Diarrheagenic Escherichia coli (DEC) are the most common bacterial cause of diarrhea in both developing and industrialized regions, primarily affecting children under 5 years of age (Gomes et al., 2016). According to the detection of the virulence factors, six classical categories or pathotypes have been described (Kaper et al., 2004). DEC pathogenesis comprises three stages: adherence and colonization, production of toxins, and diarrhea followed by inflammation (Croxen et al., 2013).

Shiga toxin-producing Escherichia coli (STEC) initial attachment to intestinal cells is mediated by adhesins, such as the long polar fimbriae (Lpf) (Torres et al., 2002). Intimate attachment is defined by the ability of STEC to induce characteristic attaching and effacing lesions on intestinal epithelia, which is mediated by proteins encoded by genes on the pathogenicity island called locus of enterocyte effacement (LEE) (Stevens and Frankel, 2014). Enteroaggregative Escherichia coli (EAEC) pathogenesis is regulated by a plasmid-encoded virulence regulator called AggR. This master regulator controls the expression of several chromosome- and plasmid-encoded genes, including aggregative adherence fimbriae (AAFs), a key virulence factor in EAEC pathogenesis responsible for intestinal cell adherence, and dispersin, with its dedicated type I secretion system (Harrington et al., 2006; Morin et al., 2013). Another important virulence factor in EAEC is plasmid-encoded toxin (Pet), a protease that induces cytotoxic and cytopathic effects on different cultured cell lines by spectrin and fodrin degradation (Villaseca et al., 2000).

STEC and EAEC infections are characterized by acute inflammation of the colonic mucosa, where these DEC pathotypes induce the secretion of several pro-inflammatory markers, such as interleukin-8 (IL-8), as a response to the activation of the MAPK, AP-1, and NF-kB signaling pathways (Steiner et al., 1998; Dahan et al., 2002; Miyamoto et al., 2006). The mechanism of pathogenesis for DEC pathotypes and other enteropathogens involves a variety of interactions between bacterial and host factors (Kitamoto et al., 2016). Under well-defined environmental conditions, expression of virulence genes at the infection site occurs, allowing the bacteria to adhere and colonize the target cells (Barnett Foster, 2013). For STEC and EAEC, virulence gene expression is not completely understood, and most studies have focused on deciphering the molecular mechanism occurring inside the bacteria, while very little is known about the environmental factors that regulate pathogenesis at a specific time or place (Jubelin et al., 2018). Considering that the expression of virulence factors in bacterial pathogens is strongly regulated by environmental conditions at the infection site, it is not surprising that gut microbiota might play a role in the regulation of pathogenic mechanisms (Baumler and Sperandio, 2016; Jubelin et al., 2018). Evidence supporting the role of specific strains from the normal gut microbiota in STEC virulence regulation has been published (Kendall et al., 2012; Pacheco and Sperandio, 2012; Curtis et al., 2014; Pacheco and Sperandio, 2015). In a previous publication, we identified two indicative bacterial species of gut microbiota significantly associated with DEC-positive diarrhea stool samples from children under 5 years of age, Citrobacter werkmanii (C. werkmanii) and Escherichia albertii (E. albertii) (Gallardo et al., 2017). In this work, we evaluated the role of these indicative species in the pathogenicity of STEC and EAEC.

DEC pathotypes have evolved from commensal E. coli, acquiring virulence traits and sensing mechanisms to express virulence factors at the right time and place for the successful colonization of intestinal cells (Kitamoto et al., 2016). Among the environmental conditions, gut microbiota composition might play a role in the regulation of pathogenic mechanisms (Tsolis and Baumler, 2020). In this study, we provided experimental evidence that indicative species of gut microbiota associated with DEC infection might regulate the expression of virulence genes in STEC and EAEC.

It is well known that changes in microbiota composition occur in diarrhea (Pop et al., 2014; Samb-Ba et al., 2014; Schubert et al., 2014; Pop et al., 2016), and this could have an impact on the environmental conditions present at the infection sites (Vogt et al., 2015; Gallardo et al., 2020). In a previous study, we identified indicative bacterial species of gut microbiota associated with DEC-positive diarrhea stool samples from children under 5 years of age (Gallardo et al., 2017). In this work, we evaluated the role of two of these indicative species, Citrobacter werkmanii (CW) and Escherichia albertii (EA), in the infection process of the DEC pathotypes STEC and EAEC.

