Background
The colonization of the gut by microbes in mammals occurs during and after birth. This process is dynamic and influenced by various factors such as lifestyle, diet, host genotype, antibiotic use, and diseases [
1]. These interactions eventually lead to the establishment of diverse bacterial populations that form a symbiotic relationship with the host [
2]. The term “dysbiosis” was coined over a century ago to describe the disruption of this symbiosis [
3]. Dysbiosis involves the imbalance of beneficial microbial input or signals and the expansion of pathogenic microbes, known as pathobionts. It is believed that dysbiosis can trigger pro-inflammatory responses and immune dysregulation, which are associated with various disease states [
4], including inflammatory bowel disease (IBD) [
5].
Inflammatory bowel disease (IBD) is characterized by chronic and recurrent inflammation of the intestines and includes two main subtypes: ulcerative colitis (UC) and Crohn's disease (CD). UC primarily involves inflammation of the colonic mucosa throughout the colon, while CD is characterized by transmural ulceration that can occur in any part of the gastrointestinal tract, with the terminal ileum and colon being most commonly affected [
6].
The pathogenesis of IBD is complex and not fully understood. Current evidence suggests that dysregulation of immune responses to intestinal flora, as well as interactions between genetic and environmental factors, play a significant role in driving the disease [
6,
7]. However, the exact underlying pathogenic mechanisms that lead to IBD development and progression are still not completely elucidated [
6,
7]. Further research is necessary to gain a comprehensive understanding of the factors involved in the pathogenesis of IBD.
Formyl peptide receptors (FPRs in humans and Fprs in mice) are members of the G-protein-coupled chemoattractant receptors (GPCRs) family [
8]. They play a role in recognizing natural and synthetic ligands and mediating the accumulation of myeloid cells at sites of infection and inflammation [
9,
10]. In humans, there are three functional FPRs: FPR1, FPR2, and FPR3 [
11,
12], while in mice, there are at least two counterparts: Fpr1 and Fpr2 [
13]. FPR2 exhibits a diverse expression pattern and functionality, as it can recognize a wide range of formylated or non-formylated chemotactic agonists derived from pathogens, host cells, synthetic peptides, and small molecules [
14‐
16]. FPR2 has been implicated in various human diseases, including infections, inflammation, and cancer [
17,
18].
The use of
Fpr2−/− mice has provided valuable models for studying human diseases, such as allergic airway inflammation [
19], lung carcinoma [
20], and colon inflammation and cancer [
21]. By examining the changes in the colon microbiota of
Fpr2−/− mice, we can gain a deeper understanding of the mechanisms underlying human diseases and develop improved therapeutic strategies. This approach allows for a more comprehensive exploration of the relationship between Fpr2 and the gut microbiota, shedding light on the pathogenesis of various conditions and facilitating the development of targeted treatments.
Mouse Fpr2 exhibits a significant role in maintaining colon homeostasis by being expressed on colon crypt epithelial cells and promoting cell proliferation [
8,
21]. This response is triggered by the chemotactic agonist fMLF, which is released by enteric bacteria like
Lactobacillus rhamnosus GG (LGG) and
E. coli [
21,
22]. In healthy individuals,
E. coli represents a small fraction, less than 1%, of the gut microbiota [
23]. However, in individuals with inflammatory bowel disease (IBD) and in animal models of gut inflammation,
E. coli becomes more dominant in the gut microbiota [
24‐
26].
E. coli strains isolated from individuals with IBD often exhibit adherence and invasive properties, displaying virulence characteristics [
27].
In the current study, we identified two serotypes of E. coli and established the connection between the commensal E. coli serotype O22:H8 and Fpr2. This finding suggests a specific relationship between commensal E. coli strain and the Fpr2 receptor, potentially implicating their involvement in the modulation of colon homeostasis and the pathogenesis of intestinal diseases.
Methods
Mice
Specific-pathogen-free (SPF) mice:
Fpr2−/− mice were generated as described previously [
19]. Cre-loxp strategy [
28] was used to deplete mouse Fpr2 gene.
