ABSTRACT
The translocation of bacteria across the intestinal epithelium of immunocompromised patients can lead to bacteremia and life-threatening sepsis. Extraintestinal pathogenic Escherichia coli (ExPEC), so named because this pathotype infects tissues distal to the intestinal tract, is a frequent cause of such infections, is often multidrug resistant, and chronically colonizes a sizable portion of the healthy population. Although several virulence factors and their roles in pathogenesis are well described for ExPEC strains that cause urinary tract infections and meningitis, they have not been linked to translocation through intestinal barriers, a fundamentally distant yet important clinical phenomenon. Using untransformed ex situ human intestinal enteroids and transformed Caco-2 cells, we report that ExPEC strain CP9 binds to and invades the intestinal epithelium. ExPEC harboring a deletion of the gene encoding the mannose-binding type 1 pilus tip protein FimH demonstrated reduced binding and invasion compared to strains lacking known E. coli virulence factors. Furthermore, in a murine model of chemotherapy-induced translocation, ExPEC lacking fimH colonized at levels comparable to that of the wild type but demonstrated a statistically significant reduction in translocation to the kidneys, spleen, and lungs. Collectively, this study indicates that FimH is important for ExPEC translocation, suggesting that the type 1 pilus is a therapeutic target for the prevention of this process. Our study also highlights the use of human intestinal enteroids in the study of enteric diseases.
INTRODUCTION
The gastrointestinal tract (GIT) not only functions in food digestion and excretion but also serves as an immunological barrier against microorganisms. However, under certain conditions, the bacteria residing in the GIT breach this barrier and migrate to extraintestinal sites, which can lead to gut-derived bacteremia and sepsis. The process by which bacteria and their products migrate from the GIT to the bloodstream and other organs is called “bacterial translocation,” which was first described as a clinical phenomenon by Wolochow and colleagues in the 1960s (1) and was investigated experimentally by Berg and coworkers in the 1980s (2–10). The latter set of studies defined several risk factors that contribute to translocation, including disruption of gut microflora by antibiotics (4, 6), damage or increased permeability of the intestinal epithelial barrier (4, 9, 11, 12), and the health status of the host (4, 5, 7, 8, 10–12). In addition, several studies have established an association between the intestinal microflora and systemic infections in immunocompromised individuals (13–18). Extraintestinal pathogenic E. coli (ExPEC), the general name prescribed to E. coli strains that cause infections distal to the intestine, is the most frequently isolated Gram-negative organism in such cases (14, 15, 17). ExPECs include E. coli associated with urinary tract infections (UTIs) (uropathogenic E. coli [UPEC]), neonatal meningitis (neonatal meningitis-associated E. coli [NMEC]), and septicemia (sepsis-causing E. coli [SEPEC]) (19–21). ExPEC strains are steadily acquiring resistance to commonly used antibiotics (22–25), and pandemic strains (sequence type 131) have been described previously (26).
ExPEC organisms are often indistinguishable from commensal organisms, colonize a sizable fraction of the healthy population (27), exhibit large genomic diversity, and carry multiple virulence factors in various combinations (21, 28, 29). Although several ExPEC virulence factors have been identified (21, 28, 29), they have not been directly linked to translocation from the intestine. This is problematic because knowledge of the bacterial factors that facilitate translocation would constitute attractive targets for the development of antitranslocative medicines. This study aims to fill this void by identifying the ExPEC factors involved in the early stages of translocation. Our findings establish a role for the mannose-binding type 1 pilus tip protein FimH in ExPEC invasion and translocation.
RESULTS
ExPEC invades the intestinal epithelium.Gut-derived bacteremia and sepsis originate from the movement of enteric bacteria from the intestinal lumen to the bloodstream, whereby a systemic infection and/or excessive immune response facilitates a life-threatening condition (13, 14). Passive translocation results when there is a physical break in the GI wall and intestinal contents, including bacteria, leach into the lymphatics and circulation (30). In contrast, active translocation occurs when the inherent properties of the bacteria allow for their passage through a GI barrier in the absence of evidence of pathology or trauma to the epithelial wall (31). We hypothesized that ExPEC would pass through an intact intestinal epithelium. To test this hypothesis, we adapted the use of a confluent and polarized Caco-2 monolayer, a system sometimes used to measure the invasion of Listeria (32, 33) and Salmonella (34), to measure that of ExPEC. This system allows the investigator to measure three fundamental steps in the process of translocation: adherence, invasion, and the actual movement through an intact intestinal barrier.
