Escherichia coli from Commercial Broiler and Backyard Chickens Share Sequence Types,Antimicrobial Resistance Profiles,and Resistance Genes with Human Extraintestinal Pathogenic Escherichia coli

Introduction

Extraintestinal pathogenic Escherichia coli (ExPEC) are a subset of E. coli—distinct from nonpathogenic commensal E. coli and intestinal pathogenic E. coli—that cause extraintestinal infections and represent a serious public health concern (Russo and Johnson, 2000, 2003). In humans, ExPEC are responsible for most cases of urinary tract infection (UTI), bloodstream infection (BSI), and other community-associated and health care-associated infections (Russo and Johnson, 2000; Weiner et al., 2016). Similarly, E. coli that cause infections in poultry are termed avian pathogenic E. coli (APEC); APEC can cause colibacillosis, a disease of significant economic impact in farmed poultry (Dziva and Stevens, 2008).

While ExPEC are genetically diverse, a small number of important clonal groups—sequence type 131 (ST131), ST69, ST95, ST73, ST127, and ST10—are responsible for a large proportion of human extraintestinal E. coli infections worldwide (Izdebski et al., 2013; Maluta et al., 2014; Riley, 2014). Some of these strains produce extended-spectrum β-lactamases (ESBLs), which mediate resistance to extended-spectrum cephalosporins, and are often also resistant to other antimicrobial agents, including aminoglycosides, fluoroquinolones, and trimethoprim-sulfamethoxazole (Coque et al., 2008; Brisse et al., 2012; Adams-Sapper et al., 2013; Kim et al., 2018). These ESBL-producing and multidrug-resistant isolates are often associated with treatment failure.

Some ExPEC isolated from humans and E. coli isolated from poultry meat demonstrate similar phenotypic and genotypic characteristics, such as antimicrobial resistance patterns, antimicrobial resistance genes, virulence genes, serotype, and sequence types. Therefore, it has been hypothesized that food animals, especially poultry, are a potential source of human ExPEC (Bergeron et al., 2012; Mora et al., 2013; Maluta et al., 2014; Yamaji et al., 2018). This zoonotic potential could represent a serious public health problem (Giufrè et al., 2012; Egervärn et al., 2014; Casella et al., 2015). Brazil is a large producer and exporter of poultry, particularly chicken meat (https://apps.fas.usda.gov/psdonline/circulars/livestock_poultry.pdf). Yet, studies comparing E. coli isolated from human clinical samples and poultry from Brazil are scarce (Maluta et al., 2014). The purpose of this study was to determine if poultry E. coli and human ExPEC from Brazil demonstrate phenotypic and genetic similarities, to evaluate whether Brazilian poultry represent a potential source of E. coli that can cause extraintestinal diseases in humans.

Materials and Methods

Results

Distribution of MLST genotypes among human and poultry E. coli isolates

We analyzed 149 human ExPEC isolates and 111 poultry E. coli isolates. The human isolates were classified into 55 distinct STs, whereas the poultry isolates were classified into 30 STs (Fig. 1 and Supplementary Table S1). Of these isolates, 10 ExPEC and 28 poultry E. coli could not be assigned to a known ST. The most common STs identified in human ExPEC were ST69, ST131, ST789, and ST73, accounting for 42 (28%) of the human E. coli isolates. The most common STs identified in poultry E. coli were CC155 (ST58, ST155) and ST117, accounting for 33 (30%) of the E. coli isolates from chickens. The Simpson's diversity index for poultry ST genotypes was 0.91, whereas the diversity index for human ST genotypes was 0.97.

OPEN IN VIEWER

FIG. 1.

Distribution of multilocus STs of Escherichia coli isolates obtained in Brazil from (A) patients with urinary tract infection or bloodstream infection, and from (B) commercial broiler or backyard chickens. Red bars represent shared STs and gray bars represent unshared STs. ST, sequence type. Color images are available online.

Overall, 75 different STs were detected among 222 E. coli isolates. Twenty STs were found exclusively among poultry isolates, while 45 STs were found only among human isolates. Ten STs were shared between human and poultry isolates (ST10, ST48, ST58, ST88, ST90, ST93, ST131, ST602, ST617, and ST1018). Thirty-five (32%) poultry E. coli and 35 (23%) human ExPEC isolates belonged to the 10 shared genotypes (Table 1).

OPEN IN VIEWER

Table 1.