Fpr2−/− mice were backcrossed for at least eight generations to wild type (WT) C57BL/6 mice before using experiments. WT (
Fpr2+/+) and
Fpr2−/− litter mates were generated by mating pairs of male and female heterozygous (
Fpr2+/−) mice. All mice were maintained under SPF environment in the facility of Frederick National Laboratory for Cancer Research (Frederick, MD). Mice were allowed for free access to standard laboratory chow/tap water. All animals were housed in an air-conditioned room with controlled temperature (22 ± 1 °C), humidity (65–70%), and day/night cycle (12 h light, 12 h dark). The mice used in the experiments were male and 2–3 months old (except in the experiments to detect crypt length of colons at different mouse ages).
Germ-free (GF) mice: The embryos were obtained from pregnant female C57BL/6 wild type (WT) mice under sterile conditions then transplanted into a germ-free mother mouse to obtain GF WT mice. GF mice were maintained in a sterile environment completely devoid of microorganisms in Gnotobiotic Facility, Frederick National Laboratory for Cancer Research, Frederick, MD. The GF mice used in the experiments were 2–3 months old, males and all experiments were performed under sterile conditions.
Cell culture
CT26 mouse colon carcinoma cell line was maintained in Dulbecco's modified Eagle's medium (DMEM) (Gibco-Invitrogen) containing 10% FBS (HyClone Laboratories, Logan, UT, USA) and 1% penicillin/streptomycin. CT26 cells were cultured in a humidified 37˚C incubator with 5% CO2.
Design of animal experiments
1.
For measuring the effect of Fpr2 deficiency on colon crypt length, male WT and Fpr2−/− mice (4–9 mice per group) were euthanized at ages of 1, 15, 30, and 90 days. The colons were harvested and used to examine the crypt length. The colon sections of WT and Fpr2−/− mice at age of 60 days were further used to examine the levels of Muc2, cathelin-related antimicrobial peptide (CRAMP), β-Defensin, PAS+ goblet cells and Ki67.
2.
Evaluation of the composition of gut microbiome in naïve mice. Male WT and Fpr2−/− mice were co-housed (2–3 WT mice plus 2–3 Fpr2−/− mice/cage) for 4 weeks after weaning, followed by separation (4–5 WT mice/cage or 4–5 Fpr2−/− mice/cage) for an additional 6 weeks. Subsequently, each mouse was individually housed in a separate cage to collect fresh fecal pellets. The fecal pellets were collected in tubes, with one tube assigned to each mouse. These fecal samples were collected for the purpose of analyzing the microbiota through 16S rRNA gene sequencing.
For analysis of fecal microbiota by 16S rRNA gene sequencing, fecal DNA was prepared using the DNA Stool Mini Kit (QIAGEN) for sequencing. Briefly, the V4 fragment of 16S rDNA was amplified by PCR using primers 515F: 5′-GTGCCAGCMGCCGCGGTAA-3′ and 806R: 5′-GGACTACHVGGGTWTCTAAT-3′ flanked by p5 and p7 Illumina Sequencing adaptors (p5 and p7), barcodes (i5 and i7), pad (to optimize melting temperature), and a link sequence. PCR products were purified and normalized using a SequalPrep Normalization Plate Kit (Invitrogen). Sequencing was performed using Mothur v.1.30.0. as described in the MiSeq 16S standard operating procedure protocol. All sequencing was performed using the National Institutes of Health Biowulf Cluster. The bioinformatics analysis of 16S rRNA gene sequencing data of microbiota was conducted using software USEARCH version 9.2.64 and QIIME version 1.9.1. Paired-end reads were merged, and stringent quality filtering was performed to remove low-quality reads. The remaining reads were de-replicated and clustered into operational taxonomic units (OTUs) with 97% sequence identity using the UPARSE algorithm. OTU assignment and creation of an OTU table were done using the usearch_global command. Taxonomy was assigned using BLASTn searches against the SILVA ribosomal RNA gene database. OTUs with fewer than 10 sequences and no BLASTn hit were removed as quality control. The OTU tables were further processed by rarefying to the sample with the lowest number of sequences, with a threshold of > 10,000 sequences. Statistical analysis was performed using ANOVA and
P values were corrected for multiple comparisons using the q-value test (0.1) [
29]. The results of fecal bacterial population were represented as a Heat-map.