We first tested if ExPEC will adhere to Caco-2 cells. ExPEC strain CP9 and E. coli laboratory strain MG 1655 (K-12) were incubated with Caco-2 cells for 1 h, the monolayer was washed, and bacterial cells were enumerated by plating on Luria-Bertani agar (LB/agar). As demonstrated in Fig. 1A, both strains of E. coli adhered to Caco-2 cells. When gentamicin was added for 2 h to kill extracellular bacteria and the monolayer was assessed for the presence of E. coli (invasion), there was an average of 10 times more ExPEC than K-12 organisms (multiplicities of infection [MOIs], 10 and 50; P ≤ 0.05) (Fig. 1B). When mock infections in which the only omission from the experiment was the presence of the Caco-2 cells were performed, neither ExPEC nor K-12 survived.
ExPEC adheres to and invades Caco-2 cell monolayers. (A) ExPEC strain CP9 and E. coli laboratory strain K-12 were seeded at MOIs of 10 and 50, and bacterial adherence was measured as described in Materials and Methods. (B) Bacterial invasion was measured by gentamicin protection assay at MOIs of 10 and 50 (see Materials and Methods). Data are means ± SEM; n = 4 to 6, assayed in triplicate; *, P ≤ 0.05.
ExPEC invasion is dependent on the type 1 pilus adhesin FimH.We reasoned that virulence factors that are important for urinary tract infection may be a starting point for identifying the bacterial factors important for the invasion and translocation through the intestinal epithelium. Therefore, we assessed if iroN, cnf-1, or fimH is necessary for ExPEC's invasion and translocation through intestinal epithelial cells. The iroN gene encodes the salmochelin siderophore receptor, which is important for iron uptake in iron-limited environments (35, 36), cnf-1 encodes a potent cytotoxin that induces the rearrangement of the actin cytoskeleton (37), and fimH encodes a type I pilus tip adhesin protein (38). In particular, FimH and allelic variants of this protein are important for the invasion of the urinary epithelium, as well as the formation of intracellular bacterial communities (IBCs), a persistent intracellular infection that forms a reservoir of UPEC that is a source of recurring infections in the urinary tract (38–42).
We used recombineering and transduction to generate deletions of iroN, cnf-1, and fimH in strain CP9, and these strains were evaluated for invasion and translocation using the Caco-2 cell system described above. Wild-type, ΔiroN, and Δcnf-1 strains all adhered (see Fig. S1A in the supplemental material) and invaded (Fig. 2A) the Caco-2 cells with similar efficiency, independent of the MOI used. The fimH mutant strain's ability to invade and adhere was reduced at all MOIs, but this reduction was 3-fold and 6-fold at the higher MOIs, respectively (Fig. 2A, P ≤ 0.05; Fig. S1A). To further evaluate the role of the mannose-sensitive adhesin FimH on ExPEC adhesion and invasion, Caco-2 cells were infected with ExPEC CP9 or E. coli K-12 at an MOI of 50 in the presence of methyl α-d-mannopyranoside (mannoside), or methyl α-d-glucopyranoside (glucoside). The mannoside reduced the percentage of ExPEC associated with the Caco-2 cells compared to the untreated control (P ≤ 0.05) (Fig. S1C; ExPEC+mannoside), whereas the total number of intracellular bacteria was reduced almost 5-fold, indicating that the invasive phenotype is indeed partially due to the loss of FimH (P ≤ 0.05) (Fig. 2B, ExPEC+mannoside).
ExPEC invasion is partially dependent on the type 1 mannose-sensitive pilus adhesin FimH. (A) Caco-2 cells were infected with ExPEC wild-type (Wt) ΔfimH, ΔiroN, or Δcnf-1 strain, and invasion was evaluated at MOIs of 10 and 50 (see Materials and Methods). (B) Caco-2 cells were infected with ExPEC strain CP9 or E. coli K-12 at an MOI of 50 in the presence of mannoside or glucoside, and invasion was measured (see Materials and Methods). Data are means ± SEM, n = 3 to 6, assayed in triplicate; *, P ≤ 0.05.
