysentery, which is characterized by severe diarrhoea with the presence of blood in the faeces. Classification of Shigella is based on serotyping, and comprise the following groups and serotypes: Serogroup A (S. dysenteriae) 12 serotypes, serogroup B (S. flexneri) 6 serotypes, serogroup C (S. boydii) 23 serotypes and serogroup D (S. sonnei) with only one serotype [11].
Diarrhoea can also be caused by protozoa such as Entamoeba histolytica and Giardia intestinalis, these zoonotic parasites are frequently transmitted by consumption of water contaminated with infective cysts [12].
The purpose of this study was to determine the prevalence of five enteropathogen diarrheogenic agents namely E. coli, Salmonella spp., Shigella spp., E. histolytica/E. dispar and Giardia intestinalis in Mexico City. Stool specimens obtained from children patients from several communities in Mexico City were examined; the association patterns between different pathogens and its correlation with occurrence of diarrhoea were also described.

Stool samples
A total of 300 stool samples were obtained from children patients with diarrhoea from different hospitals in Mexico City (patient group); also, 80 samples were obtained from children attending schools in the surrounding area, who did not had diarrhoea in the previous 45 days (control group). All subjects aged > 2 to < 12 years, and samples were collected from September 2004 through December 2006.
The selection criteria for inclusion of children patients with diarrhoea was having at least 3 or more soft, semisolid or liquid bloody faeces within 24 hours. Also, selection was made on the basis of a questionnaire filled up by all subjects with assistance of a relative over 18 years old, providing information regarding other gastrointestinal disorders, non-related diseases, travelling, frequency of diarrhoeal episodes, previous and current antibiotic treatment; as well as general data such as age, gender and place of residence. The control and patient groups were matched for age and sex.

Bacteriology
Stool samples were streaked on the surface of MacConkey agar (DIBICO, Mexico) for obtaining E. coli isolates and on sodium deoxycholate agar for the selection of Shigella and Salmonella, and were incubated overnight at 37°C. All samples were tested for Shigella by using colony morphology, biochemical properties, and agglutination with specific antisera (Serobac, BioRad).
Multiplex PCR assays for the detection of E. coli and Shigella spp. were performed following a previously reported method [13], which selectively amplifies specific regions of several virulence genes: ETEC (eltB 322 bp, and estA 147 bp) VTEC (eaeA 376 bp, vt1 130 bp, and vt2 298 bp), EPEC (eaeA 376 bp, and bfpA 367 bp) EIEC (ial 320 bp), E. coli O157:H7 (fliC H7 625 bp, and O157, 500 bp) present in diarrhoegenic E. coli; however, although this method is also capable of detecting Shigella spp., it can not distinguish it from EIEC, since the amplification target used by this method is a region of the invasion-associated locus (ial), common to both species. Therefore, the presence of Shigella spp. was also confirmed by using specific antisera.
In order to molecularly serotype Salmonella spp., we used two previously described methods [9,10] coli ATCC 11775 (negative control without virulence genes), Salmonella typhimurium, Salmonella paratyphi B and Salmonella infantis. All primers used were obtained from Sigma Genosys (Sigma). Gel electrophoresis was photodocumented using a Gel Logic 100 Imaging system (KODAK). Molecular sizing of the amplicons was performed using KODAK Molecular Imaging Software.

Parasitology
Determination of E. histolytica/E. dispar and G.intestinalis was done using Faust method. Protozoa were concentrated by centrifugal-floatation (500 × g 2 min) using zinc sulphate as the diluent (specific gravity 1.19) and observed with a light microscope at 40× [14].

Statistical methods
Differences between isolation rates among patient and control groups were evaluated by the χ 2 test.

Stool samples
Sex distribution was similar in both groups: 55% of population were females and 45% were males. Sex distribution did not show to play an important role regarding enteropathogen prevalence. Regarding the patient group, 58% presented abdominal pain, vomit and fever (> 39°C); and 20% required oral rehydration. Ten percent of the stool samples obtained from this group presented blood.
The predominant pattern was comprised by a single enteropathogen infection (48%), followed by co infection involving two pathogens (24%), and being the less common co infection by four pathogens (12%) ( Table 2).
In the control group, 27.5% stool samples were free of any of the bacterial enteropathogens analyzed, 62.5% were positive for one and 10% were positive for two enteropathogens (E. histolytica/E. dispar and G. intestinalis).