3.
For DSS-induced colitis, male WT and Fpr2−/− mice (10–12 mice per group) were administered 3% dextran sulfate sodium (DSS) (M.W. = 36,000–50,000, MP Biomedicals, LLC) in their drinking water for a duration of 5 days. Following this, they were provided with normal water for a period of 7 days to observe mouse survival. Furthermore, additional groups of WT and Fpr2−/− mice was administered 5% DSS for 5 days, after which they were euthanized. The colons from these mice were harvested and sectioned for the assessment of colon mucosal damage.
4.
To evaluate the recovery of colon epithelial cells, a group of male WT mice and a group of Fpr2−/− mice (6 mice per group) were subjected to a treatment with 5% DSS for a duration of 3 days, followed by a period of normal water consumption for 4 days. After this treatment period, the mice were euthanized, and their colons were harvested and sectioned for further analysis.
The levels of Ki67, which is a marker of cell proliferation, were examined to assess the regenerative capacity of the colon epithelial cells. Additionally, the presence of PAS + goblet cells, which produce mucus, Muc2 (a key mucin protein), and IL-1β (an inflammatory cytokine), were also evaluated. These measurements were performed to gain insights into the recovery and functionality of colon epithelial cells in both WT and
Fpr2−/− mice following the DSS treatment and subsequent recovery period.
5.
For quantification of colony forming units (CFU) of E. coli in feces, the mice given with 5% DSS for 5 days were euthanized and the colon with cecum (avoid fecal leakage) were harvested. Feces (50–70 mg per mouse) were used for culture of E. coli on Violet Red Bile Lactose agar (VRBL, EMD Millipore Corporation) to compare the number of E.coli between WT and Fpr2−/− mice. Single colonies on VRBL were harvested and amplificated in LB Broth, then extracted for E. coli DNA with the DNeasy UltraClean Microbial Kit (GIAGEN, MD). E. coli were identified with PCR and 16S rRNA gene sequencing.
Conventional PCR for genes of 16S rRNA and LpfA of E. coli
For the amplification of E. coli 16S rRNA and LpfA gene, a conventional PCR method was employed. Purified E. coli DNA was quantified using an NP-1000 Spectrophotometer (Thermo, MD) and adjusted to the same concentration for each sample, which was 1 µg/µl. For the amplification of E. coli 16S rRNA, the following primer sequences were used: primers 5′-TGG CTC AGG ACG AAC GCT GGC GGC-3′ (sense) and 5′-CCT ACT GCT GCC TCC CGT AGG AGT-3′ (antisense) were designed to yield a 348-bp product. The PCR amplification condition consisted of an initial denaturation step at 95 °C for 5 min, followed by 30 cycles of denaturation at 95 °C for 45 s, annealing at 58 °C for 1 min, extension at 72 °C for 45 s, and a final extension step at 72 °C for 10 min. The PCR products were separated and visualized on 1.5% agarose gels through electrophoresis, followed by staining with ethidium bromide. Similarly, for the amplification of E. coli LpfA gene, the following primer sequences were used: Sense primer: 5′-AGTTGGTGATAAATCACCAT-3′, Antisense primer: 5′-GTGCTGGATTCACCACTATTCATCG-3′. These primers were designed to yield a 222-bp product specific to the LpfA gene. The PCR amplification condition and product visualization were the same as mentioned above.
Histologic and immunohistochemical staining
To prepare the colon tissues for analysis, they were embedded in Optimal cutting temperature compound (OCT) and frozen. The frozen tissues were then sectioned into slices that were 10-μm thick. These sections were subsequently fixed in 8% neutral buffered formalin for 30 min to preserve their structure. After fixation, the sections were washed three times with distilled water (ddH2O) to remove any residual formalin. Hematoxylin and eosin (H&E) staining was performed on the sections to visualize the tissue morphology. The H&E-stained sections of the colon tissue were observed using an Olympus microscope equipped with a DP80 camera. Images of the crypts were captured for further analysis. The length of the crypts was measured using ImageJ (a Java-based image processing program developed at the National Institutes of Health).