Cultured transformed cancer cell lines, similar to the Caco-2 cells used in this study, are often the default infection model system for the study of enteric pathogens. To provide a more physiologic model system to further test the hypothesis that ExPEC adheres to and invades the intestinal epithelium, a process dependent on FimH, we developed the use of organotypic human intestinal enteroids to measure ExPEC invasion. Human enteroids are untransformed, composed of all the major cell types found in villus intestinal epithelium (enterocytes and Paneth, goblet, enteroendocrine, and stem cells) (43), display physiological responses upon infection (44–46), and can be grown in two dimensions (2-D) or 3-D. Here we used monolayers generated from multilobular 3-D human jejunal intestinal enteroids (HJIEMs) to assess the invasive capacity of ExPEC (see Materials and Methods). Wild-type, ΔfimH, or fimH-complemented ExPEC strains were applied (MOI, 100) to HJIEMs, and adhesion and intracellular invasion were measured. Whereas all ExPEC strains adhered to the HJIEMs (Fig. 3A), the strain lacking fimH was reduced by approximately 1 log in its ability to invade, a phenotype that was partially complemented with expression of FimH from a plasmid (Fig. 3B, ΔfimH) (P ≤ 0.05). Consistent with a role for FimH in this process, the invasion of enteroids was also inhibited by the addition of mannose (Fig. 3C, ExPEC with mannose) (P ≤ 0.05). These results mirror very closely those observed for wild-type ExPEC and the interaction with Caco-2 intestinal epithelial cells.
FimH mediates ExPEC invasion of human jejunal intestinal enteroid monolayers (HJIEMs). HJIEMs were infected at an MOI of 100 with ExPEC wild-type (Wt) or ΔfimH strain or pUC19:ΔfimH ExPEC strain CP9, and adherence (A) and invasion (B) were measured (see Materials and Methods). (C) The effects of mannose on wild-type (Wt) ExPEC and ΔfimH mutant strains were evaluated as described in Materials and Methods. Data are means ± SEM; n = 4 to 8, assayed in triplicate; *, P ≤ 0.05.
ExPEC translocation through a polarized intestinal monolayer is mediated by FimH.The finding that ExPEC could invade intestinal epithelial cells suggests that it may also be able to translocate through this barrier. We first tested if a 6-hour incubation of ExPEC with Caco-2 cells leads to a loss of intestinal cell barrier integrity as measured by a potential reduction in the transepithelial electrical resistance (TER). Polarized Caco-2 monolayers normally have resistances above 200 (Ω cm2) when grown on Transwell inserts (47, 48). At an MOI of 10, the TER of Caco-2 cells remained unchanged throughout the experiment for both ExPEC and E. coli K-12 and did not differ from cells that were not infected (Fig. 4A). However, at an MOI of 50 and starting about 3 h into the infection, the TER of the K-12 cells slightly increased with time, whereas that of the ExPEC-infected cells was reduced, a finding not observed for the uninfected control (Fig. 4B, ExPEC and K-12) (P ≤ 0.05). These results suggest that ExPEC strain CP9 does not disrupt the integrity of the epithelium at lower infection doses, such as an MOI of 10. It also indicates that we can test if ExPEC is capable of translocating through these cells in the absence of the physical disruption of the monolayer.
ExPEC translocation through a polarized intestinal monolayer is mediated by FimH. The transepithelial electrical resistance (TER) was measured at an MOI of 10 (A) or 50 (B) for Caco-2 cells infected with ExPEC strain CP9 or E. coli K-12 (see Materials and Methods). (C) Caco-2 cells were infected with strain ExPEC CP9 or E. coli K-12 at an MOI of 10, and levels of translocation were measured 6 h postinfection as described in Materials and Methods. (D) Translocation levels for wild-type (Wt) and ΔfimH, ΔiroN, and Δcnf-1 mutant strains at an MOI of 10 were evaluated after 6 h. Data are means ± SEM; n = 3 to 6, assayed in triplicate; *, P ≤ 0.05; ***, P ≤ 0.001.