Multiplex PCR and detection of Enterobacteriaceae
Only 195 E. coli strains were detected in the stool samples (160/300 from patient group and 35/80 from control group); 46.6 % (140/300) of cases and 56.2% (45/80) of controls stools contain no E. coli. Analysis by multiplex PCR shows that only 101 out of 195 E. coli were positive (97/160, 60.6%) from patients and 4/35 (11.4%) from control group. Eight isolates from patients were further identified as Shigella ( Table 1). The prevalence of diarrhoegenic E. coli in both groups was significantly different (P < 0.001). The frequencies of the diarrhoegenic E. coli positive for one of the targeted genes are shown in Table 3. Regarding Salmonella spp., the most common species in the patient group was S. ohio (28.3%), followed by S. typhimurium (16.3%) and S. infantis (8.0%); compared with S. ohio (2.5%) and S. typhimurium (1.2%) in the control group (Table 1).

Discussion
The current study used a variety of diagnostic methods which helped estimate 100% prevalence of the enteropathogens in stool samples from children presenting diarrhoea symptoms. In addition, a high rate for multiple infections 156/300 (52%) was observed. However, not only the patient group had enteropathogens, also 50/80 (62.5%) members of the control group were positive for either one or two parasites (E. histolytica/E. dispar and Giardia intestinalis). Although only 4 (5%) of the control group were positive for diarrhoegenic E. coli (Table 3), they did not have any symptoms by the time this study was conducted. One of the strains detected in the controls was E. coli O157:H7, which has been recently described as an emerging pathogen worldwide [6,15]. The high rate of enteropathogens detected in this study among both groups, reflects the importance of monitoring on a daily basis the most vulnerable population, such as low economic level children groups.
E. histolytica/E. dispar were the most frequently found enteropathogen in both, the patient (70.3%) and the control group (43.7%) ( Table 1). Globally, amoebiasis is wide spread in approximately 20% of the world's population; 10% of those individuals get sick, and 0.1-0.25% of them die. E. histolytica constitutes the third cause of death for parasitic diseases [16]. On the other hand, G. intestinalis a protozoa which causes symptomatic infections mainly in children under 12 years old was observed in 33% of the patient group, and 20% for the control group (Table 1). These results are in agreement with a previous study that reported G. intestinalis in 29.9% of a population group in Mexico City [17]. An interesting observation is that although 63.7% of the members of the control group were positive for E. histolytica/E. dispar and/or G. intestinalis (Table 1), none of them presented symptoms of disease.
We used three multiplex PCR methods that accurately detected specific genetic virulence markers of Salmonella spp., Shigella spp., and E. coli groups ETEC (eltB and estA gene fragments), EPEC (eaeA and bfpA gene fragments), VTEC (eaeA and vt1or vt2 genes), EIEC/Shigella (ial region), and E. coli O157:H7 (fliC H7 flagellar antigen and a characteristic O157:H7 DNA fragment previously described) [18]. EIEC and Shigella are closely related regarding some of their virulence factors and phenotypic properties [19,20], the multiplex PCR method employed in this study can not be used alone for a definitive discrimination between these two enteropathogens since it recognizes the ial locus, which can be found in both, EIEC and Shigella spp. Therefore, ial positive amplification isolates were considered as positive for EIEC if they showed no agglutination when challenged against specific antiserum for Shigella spp. Also, because eltB and ial amplicons are very close regarding their molecular size (322 and 320 bp respectively) (Fig. 1), two separate PCR reactions were done, each containing specific primers against eltB or ial in order to avoid the overlapping of amplification products, that might lead to false positives. Diarrhoegenic E. coli was present in 32% of the patient group and 5% of the control group (Table 3); these data are consistent with a previous study conducted in Hanoi, Vietnam, where diarrhoegenic E. coli was detected by multiplex PCR, in 22.5 % of the patient group and 12% for the control group [21]. The presence of diarrhoegenic E. coli groups has been reported world wide and recognized as one of the major causes of deceases involving enteropathogens in children [22][23][24][25][26]. In Mexico and other develop-ing countries in the world, ETEC is the most prevalent diarrhoegenic E. coli group [27,28]. It is considered as an important pathogen in children, especially during the first six months, where the isolation rate ranges from 10 to 30% [29]. In this study, the presence of ETEC in the patient group (13.3%) was the highest compared to the other E. coli groups detected (Table 1). However, this rate is lower than what has been reported in some parts of the world, which reaches up to 20.7% [30,31]. It seems that, regardless of the location, the presence of eltB gene (encoding thermo labile toxin) is a definitive advantage for ETEC, since in our study, the virulence gene distribution rate for eltB was high (6.6%). The same observation was reported by another studies conducted in Sweden [13] and Vietnam [21], where eltB was the most commonly found ETEC virulence gene in the studied groups. The second most commonly found diarrhoegenic E. coli strain was EPEC (9.3%), which has been reported responsible for a high rate of mortality and morbidity among children, especially in developing countries where poor sanitary conditions prevail [6]. For example, in countries such as Mexico [32,33] and Brazil [34] up to 40% of diarrhoeal episodes in children are due to EPEC. Interestingly, regarding virulence gene homogeneity distribution, it seems that the situation for EPEC is not the same as for ETEC. A higher percentage of eaeA+/bfpA-isolates (atypical strains) was found compared to eaeA +bfpA+ (typical strains) (Table 3), which differs from others studies conducted in Vietnam, where most of the isolates were eaeA +bfpA+ [21].