For the assessment of colon mucosal damage, histopathological scoring was performed. The scoring system typically involves assigning grades ranging from 0 to 5 based on several criteria, including the extent of colon tissue affected, the extent of crypt damage, and the quantity and dimension of inflammatory cell infiltration [
30]. The scoring system provides a standardized way to evaluate the severity of colon mucosal damage. In this study, 4–8 mice were used for each mouse group, and the histopathological scoring was performed on multiple sections from each mouse to obtain representative data.
For detection of bacteria attaching to colon mucosa, Bacterial Gram Staining Kit (Abcam, MA) was used following the manufacturer’s protocol. Sections were dehydrated in absolute alcohol, cleared in xylene, and then mounted in a synthetic resin.
To detect goblet cells in the colon crypts, the slides were stained with the Alcian Blue Periodic Acid-Schiff (PAS) Stain Kit. 23–25 crypts from 4 mice per mouse group were analyzed for the quantification of PAS + cells.
Immunofluorescence staining
To perform immunofluorescence staining on the colon tissue sections, the frozen sections embedded in OCT were fixed in 8% neutral buffered formalin for 30 min. After fixation, the sections were washed three times with distilled water (ddH2O) to remove residual formalin. For the staining process, primary antibodies specific to the target proteins of interest, such as anti-mouse Muc2 (ab76774, Abcam), CRAMP (sc-66843, Santa Cruz), or β-Defensin 2 (ab203077, Abcam), Ki67 (ab16667, Abcam), IL-1β (AF-401-NA, R&D) antibodies and anti E. coli antibody (ab25823, Abcam), were applied to the sections. Following the primary antibody incubation, secondary antibodies conjugated with biotin (ab6720, ab208000, ab207997, ab207996, Abcam and BAF109, R&D), were applied to the sections followed with streptavidin-FITC (405202, BioLegend) or PE (405204, BioLegend), allowing for the visualization of the target proteins. The 6-diamidino-2-phenylindole (DAPI) was used to stain the cell nuclei. The fluorescence intensity of all immunofluorescence staining was measured by ImageJ.
Fluorescence in situ hybridization (FISH)
Detection of bacteria on colon mucosa was performed with freshly frozen, OCT-embedded, and sectioned slides (10-μm thick). The slides were fixed in 4% neutral buffered formalin for 5 min, washed twice with 1 × PBS (DEPC treated) and pre-hybridized with 200 μl 3% BSA at 37 °C for 2 h. The sections were placed in the Slide Griddle (Model SG96P, MJ Research, Inc., MA) and in 96 V PTC Thermal Cyclers (MJ Research, Inc., MA) and heated to denature at 84 °C for 5 min, 37 °C for 3 min before adding FISH probe (0.1–0.4 μm in Hybridization buffer) and incubated overnight at 40 °C. The slides were washed with washing solution (2 × SSC/0.1% Tween 20) for 15 min at RT followed by treatment with 75% and 100% ethanol, then air dried. DAPI-Antifade solution was added in the dark for 10 min to stain nuclei. The sections were then mounted in Resolve (Thermo Scientific, MI, USA) and stored at 4 °C. The probes conjugated to CY3 were used to detect bacteria adhering to colon epithelial cells. The universal bacterial EUB338 FISH Probe (5′-CTGCCTCCCGTAGGAGT-3′) (Creative Bioarray, NY) conjugated with CY3 was used to detect total bacteria. A ‘non-sense’ probe (5′-CGACGGAGGGCATCCTCA-3′) conjugated with CY3 was used as a negative control for EUB338. EC1531 FISH Probe (5′CACCGTAGTGCCTCGTCATCA-3′) conjugated with CY3 (Integrated DNA Technologies) was used to detect E. coli.