To test this hypothesis, Caco-2 cells were grown on a Transwell insert seeded for 21 days until polarized (as measured by TER). ExPEC or E. coli K-12 (MOI, 10) was added to the apical surface (top chamber) of the monolayer for a 6-h incubation, and passage of bacteria to the basolateral side (bottom chamber) was measured by plating the medium on LB/agar. As demonstrated in Fig. 4C, some level of basal translocation was observed for E. coli K-12; however, approximately 2 and 3 orders of magnitude more ExPEC isolates passed from the apical to the basolateral side at the later time points, respectively (Fig. 4C, 4 h, P ≤ 0.05; 6 h, P ≤ 0.001). Collectively, these findings indicate that ExPEC can translocate through a polarized epithelial layer of intestinal origin grown on Transwell inserts. The finding that ExPEC, a leading cause of human bacteremia and gut-derived sepsis, can both invade and pass through an intestinal epithelium is a novel observation for this pathogen. Having identified FimH as a determinant of ExPEC's invasion, we next assessed the ExPEC mutant strains' ability to translocate through Caco-2 cells seeded on a Transwell system with apical and basolateral compartments. The mutant strains lacking iroN and cnf-1 migrated across the Caco-2 cell monolayer as well as the wild-type strains after 6 h; however, ExPEC lacking fimH did translocate at approximately 1 order of magnitude less than the wild type (P ≤ 0.05) (Fig. 4D, 4 and 6 h ΔfimH).
FimH is important for ExPEC translocation in vivo.Berg and colleagues defined translocation as the movement of bacteria from the intestinal lumen to extraintestinal sites, including the mesenteric lymph nodes and major organ systems (2–4). The Caco-2 and enteroid cell culture models indicate that ExPEC can invade and pass through an intact intestinal epithelium in a process partially dependent on FimH. To further test this hypothesis, we utilized a murine model of chemotherapy-induced ExPEC translocation that we developed (49) to evaluate if the loss of fimH affected ExPEC translocation during a developing infection. Mice were subjected to gavage with the wild-type, ΔfimH mutant, or fimH-complemented strain, colonization was established, and chemotherapy was administered as described previously (49). As shown in Fig. 5A, the levels of each mutant ExPEC strain shed into the feces from days 1 to 7 were comparable to wild-type levels, with a characteristic drop in the levels of colonized ExPEC after day 1 as previously observed (49). At 8 days, the mice were sacrificed and necropsied, and tissues were examined for the levels of ExPEC. Both the small and large intestines had a mean level of ExPEC ΔfimH that was slightly lower than the wild-type and the complemented strain levels, but this was not significant (Fig. 5B and C). When assessed for translocation, the fimH mutant displayed on average ∼1-log reduction in the mean CFU per gram of tissue in the kidneys, spleen, and lungs (kidneys, ΔfimH strain, P ≤ 0.05; spleen, ΔfimH strain, P ≤ 0.05; lungs, ΔfimH strain, P ≤ 0.05) (Fig. 5D to F). This phenotype was restored with the fimH-complemented strain (kidneys, ΔfimH strain, P ≤ 0.05; spleen, ΔfimH strain, P ≤ 0.01; lungs, ΔfimH strain, P ≤ 0.01) (Fig. 5D to F). Although there was a definite decrease in the levels of the fimH mutant strain in the liver and these levels were restored with the complemented strain, these trends were not significant (Fig. 5G). To determine if the higher numbers of wild-type isolates found in the kidneys, spleen, and lungs were due to some tropism for these tissues, we performed an intraperitoneal infection with the wild-type or fimH mutant strain into animals and assessed the levels of ExPEC in tissues. There was no difference in the levels of ExPEC in each organ between the wild-type and mutant strains (Fig. 5H).
FimH is important for ExPEC translocation of an immunocompromised host. (A to G) BALB/c mice were subjected to oral gavage with wild-type (Wt), fimH mutant, or complemented pUC19:ΔfimH ExPEC strains at 109 CFU/100 μl. Mice were given cyclophosphamide at a total dose of 450 mg/kg (three 150-mg/kg doses on days 1, 3, and 5) by intraperitoneal injection. Intestinal colonization was measured by fecal homogenates (days 1, 3, 5, and 7) (A) or intestinal homogenates (8 days postinfection immediately after euthanasia) (B, C) plated on LB/agar plus 10 μg/ml CM. (D to G) Translocation out of the GI tract was determined by the presence of bacteria in organs as described in Materials and Methods. Data are means ± SEM; n = 14 to 16 for each group. *, P ≤ 0.05; **, P ≤ 0.01. (H) For an intraperitoneal infection, BALB/c mice were given 107 CFU/100 μl of wild-type (Wt) ExPEC or fimH mutant strains. Colonization was measured by organ homogenates plated on LB/agar plus 10 μg/ml CM (see Materials and Methods). Data are means ± SEM; n = 8 for each group.