Distribution of Escherichia coli Isolates Among the 10 Shared Genotypes Detected in Human Clinical Samples and Poultry Samples

ST/CC	No. (%) of E. coli isolates
	Human clinical isolates			Poultry isolates
	Hospital-acquired BSI/UTI, n = 68	Community-acquired UTI, n = 81	All, n = 149	Broiler chickens, n = 60	Backyard chickens, n = 51	All, n = 111
131/131	6	5	11 (7.4)	2	0	2 (1.8)
10/10	3	3	6 (4.0)	1	2	3 (2.7)
93/168	2	0	2 (1.3)	0	3	3 (2.7)
58/155	2	3	5 (3.4)	5	9	14 (12.6)
48/10	1	0	1 (0.7)	0	2	2 (1.8)
1018/58	0	1	1 (0.7)	3	0	3 (2.7)
90/23	1	2	3 (2.0)	0	1	1 (0.9)
602/446	2	0	2 (1.3)	3	0	3 (2.7)
617/10	0	2	2 (1.3)	1	0	1 (0.9)
88/23	2	0	2 (1.3)	2	1	3 (2.7)
Total	19 (28)	16 (20)	35 (23)	17 (28)	18 (35)	35 (32)

BSI, bloodstream infection; CC, clonal complex; ST, sequence type; UTI, urinary tract infection.

Phylogenetic relationship between human and poultry E. coli isolates

The E. coli STs were compared phylogenetically based on a minimal spanning tree using eBURST (goeBURST). Figure 2 shows the distribution of STs found in human clinical isolates (community-acquired or hospital-acquired) and in poultry (commercial broiler chickens or backyard chickens). STs from human clinical isolates are distributed across the entire spectrum of the goeBURST diagram. Putative founder STs in both human ExPEC and poultry E. coli belonged to CC10, CC23, CC131, CC155, CC168, CC446, and CC648. Putative founder STs that contained human isolates only belonged to CC86, CC95, and CC141, whereas founder STs that contained avian isolates only belonged to CC350.

OPEN IN VIEWER

FIG. 2.

A population snapshot of 222 Escherichia coli isolates, diagrammed based on a minimal spanning tree using optimized eBURST (goeBURST), based on PHYLOViZ software (www.phyloviz.net). Genetic relationships are shown among E. coli isolates obtained in Brazil from community-acquired (dark red) or hospital-acquired (light red) patients, and from commercial broiler chickens (dark green) or backyard chickens (light green). Each ST is represented by a circle; the size of the circle is proportional to the number of isolates. Representative STs are shown as numbers without the “ST” prefix; numbers in squares represent the shared genotypes between human ExPEC and poultry E. coli. The links between circles depict the number of allelic mismatches between STs as follows: black lines, one mismatch; dark gray lines, two mismatches; and light gray lines, three or more mismatches. The CC is represented in blue font when more than two STs belong to one CC. CC, clonal complex; ExPEC, extraintestinal pathogenic E. coli; ST, sequence type. Color images are available online.

Discussion

It has been suggested that poultry are a potential reservoir for E. coli that cause extraintestinal disease in humans, since shared genotypic and phenotypic characteristics have been found among human ExPEC and E. coli isolated from poultry (Vincent et al., 2010; Overdevest et al., 2011; Bergeron et al., 2012; Mora et al., 2013). Here, we analyzed the genotype and antimicrobial susceptibility profile of E. coli isolated from human infections and poultry sources that were collected in Southeast Brazil, to assess the potential and degree to which poultry may serve as a reservoir for ExPEC.

We found that 10 genotypes were shared among human ExPEC and poultry E. coli. Of the 10 shared genotypes, ST131 and ST10, recognized worldwide as pandemic lineages of ExPEC (Adams-Sapper et al., 2013; Riley, 2014), were the most common STs among human UTI and BSI E. coli isolates. This result is consistent with studies from Brazil and the United States, which compared genotypes of E. coli isolated from human UTI and/or BSI and avian sources (Maluta et al., 2014; Yamaji et al., 2018). Other shared genotypes—such as ST48, ST58, ST93, ST602, and CC23—while not considered pandemic lineages, are of interest because members of these STs have previously been detected in poultry and implicated in extraintestinal human diseases worldwide (Oteo et al., 2010; Bergeron et al., 2012; Maluta et al., 2014; Manges et al., 2015; Day et al., 2016; van Hoek et al., 2016; Liu et al., 2017). Some of these isolates shared more genetic characteristics than merely ST lineage membership, such as resistance genes.

ST69, ST73, and ST95—other important ExPEC pandemic lineages—were among the most common STs detected in human clinical E. coli in this study, accounting for 17% of all urine and blood isolates. However, we did not find these STs in poultry, which is in accordance with other reports (Maluta et al., 2014; Manges et al., 2015). The lack of detection of these STs in poultry isolates in this work may be due to the limited number of samples tested, as other studies from Brazil, the United States, and Europe have reported these STs in food animal sources, including poultry (Nandanwar et al., 2014; Day et al., 2016; Cunha et al., 2017; Yamaji et al., 2018). Of the 20 STs that were unique to poultry in this study, 15 (ST23, ST40, ST57, ST117, ST155, ST297, ST359, ST371, ST428, ST533, ST624, ST683, ST695, ST1196, and ST5005) have been reported previously from human ExPEC isolates deposited in the MLST database (https://enterobase.warwick.ac.uk/species/index/ecoli). Thus, there appears to be additional genotype overlap between poultry strains characterized in this study and ExPEC that have been found to cause disease in humans. These reported differences in genotype distribution of shared poultry and human E. coli isolates may be related to differences in sampling design, sample size, as well as geographic and temporal study framework.