Enzyme-linked immunosorbent assay (ELISA)
To measure fecal Muc2, β-definsin-2 (DEFb2), cathelin-related antimicrobial peptide (CRAMP), and Lipopolysaccharides (LPS), freshly harvested feces were homogenized in PBS containing 1% NaN
3, 20 mM dithiothreitol, and a protease inhibitor mixture (P8340; Sigma, 1:200 dilution) [
31]. Fecal suspensions were centrifuged at 15,000 RPM for 10 min at 4 °C. The supernatants were collected and inactivated at 60 °C for 30 min, and their protein concentrations were determined using a DC Protein Assay Kit (Bio-Rad). The concentrations of Muc2, DEFb2, CRAMP and LPS in the feces were determined with ELISA specific to Muc2 (Cloud-Clone Corp., TX, USA), DEFb2 (Cloud-Clone Corp., TX, USA), CRAMP (MybioSource, CA, USA), and LPS (MybioSource, CA, USA) respectively. The Muc2 concentration was expressed in ng per 1 mg (mg) protein of stool. The concentrations of DEFb2 and CRAMP in the feces were expressed in pg per 1 mg (mg) protein of stool. The concentrations of LPS in the feces were expressed ng/g stool.
To measure IL-1β in the mucosa, colons were washed with PBS containing antibiotics (penicillin/streptomycin) and the distal 3 cm were isolated and further cut into1–2 mm sections. Colon sections were covered with RPMI media (1 ml), containing 1% FBS and penicillin/streptomycin overnight in a humidified 37˚C incubator with 5% CO2. Cell-free supernatants were harvested and their protein concentrations were determined using a DC Protein Assay Kit (Bio-Rad) [
32]. The samples were stored at − 80 °C for ELISA assay. IL-1β concentrations were measured by ELISA (Thermo Fisher). LPS in serum was measured with mouse LPS ELISA Kit (MybioSource, CA) and the concentration of LPS in serum were expressed in ng/ml.
Identification of E. coli
Single colonies of E. coli on Violet Red Bile Lactose agar were harvested and smeared onto slides and stained with Gram stain Kit (abcam, MA) and examined with Fluorescence in situ hybridization (FISH) with EC1531 probe and PCR to identify E. coli. The colonies of E. coli identified were selected for 16S rRNA sequencing. Single colonies with different 16S rRNA sequences were then used for whole genome sequencing. E. coli serotypes O22H8 and O91H21 were defined based on databases and kept in − 80 °C for further use.
Adhesion of colon epithelial cells by E. coli
Mouse epithelial CT26 cells (1 × 106 cells/ml) were cultured with E. coli (1 × 107 CFU/ml) at a multiplicity of 10 bacteria per cell in 14 ml polypropylene round-bottom tubes (FALCON, 352059) at 37 °C with shaking at 60 RPM for 2 h. The cells were smeared onto slides, stained with a three-step staining kit (Thermo Scientific Richard-Allan Scientific), and observed under an Olympus microscope with a DP80 camera.
E. coli-induced death of colon epithelial cells
One million CT26 cells were seeded in 35 mm dishes with 14 mm coverslips in the bottom and co-cultured with E. coli O22:H8 and O91:H21 at a multiplicity of 10 bacteria per cell for 6 h at 37 °C. The cells were washed two times with PBS, fixed with 4% neutral buffered formalin for 5 min and then stained with Live/Dead Viability/Cytotoxicity Kit for mammalian cells. The slides were observed under an Olympus microscope with a DP80 camera.
Chemotaxis assays
Cells were performed with 48-well chemotaxis chambers and polycarbonate filters (8-µm pore size) (NeuroProbe, Cabin John, MD). The results are expressed as the mean ± S.D. of the chemotaxis index, which represents the fold increase in the number of migrated cells, counted in three high power fields (×400), in response to chemoattractants over spontaneous cell migration (to control medium).
Testing E. coli in germ-free (GF) mice
GF mice were orally inoculated with E. coli O22:H8 or O91:H21 (2 × 108 CFU live E. coli per mouse). Five days after inoculation, all mice were sacrificed, or the mice were given 3% DSS for 4 days and colons were harvested.