DISCUSSION
Bacterial translocation occurs in healthy individuals and is very important for the differentiation and maturation of the host's immune system (50). However, in those with weakened immune systems, either because of severe illness or medical treatments that render the patient immunocompromised, or because gut barrier integrity has been disrupted, the potential for translocation of resident enteric bacteria into the blood is high (13–18). Sepsis is responsible for approximately 1.6 million hospital stays in the United States per year (11). Although not all cases of sepsis originate from gut enterics, the most frequently isolated Gram-negative bacterium from the blood of such patients is E. coli (17, 51, 52), generally classified as ExPEC. Particularly concerning is the notion that humans, pets, or livestock may be asymptomatic carriers of these bacteria and thus constitute a hidden reservoir for high-risk individuals, with evidence of transmission between family members (53, 54). There are major knowledge gaps in our understanding of the bacterial factors that promote transmission, colonization, and the translocation of these bacteria. The results presented in this study advance this field by demonstrating that (i) ExPEC adheres to and invades a polarized Caco-2 intestinal monolayer as well as monolayers derived from human intestinal enteroids, (ii) this process is dependent on FimH, (iii) ExPEC translocates through polarized intestinal epithelial cells seeded onto a Transwell insert, (iv) translocation is also FimH dependent, and (v) FimH mediates translocation of ExPEC from the intestine to the peripheral organs upon immunosuppression in a murine host.
ExPEC, like most members of the Enterobacteriaceae family, produces type 1 pili, which aid in the attachment to mucosal epithelial surfaces. These pili are encoded by the fim operon, which consists of nine genes that are transcriptionally controlled by phase variation of an invertible promoter element (reviewed in reference 55). FimH, the tip protein, is well known to facilitate adhesion to the urinary epithelium and be a prominent player in UPEC pathogenicity (56–62). Adhesion via FimH is largely driven by the direct binding to mannose moieties on receptors expressed on the surface epithelium. The mannose-binding pocket of FimH does not vary among sequenced UPEC strains (63, 64), but residues outside this pocket can influence binding affinity and are evolving under positive selection in clinical UPEC strains (64). FimH is also required for UPEC to enter the bladder epithelium and persist in the form of IBCs (38–42). Because of its role in UPEC virulence, several therapeutic strategies have targeted FimH, including vaccines (65, 66), mannosides as competitive inhibitors of FimH binding (67–69), and compounds that inhibit FimH assembly (70), some of which can treat chronic cystitis caused by circulating UPEC strains in animal models (71). Our finding that FimH is also important for the early stages of translocation now implicates this pilus protein in the onset of gut-derived bacteremia. Since the GIT is the likely initial reservoir for these pathogens, future work should give strong consideration to the testing and use of these therapeutic measures in the GIT microenvironment as a means to prevent translocation.