Considering that pathogenic bacteria can use microbiota-derived sources of carbon and nitrogen as nutrient and regulatory signals to promote their own growth and virulence (Baumler and Sperandio, 2016), our first approach was to evaluate the effect of the supernatant of C. werkmanii and E. albertii on the gene expression of virulence factor of STEC and EAEC by RNA-seq experiments. In STEC, an increase was observed in the expression of the gene coding for the master regulator of LEE pathogenicity island (ler gene) and virulence genes regulated by Ler (Table 1), a situation that could be indicative of a positive regulation of STEC virulence in vitro (Honda et al., 2009; Garcia et al., 2012). A similar situation was previously reported, where, after incubating STEC with B. thetaiotaomicron or E. faecalis, both bacteria of gut microbiota present in healthy people, there was an increase in ler gene expression and several virulence genes in STEC (Curtis et al., 2014; Iversen et al., 2015). Along with this information, we observed an increase in the genes coding for different subunits of Lpf fimbriae (genes lpfA, lpfB, lpfD, and lpfE), a relevant virulence factor for STEC adherence to intestinal cells (Torres et al., 2002). For EAEC, we found a downregulation of the gene coding for the protein FliC, belonging to flagellin, which can induce IL-8 secretion in intestinal cells by interacting with the cellular receptor TLR5 (Khan et al., 2004). The regulation of the fliC gene expression by metabolites has been previously reported. E. coli incubation with short-chain fatty acids (SCFAs) at 20 mM produced an increase in FliC protein levels along with a higher bacterial motility (Tobe et al., 2011); however, colonic levels of SCFA (65 mM) downregulated genes associated with motility (fliC) in 11 out of 15 E. coli strains tested (Zhang et al., 2020). Furthermore, we found an increase in the expression of the gene coding for Pet toxin, an autotransporter-secreted protein with serine protease activity that has been described as having enterotoxic and cytotoxic effects (Navarro-Garcia et al., 2010). It has also been reported that Pet can induce proinflammatory cytokine secretion as IL-8 and TNF-α in macrophages (Rocha-Ramirez et al., 2016). Together, these results suggest that secreted components by indicative bacteria from intestinal microbiota of DEC-positive diarrhea can regulate the expression of virulence factors in STEC and EAEC and suggest a higher virulence of these DEC pathotypes in the presence of EA or CW SP.

Previous studies have shown that the microbiota shapes the chemical environment of the gut, producing metabolites that have an effect on the host and other microbes (reviewed in Vogt et al., 2015). It has been reported that spent culture medium from human fecal microbiota inhibits virulence gene expression in enteropathogens; for example, in STEC, the spent culture medium repressed the expression of the stx2 gene (de Sablet et al., 2009). Also, isolated Peptostreptococcus sp. and Lactobacillus sp. from human fecal samples secreted unidentified metabolites that inhibited Vibrio cholerae growth (Silva et al., 2001). Microbiota could also have effects on the host, such as modulating immune functions or altering the expression of host receptors for bacterial toxins (Vogt et al., 2015). For example, SCFAs (generated by fermentation of dietary fibers by intestinal commensal microbiota) can elicit the production of cytokines, including TNF-a and IL-6, and chemokines, such as CXCL1 and CXCL2, by colonic epithelial cells, which induce a protective immune response against pathogens by promoting the recruitment of neutrophils and the development of Th17 cells (Kayama and Takeda, 2020). When we studied the effect of EA and CW SP on the ability to induce IL-8 secretion by epithelial intestinal cells, we found that CW SP can induce IL-8 secretion without the presence of DEC pathotypes; no IL-8 secretion was observed when EA SP was added to T84 cells (Figure 2). A previous study using supernatants from four potentially probiotic E. coli strains showed that one of them can induce IL-8 secretion on intestinal cells in culture, and this could be attributable to differences in bacterial surface molecules (Zimmermann et al., 2018). Cell extract from E. coli Nissle can also induce IL-8 production in a dose-dependent way, suggesting that secretory proteins, virulent or not, may play a role in IL-8 production (Lammers et al., 2002). This evidence suggests that some bacterial metabolites present in CW SP can activate this signaling pathway; however, the mechanism involved in this process has not been discovered yet. For this reason, we decided to evaluate the effect of EA SP and not CW SP on the ability of STEC and EAEC to induce IL-8 secretion on intestinal epithelial cells.

We found that incubation of EA SP with EAEC or STEC induced a higher secretion of IL-8 compared to the control condition (Figure 3). Interestingly, this increase in IL-8 secretion can also be observed with the supernatant obtained from five clinical strains of E. albertii (Figure 4), which suggests that the effect is associated with E. albertii species and not with a particular strain. The secretion of IL-8 by intestinal cells creates a chemotactic gradient that promotes the migration of neutrophils to the infection site (Godaly et al., 1997). It has been reported that the presence of neutrophils at infection sites can favor EAEC colonization, producing an increase in adherent bacteria to cells in culture and promoting pathogenesis (Boll et al., 2012). It has also been reported that pathogenic bacteria, by eliciting inflammation, can alter the intestinal environment and create a fresh niche of respiratory nutrient suitable for their expansion in the gut (Baumler and Sperandio, 2016). Considering that different supernatants from E. albertii cannot by themselves induce an inflammatory response, the effect observed could be associated with a higher expression of virulence factors in DEC pathotypes that are capable of causing an inflammatory response in intestinal cells.