Ethical statement
Animal protocols were in accordance with the recommendations of the US NIH Guide for the Care and Use of Laboratory Animals (National Academies Press, 2011) and were approved by the Frederick National Laboratory for Cancer Research Animal Care and Use Committee (ASP No. 21-464).
Statistical analysis
All experiments were performed at least three times with reproducible results. Statistical analyses, unless specified, were performed by GraphPad Prism 9 (GraphPad Software, San Diego, CA). A P value of < 0.05 was considered statistically significant.
Discussion
The microenvironment within the colon mucosa plays a crucial role in maintaining the balance of the gut microbiota. In this study, various defects associated with Fpr2 deficiency were observed, including shortened colon crypts, reduced production of Muc2 and antimicrobial peptides by epithelial cells, increased production of the pro-inflammatory cytokine IL-1β in the colon mucosa of naïve mice, and gut dysbiosis. There is mounting evidence to suggest that dysbiosis, characterized by an imbalance in the gut microbiota, is closely associated with inflammatory bowel disease (IBD) [
5,
37]. Dysbiosis patterns commonly observed in IBD patients are characterized by a reduction in the diversity of commensal bacteria, particularly Firmicutes, and a relative increase in species belonging to Enterobacteriaceae [
5,
38‐
40]. Multiple factors can disrupt the beneficial members of the gut microbiome, including antibiotic use, psychological and physical stress, radiation, altered gut peristalsis, and dietary changes [
41]. Genetic deficiencies, such as mutations in the nucleotide-binding oligomerization domain-containing protein 2 (NOD2), have also been observed to result in gut dysbiosis in patients [
42‐
48]. Muc2, serving as a primary barrier between the gut microbiome and the intestinal epithelium, plays a crucial role in maintaining gut homeostasis [
49] and
Muc2−/− mice develop severe colitis due to the absence of protective mucous layers [
50]
. Similarly, deficiencies in cathelin-related antimicrobial peptide (CRAMP), reduced expression of β-defensin 2, and alterations in the gut homeostatic protein Fam3D have also been shown to disrupt the balance of the gut microbiota [
29,
51‐
55]. In this study, we demonstrate that Fpr2 is essential for inducing epithelial cell growth and maturation, leading to the release of Muc2 and antimicrobial peptides, thereby maintaining the homeostasis of the colon mucosal microenvironment.
Enterobacterial blooms are frequently observed in cases of gut dysbiosis [
3].
E. coli, a member of
Enterobacteriaceae family, Proteobacteria phylum
, typically represents a minor fraction of the microbiome in a healthy human colon [
3,
56]
. However, an elevated presence of
E. coli, particularly mucosa-associated genotoxin-positive
E. coli, has been consistently found in a significantly higher proportion of patients with Crohn's disease (CD) or ulcerative colitis (UC), two forms of inflammatory bowel disease (IBD) [
38,
57‐
64]
. In mouse models, deficiencies in CRAMP and reduced expression of β-defensin 2 have been shown to disrupt the balance of the gut microbiota, resulting in the overgrowth of
E. coli in the colon [
29,
51‐
54]
. Histological analysis of human tumor samples has revealed extensive infiltration of inflammatory cells in pks-positive
E. coli-infected HCT116 tumors [
65], and the growth of tumors has been shown to be enhanced by colibactin-producing
E. coli in xenograft and AOM/DSS-induced tumor models [
66]
. Therefore, an increase in
E. coli counts serves as a significant indication of colon dysbiosis with potentially harmful consequences [
4,
67]
.
Recently, there has been significant attention given to the role of normal intestinal flora, including
E. coli, in the recovery from colitis [
68‐
71].