It is well established that the genetic heterogeneity of E. coli as a species allows this pathogen to infect a variety of different human tissues, resulting in vastly different clinical outcomes or conditions. The combination of distinct virulence factors and antibiotic resistance elements, driven together primarily by horizontal transfer mechanisms, generates unique E. coli strains that may be not only more virulent but also selected for upon antibiotic treatment. The extensive pangenome of E. coli, which may contain as many genes as the human genome itself, offers a plethora of gene combinations that provide for an astounding level of adaptability (28). In addition to this versatility, there is now a growing body of evidence that virulence factors also harbor intergene allelic heterogeneity that may confer special pathoadaptive features to E. coli, a finding that has been observed with FimH and may have direct implications on the topic presented in this report (72, 73). For example, in addition to the fim operon capable of being acquired by horizontal transfer, this region of the genome undergoes de novo mutation at a higher rate than that of housekeeping genes, of which some changes seemingly enhance the binding of FimH to mannosylated receptors (74). Reports suggest that upwards of 30% of fimH genes harbor a different allele, a level of genetic diversity that is the basis for organizing strains into distinct sequence types (75, 76). Different FimH variants display different binding affinities to monomannose moieties (77), and these differences are factors in determining the environmental niche, intestinal or urinary, that E. coli may preferentially occupy, possibly by enhancing colonization, especially in the urinary tract (72). More-recent work suggests that some fimH alleles are undergoing positive selection that leads to enhanced fitness (64), with at least some of these alleles altering pilus tip conformation without directly acting on mannose contact residues (78). Not well understood at this point is whether different fimH alleles, which confer an altered functionality on the pilus tip protein, also influence the colonization or translocation of ExPEC through the gastrointestinal epithelium. Given the preponderance of different alleles in circulating ExPEC strains, the segmentation of certain alleles associated with highly virulent and resistant superbugs (74), and the already important role for them in uropathogenesis, understanding the effect of these alleles in gut-derived sepsis would appear to be a pressing need. It has been recently shown, using ExPEC strain UTI89 in a murine model, that type 1 and F-17-like pili enhance intestinal colonization (79). M4284, a high-affinity biphenyl mannoside that targets FimH, when used to treat a UTI also lowered intestinal colonization (79). In this study, colonization of the intestines by ExPEC CP9 lacking FimH was slightly lower on average than that by the wild type, but the greatest attribute hampered was its ability to translocate. Both studies highlight the gut as a reservoir for ExPEC infection and FimH as a therapeutic target to combat these infections.
The two issues left unresolved in this work are the determination of the mechanism by which ExPEC enters and translocates through the intestinal epithelium and the other virulence factors that induce this process. Understanding this mechanism and identifying the additional key determinants are of the utmost importance, since knowledge of this process can be harnessed to design smart drugs that prevent it. The data presented here suggest that ExPEC can enter intestinal epithelial cells, a process that is in part but not exclusively FimH dependent. One may put forth three general models that may explain how enteric commensals or pathobionts translocate from the GI lumen to the blood in immunocompromised patients, none of which are mutually exclusive. The first involves overt destruction of the intestinal epithelium and complete loss of barrier integrity. In this scenario, a reduced mucin layer and/or compromised epithelial junctions provide a clear path for commensals to reach the basolateral compartment. Barrier loss can be induced by medical treatments or trauma; for example, some cancer chemotherapies lead to increased permeability in the intestinal epithelium (80). Barrier loss may also be bacterially induced, and there is growing evidence that E. coli makes enzymes that “chip away” at host barriers. For example, ExPEC produces a protein termed surface and secreted lipoprotein E (SslE), which degrades mucin and is very important for translocation in a murine chemotherapy-induced model of gut-derived sepsis (49). Many clinical strains of E. coli also make α-hemolysin, a secreted lipase, which is reported to generate intestinal “focal leaks,” small holes in an otherwise-intact epithelium through which bacteria can paracellularly traverse enterocytes (81). In the two culture model systems (Caco-2 and HJIEMs) reported here, there is evidence of epithelial cell destruction (as measured by a drop in TER), but only at high MOIs and later time points postinfection. At least for ExPEC strain CP9, adherence, invasion, and translocation occur before a drop in TER. Nevertheless, the finding that SslE and hemolysin may aid in translocation, along with a role for FimH as defined in this report, implies that at least some of the translocation is a specific process mediated by distinct virulence factors.
A second potential way enteric bacteria may access the underlying lamina propria is if they are taken up and brought there by host immune cells such as dendritic cells (82) or enter through specialized enterocytes within the follicle-associated epithelium of the Peyer's patches (M cells) (83). Since glycoprotein 2 (GP2) can mediate the binding of FimH-containing bacteria and the subsequent transcytosis from the apical to basolateral compartment (83), our finding that FimH is important for translocation is consistent with this observation. Shigella species, which are closely related to E. coli, are thought to enter M cells to access the underlying lamina propria (84). Finally, a third potential mechanism is that ExPEC squeezes between cells to access the basolateral compartment. In this route, referred to as paracellular invasion and commonly associated with Campylobacter jejuni, it is thought that the pathogen disrupts tight and adherens junctions to facilitate transmigration (85).