Identical antimicrobial resistance profiles were more commonly shared by different STs across human ExPEC and poultry E. coli. However, among the 10 shared genotypes, ST58 and ST88 isolates shared identical antimicrobial resistance profiles, and ST58 isolates also shared the bla _TEM-1 β-lactamase gene. Interestingly, commercial broiler chicken isolates were significantly more likely to be MDR than human clinical or backyard chicken isolates. The fact that the poultry industry uses antimicrobial agents in animal feed to maintain health and productivity, which presumably selects for antimicrobial-resistant bacteria (Van Boeckel et al., 2015), may explain the higher level of antimicrobial resistance in commercial broiler chicken isolates compared with backyard chicken isolates. However, one study from the United States found that human UTI isolates were more likely to be antimicrobial resistant than retail meat isolates, which included poultry (Yamaji et al., 2018). The contrasting results from our study compared with the U.S. study may be due to differences in the source of poultry isolates (cloaca vs. meat) or sanitary management at the farms of origin.

The bla _TEM-1 gene was prevalent in β-lactam-resistant isolates from both humans and chickens. In contrast, many of the ESBL/pAmpC gene variants identified in human and poultry isolates were different; the bla _CTX-M-15, bla _CTX-M-27, and bla _CMY-2 genes were found only among human ExPEC. Interestingly, bla _SHV-2 was the most common ESBL gene detected in poultry isolates. To our knowledge, this is the first description of the bla _SHV-2 gene in E. coli from poultry in Brazil (Egervärn et al., 2014; Ferreira et al., 2014; Casella et al., 2015). While groups from Canada, Europe, and Japan have detected bla _SHV-2 in E. coli isolated from avian and human samples, it is not a commonly reported β-lactamase gene in E. coli (Blanc et al., 2006; Asai et al., 2011; Leverstein-van Hall et al., 2011; Peirano et al., 2017).

bla _CTX-M-2, bla _CTX-M-8, and bla _SHV-2 were the ESBL genes in common among human clinical and poultry isolates, and most of the strains containing these genes were from distinct STs and CCs. Taken together, these results suggest that (1) some ESBL genes may be transferred from poultry E. coli to ExPEC, since all the β-lactamase genes detected in poultry isolates were observed among human isolates but not the contrary; while it is also likely that (2) ESBL genes in human ExPEC and poultry E. coli are acquired from different bacterial sources, perhaps due to existence in different intestinal bacterial communities within each host (Sommer et al., 2009), or (3) humans and poultry may acquire ESBL-producing E. coli from different sources due to other exposures, such as vegetables, water sources, or wildlife (Alcalá et al., 2016; Dropa et al., 2016; Müller et al., 2016; Gomi et al., 2017; Batalha de Jesus et al., 2019).

Most of the human ESBL-producing strains belonged to the ST10 CC (excluding ST10 itself). Since ST131 is one of the most common CTX-M-producing lineages (Brisse et al., 2012; Day et al., 2016; Merino et al., 2016), it was surprising that only three (27%) of 11 ST131 isolates in this study carried ESBL genes. Notably, only two ST131 E. coli were isolated from poultry; one was pan-susceptible and the other was resistant to gentamicin. This is in contrast to the 11 ST131 E. coli isolated from humans, which were each resistant to two or more antimicrobial agents. This again suggests that either poultry and human ST131 E. coli acquire resistance genes in the gastrointestinal tract from different types of bacteria (Sommer et al., 2009), or that human and poultry ST131 E. coli have different reservoirs, and represent phylogenetically distinct populations. Our second hypothesis is supported by a study which reported that ST131 strains isolated from human clinical samples and food animals did not show similarity by pulsed-field gel electrophoresis (Johnson et al., 2012).

In conclusion, while human and poultry E. coli isolates showed genetic diversity, some genotypic and phenotypic characteristics overlapped, reinforcing the zoonotic potential of avian E. coli. Since the ExPEC isolates were not necessarily obtained from people residing close to the chicken production sites, we cannot conclude that the ExPECs we found in this study originated from poultry. Additional research is needed to understand the potential direction and pathways of transmission of ExPEC. A better understanding of ExPEC reservoirs could lead to the identification of effective control interventions to disrupt transmission of these bacteria and their resistance genes.