E. coli has been shown to promote the healing of colitis-related mucosal damage in the colon through the activation of the TLR4/NF-κB signaling pathway [
68]. Additionally,
E. coli has been found to protect mice from
Citrobacter rodentium infection and DSS-induced colitis [
69]. In our previous study, we demonstrated the expression of Fpr2 on mouse colonic epithelial cells, which can be activated by fMLF, a chemotactic peptide ligand derived from
E. coli. We observed that the intake of DSS increased the expression of Fpr2 in both immature and mature epithelial cells of colonic crypts, suggesting the importance of Fpr2 in enabling epithelial cells to respond to locally and systemically available ligands in pathological conditions.
E. coli O22:H8 produces ligands for both Fpr1 and Fpr2, which induce inflammatory cell migration [
9] and promote the restoration of chemically induced mucosal damage [
68]. In the current study, we identified two
E. coli serotypes, O22:H8 and O91:H21, through whole genome sequencing.
E. coli O22:H8 was present in the feces of both WT and
Fpr2−/− mice and, as a commensal bacterium, did not cause significant damage to colon epithelial cells compared to
E. coli O91:H21, both in vitro and in vivo.
Infection with
E. coli O22:H8 stimulated the upregulation of Fpr2 expression, and its products induced the migration and proliferation of colon epithelial cells through Fpr2. However, Fpr2 deficiency led to an increased population of
E. coli in the mouse colon and delayed the recovery of damaged colon epithelial cells, indicating the involvement of Fpr2 expression in the effects of commensal
E. coli. Nonetheless, the overgrowth of
E. coli O22:H8 observed in
Fpr2−/− mice has pathological significance, such as a significant reduction in inflammatory cell accumulation at
E. coli-infected sites and increased mortality in
Fpr2−/− mice [
9]. Thus, the Fpr2 deficiency-induced overgrowth of
E. coli O22:H8 appears to be harmful to the host.
In contrast to
E. coli O22:H8,
E. coli O91:H21 represents a relatively small proportion of
E. coli colonies isolated from the mouse feces in our study. It has been reported that
E. coli O91:H21 can be found in food, animals, or the environment and may cause severe diseases, including hemolytic-uremic syndrome [
36].
E. coli O91:H21 consists of many strains, some of which carry stx genes [
36]. Both O91:H21 and O22:H8 belong to the group of Shiga toxin-producing
E. coli (STEC) strains and express long polar fimbriae (lpfA) [
36], which is a potential adherence factor originally described in
Salmonella [
72,
73]. Four genetic variants of lpfA, namely lpfA (O157/OI-141), lpfA (O157/OI-154), lpfA (O26), and lpfA (O113), have been identified in Shiga toxin-producing
E. coli (STEC) [
74]. LpfA is associated with the adherence and invasion capacity of bacteria to epithelial cells [
75], and its expression in
E. coli strains is important for pathogenicity [
35,
76]. In our study, the gene for Long Polar Fimbriae Type I (lpfA), a recognized marker for virulent isolates of pathogenic
E. coli, is more readily amplified by PCR in
E. coli O91:H21 compared to the O22:H8 strain, although whole genome sequencing results revealed that both
E. coli O91:H21 and
E. coli O22:H8 have the same number of base pairs (in PPM) for the lpfA gene. When comparing symptomatic strains of STEC O91:H21, such as ATCC 51435 and ATCC 51434, which are known to be highly virulent in an experimental infection mouse model, with asymptomatic strains of STEC O91:H21 isolated from a STEC outbreak in Korea, the asymptomatic STEC O91:H21 isolates exhibited a significantly reduced adherence phenotype and cytopathic effects due to the transcriptional repression of the genes encoding type-1 fimbriae in the asymptomatic isolates [
77,
78]. This suggests the importance of fimbriae for
E. coli in adhering to and invading colon epithelial cells. Furthermore, analyses using the Virulence Factor Database (Vfdb) [
34] suggest the presence of three virulence factors in O91:H21 but not in O22:H8, which are produced by a generic Type II secretory machinery associated with enterotoxicity in
E. coli [
79]. Our studies in GF mice clearly demonstrate the more proliferative and invasive nature of
E. coli O91:H21 in the colon. However, further investigation is required to determine the origin of
E. coli O91:H21 in mice and whether the isolate we obtained belongs to a mutated variant of the originally harmless strain.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.