The discussion of the mechanism of translocation must consider the proposed receptor or receptors needed to adhere to the intestinal epithelium. In the case of UPEC and the urothelium, FimH binds to uroplakins, mainly UP1a, found on the apical surface of the umbrella cells lining the lumen (86–88). The FimH-UP1a-dependent invasion of urothelial cells involves lipid rafts and its components caveolin-1 (89) and Rac1 (90). Alternatively, UPEC can also invade the urothelium in a zipper-like fashion caused by FimH interacting with α3β1 integrin, which activates a downstream signaling cascade involving focal adhesion, Src, and phosphoinositide 3-kinases, Rho-family GTPases, α-actinin, vinculin, and clathrin (38, 91, 92). This actin-dependent entry into the host cell can also be modulated by microtubules (93). After the FimH-induced internalization, UPEC can be found in exocytic vesicles (94) or in the cytoplasm, where it rapidly replicates to form IBCs supporting the establishment of a persistent infection (38–42). It appears that uroplakins are highly specific to urothelium; therefore, it is more than plausible that FimH-mediated invasion of intestinal epithelial cells involves integrin receptors; and we have preliminary data to support this hypothesis (unpublished data). There still remain several unanswered questions that warrant further investigation; however, our work has begun to unveil how ExPEC virulence determinants contribute to gut-derived sepsis.
MATERIALS AND METHODS
Bacterial strains.In this study, we used ExPEC strain CP9 (serotype O4:K54:H5) (95), a blood isolate from a bacteremic patient (96, 97), generously provided by James Johnson (University of Minnesota). This strain was selected because its virulence in mouse models is well characterized, some virulence factors for survival have been identified (97–100), and Green et al. (49) recently showed that it can translocate out of the mouse intestinal tract.
CP9 and K-12 MG1655 (provided by Christophe Hermann, Baylor College of Medicine) were modified through the addition of a chloramphenicol (CM) resistance and green fluorescence protein (GFP) cassette by classical P1 phage transduction. The same procedure was used to generate a knockout of fimH (ΔfimH) in ExPEC CP9 (referred to as the ΔfimH mutant). The complementing strain (pUC19:ΔfimH) was generated by amplifying the fimH gene using isolated wild-type CP9 chromosomal DNA using primers FimH1-FWD (5′-GTATACGGATCCATGAAACGAGTTATTACCC-3′), which included a BamHI site, and FimH2-REV (5′-GTATACGAATTCTTATTGATAAACAAAACGTCAC-3′), which included an EcoRI site. The product was cloned into pUC19 (New England BioLabs, Ipswich, MA, USA) and transformed into chemicompetent E. coli DH5-α cells (New England BioLabs, Ipswich, MA, USA) selected for on Luria-Bertani agar (LB/agar) with 100 μg/ml ampicillin (AMP; EMD Millipore, Darmstadt, Germany) coated with X-Gal (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside) and IPTG (isopropyl-β-d-1-thiogalactopyranoside; Sigma-Aldrich, St. Louis, MO, USA) for blue-white colony screening. The white colonies were screened using M13 forward and reverse primers for pUC19. The vector carrying fimH was isolated and then used to transform electrocompetent ΔfimH cells, and FimH was expressed by IPTG induction.
CP9 ΔiroN and Δcnf-1 mutants were generated by recombineering utilizing the λ red system. The wild-type strain was transformed with AMP-resistant pKD46, which harbors the λ red recombinase enzymes, and selected for accordingly. PCR products were made using primers that corresponded to upstream and downstream regions flanking the gene of interest and FLP recombination target (FRT) sites using pBA169CM:FRT as the template DNA (IroN1-FWD, 5′-CTCCAGCCTTCGGAAGCCAGAAGAATGGCTTTAATACGATCATATGAATATCCTCCTTAG-3′); IroN2-REV, 5′-CCGCTTTTACGCGGGCAGTGCGCCTGAAACACTACGATCAGTGTAGGCTGGAGCTGCTTC-3′; CNF-11-FWD, 5′-CATATGAATATCCTCCTTAGCCCGAAATACTGTATTTCTCAGCCATCAGTACAGCACTTTCATATGAATATCCTCCTTAG-3′; and (CNF-12-REV, 5′-GTGTAGGCTGGAGCTGCTTCCAGATCGCAGCACTGGAGCGTAAAGCCAGTGATGACGAGGGTGTAGGCTGGAGCTGCTTC-3′). These products were purified, digested, and transformed into electrocompetent cells grown in the presence of l-arabinose to induce recombinase enzymes. Transformed colonies were selected for on LB/agar with 10 μg/ml CM (EMD Millipore, Darmstadt, Germany), isolated, and screened for the recombineered product by PCR using amplifying primers complementary to regions upstream and downstream of the CM:FRT insertion site: IroN3-FWD (5′-ACAAACGCAATAACATCGCG-3′), IroN4-REV (5′-CCACCCTCACTATACTGAAA-3′), CNF-13-FWD (5′-GAAATACTGTATTTCTCAGC-3′), and CNF-14-REV (5′-CTCCAGTGCTGCGATCTGG-3′). After mutants were confirmed, pDK46 was removed from each strain by incubating streak plates overnight at 42°C. Colonies were sequentially tested for AMP sensitivity after pDK46 removal. For experiments, strains were grown from a single isolate in Luria broth at 37°C with aeration and diluted to the appropriate concentration.