VTEC E. coli, an enteropathogen distributed worldwide, that has been more extensively studied in USA [35] and χ 2 represents statistical difference between both groups studied (P < 0.001). Abbreviations are the same as indicated in Table 1. 2.5% Agarose gel electrophoresis of E. coli amplicons obtained from diarrhoeal samples (run conditions 120 V, 94 mA, 45 min.) Europe [36] was also detected in this study. One strain of VTEC was detected in the controls whereas in the patient group it had a prevalence rate of 8.6% (26/300) (  (Table 3). These results differ from those reported by Svenungsson et al., [13] where vt1 was in fact the less commonly found genotype in their samples.
E. coli O157:H7 is an emergent pathogen causing the haemolytic -uremic syndrome, and is considered as an important foodborne source of intestinal infection causing more than 73,000 diarrhoeal episodes in the United States every year [15,37]. We found one E. coli O157:H7 in the control group, however, no symptoms of diarrhoea were reported by the time this study was conducted. There is no explanation for this, but it might be possible that the infection was in an early stage when the stool sample was collected, or that the O157:H7 was a non-producing toxin strain. EIEC was the least detected in the patient group (1%) ( Table 1), which is in agreement with a previous study conducted in a surrounding area of Mexico City where prevalence was 0.85% [28]. It has been estimated that EIEC in developing countries including Latin American is rare [5,38]. The prevalence rate for Shigella sonnei (1.6%) and Shigella flexneri (1%) in the current study was very low (Table 1) when compared other studies [13,21]. However, in other parts of the world, several Shigella outbreaks have been described, and the main feature is their ability to spread in the population, due to its very low infection doses (~10 bacterial cells per host) [19,20]; therefore, a low prevalence of cases of Shigella should not be underestimated.
To detect Salmonella spp., two separate reactions containing primers complementary to all internal variable regions of the fliC and fljB genes, which codifies phase I and phase II flagellar antigens (Fig. 2), were used. Both multiplex PCR systems have been previously tested in hundreds of samples from different parts of the world, and were reported to be accurate, fast and reliable alternative to other traditional diagnostic methods used in the clinical laboratory [9,10]. In our study, both stool sample groups showed S. ohio and S. typhimurium as the predominant serovars; nevertheless, many other serovars were present in both groups of samples in a considerable proportion (Table 1), which suggest that Salmonella diversity is more in comparison to E. coli, for which ETEC was considerably predominant (Table 1). Salmonellosis is one of the biggest challenges to public health all over the world.
In the year 2000, a study conducted in Mexico showed S. typhimurium and S. ohio as the most and less commonly found serotypes of Salmonella respectively [39]. This may be related to the way the outbreak was studied, since we focused on most common serotypes in children (age < 12 years) and not in the whole population. This study is intended to investigate some interesting observations that deserve further studying. The results revealed that all stool samples from the patients and the majority of the control group were positive for at least one enteropathogen and that 52% of stool samples had multiple enteropathogens. All of the 300 stool samples tested except 2 had either E. histolytica/E. dispar or G. intestinalis and when an enterobacteria was present, it was most likely to be a Salmonella serotype. In addition, only 12 out of 300 stool samples tested E. coli presence was in association with a protozoa in the absence of any type of Salmonella spp. (Table 2).
(A) 2.5% Agarose gel electrophoresis of Salmonella phase I flagellar amplicons obtained from diarrhoeal samples It has been reported that S. typhimurium virulence is enhanced when exposed to the rumen protozoa [40]; this hypervirulent phenotype is linked to the presence of the Salmonella genomic island SGI1 integron [41]. Also, it has been shown that SGI1 is widely spread in other S. enterica serovars; and although rumen possesses different protozoa diversity than the parasites studied here, some Salmonella isolates from human infections have shown to posses such island [42]. It is tempting to speculate that a similar virulence enhancement process might occur in human digestive trait; possibly through E. histolytica/E. dispar or G. intestinalis. Whether those Salmonella strains studied in this work possesses or expresses such genomic island needs further study.

Conclusion
The study of enteropathogen associations may lead to a better understanding of the etiology of diarrhoea, and therefore in prescribing more suitable treatments in cases of outbreaks. This study revealed that 52% of the patients were infected by more than one enteropathogen, notably E. histolitica/E. dispar and Salmonella ohio. These results are useful for clinicians in improving the empiric treatment used in such cases.