Adherence to intestinal cells is a key step in DEC pathogenesis, a process mediated by bacterial adhesins. For STEC and EAEC, several adhesins have been described; some with the ability to participate in other steps of pathogenesis, such as the induction of an inflammatory response in infected cells or tissues (Harrington et al., 2005; Boll et al., 2012; Farfan et al., 2013). Of the genes coding for proteins involved in adherence, RNA-seq data revealed that 35% and 8.5% of these genes were upregulated for STEC and EAEC, respectively (Figures 1A, B). Surprisingly, we did not observe any differences in the adherence of STEC and EAEC to T84 cells in the presence of EA SP (Figure 6). These observations can be explained by the ability of STEC and EAEC to adhere to intestinal cells in large numbers (approximately 5 × 10⁵ CFU/well). The upregulation in the expression of genes found in RNA-seq experiments was likely insufficient to significantly increase the adherence to cells. Further experiments to correlate the RNA-seq data and the adherence of STEC and EAEC observed, such as protein levels of adhesins or their expression on the bacterial surface, are needed.

Previous experiments from our group have reported that Lpf is also one of the factors that mediate adherence of STEC to intestinal cells and the inflammatory response through NF-kB pathway activation, caused by this DEC pathotype (Torres et al., 2002; Farfan et al., 2013; Vergara et al., 2014). To evaluate the role of Lpf in the induction of IL-8 secretion by STEC incubated with EA SP, we used a double mutant in the lpfA1 and lpfA2 genes (strain AGT210), which adhered similarly to the wild-type strain STEC 86-24, but induced a reduced amount of IL-8 secretion by infected T84 cells (Farfan et al., 2013). We found that strain AGT210 did not present an increase in IL-8 secretion in the presence of EA SP (Figure 7), suggesting that Lpf fimbriae might be participating in the increase of IL-8 observed in the presence of EA SP. As we discussed above, we did not note a higher STEC adhesion to intestinal cells (Figure 6). Receptors involved in the recognition of Lpf fimbriae and responsible for the activation of an inflammatory response have not been discovered yet. Previous studies have explored possible receptors for these fimbriae, and found that Lpf is capable of binding to extracellular matrix proteins such as fibronectin, laminin, and collagen IV (Farfan et al., 2011), but its participation in the activation of an inflammatory response by these bacteria has not been explored. More studies are needed to correlate the induction gene expression of Lpf fimbriae coding genes, the expression of these fimbriae at the surface of bacteria, and the effect on adherence to cells as well as the inflammatory response induced by this adhesin. Pet toxin is a key virulence determinant in EAEC pathogenesis (Villaseca et al., 2000). In this study, we found that EA SP induced a higher secretion of Pet toxin by EAEC than the control conditions (Figure 8). Considering this toxin has been associated with inducing the secretion of inflammatory cytokines (Rocha-Ramirez et al., 2016), we suggest that Pet could be participating in the increase in IL-8 observed. This increase in Pet toxin secretion was even higher in the presence of the supernatant obtained from clinical EA strain (E. albertii 0621). A similar situation was noted in a previous study with the metabolite succinate, originally detected in the supernatant from B. tethaiotaomicron, which is responsible for the increase in virulence factor EspA in STEC (Curtis et al., 2014). Further studies evaluating the role of Pet toxin in EAEC infection to cells in the presence of EA SP are needed.

In this work, our results have shown that the supernatant from E. albertii, a bacterium from intestinal microbiota associated with DEC infections, can induce changes in gene expression for two DEC pathotypes, including regulators of virulence, toxins, and fimbriae. These changes had an effect on the inflammatory response caused by these bacteria, causing an increase in IL-8 levels secreted by intestinal cells after the infection with STEC or EAEC. Together, our data suggest that metabolites secreted by E. albertii modulate the virulence of STEC and EAEC, producing changes in gene expression and affecting the inflammatory process caused by DEC pathotypes. Further experiments to identify the metabolites in the EA SP might be useful to enhance our knowledge of STEC and EAEC pathogenicity, and of the mechanism and signaling molecules involved in the crosstalk between the microbiota and enteropathogens.

The RNA-seq data discussed in this publication have been deposited in NCBI's Gene Expression Omnibus (Edgar et al., 2002) and are accessible through GEO Series accession number GSE197797, https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE197797).

MI performed the experiments, data analysis, interpreted the data and participated in manuscript writing. JL performed the experiments, data analysis. RV participated in study design and data analysis. PG participated in data analysis. JO participated in data analysis. MF participated in study design, interpretation of data, manuscript writing and final approval of the manuscript. All authors contributed to the article and approved the submitted version.

This work was supported by an ANID grant 1200994 to MF. MI and PG were supported by a doctoral fellowship from ANID. We thank Dr. Tânia A. T. Gomes (Federal University of Sao Paulo, Brazil) for the E. albertii clinical strains and Dr. Fernando Navarro-Garcia (CINVESTAV, Mexico) for help and guidance with the Pet secretion experiments.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.