Cell and tissue culture.Caco-2 colonic cells (clone C2BBe1; ATCC CRL-2102) were maintained in 75-cm2 flasks in Dulbecco's modified Eagle's medium (DMEM; Mediatech, Inc., a Corning subsidiary, Manassas, VA, USA) supplemented with 10% fetal bovine serum (FBS; Atlanta Biologicals, Flowery Branch, GA, USA) and a 1% mixture containing 100 U/ml penicillin–100 μg/ml streptomycin (Thermo Fisher Scientific, Boston, MA, USA). Cells were subcultured weekly using Ca2+/Mg2+-free Hanks balanced salt solution (Mediatech, Inc., a Corning subsidiary, Manassas, VA, USA) and 0.25% trypsin–0.2% EDTA (Mediatech, Inc., a Corning subsidiary, Manassas, VA, USA). Then, cells were seeded at a density of 5 × 104 cells/ml at 37°C in the presence of 5% CO2 in a humidified incubator.
Human jejunum intestinal enteroids (HJIE) were derived from jejunal intestinal crypts extracted from tissues obtained from patients undergoing bariatric surgery biopsies using the protocol of Sato et al. (43) and as previously described (44). Briefly, minced intestinal tissue was prepared and placed in a 6-well plate with approximately 3 ml of complete chelating solution (CCS; consisting of 100 ml incomplete chelating solution [500 ml Milli-Q H2O, 2.49 g Na2HPO4 · 2H2O, 2.7 g KH2PO4, 14 g NaCl, 0.3 g KCl, 37.5 g sucrose, and 25 g d-sorbitol], 400 ml Milli-Q H2O, and 40 mg DL-dithiothreitol) with 120 μl of 0.5 M EDTA, and the plates were orbitally shaken (∼250 rpm) at 4°C for 30 min. Two milliliters of FBS was added and mixed by pipetting, and the tissue was allowed to settle. Crypts were isolated as described previously (44) and resuspended in 2 ml of complete medium without growth factors (CMGF−), which consisted of advanced Dulbecco's modified Eagle medium/F-12 (Gibco, Carlsbad, CA, USA) supplemented with 1% 100 U/ml penicillin–100 μg/ml streptomycin, 1% 1 M HEPES (Thermo Fisher Scientific, Boston, MA, USA), and 1% 100× GlutaMAX (Thermo Fisher Scientific, Boston, MA, USA), followed by a final spin at 200 × g for 5 min at 4°C. The supernatant was removed, and the pellet containing the crypts was resuspended in growth factor and phenol red-free BD Matrigel basement membrane matrix (BD Biosciences, San Jose, CA, USA) on ice to avoid solidification (30 μl). Once plated, the Matrigel-enteroid culture was allowed to solidify at 37°C for 10 to 15 min, and the cultures were overlaid with 500 μl of complete media with growth factors (CMGF+, described below). All CMGF reagents were purchased from Thermo Fisher Scientific, Boston, MA, USA, or R&D Systems, Inc., Minneapolis, MN, USA. Complete medium with growth factors (CMGF+) consisted of CMGF− with 100 ng/ml of epidermal growth factor (EGF), 20% R-spondin-conditioned medium (Trevigen cell line HA-R-Spondin-Fc293T), 10% Noggin conditioned medium produced according to the methods described by Heijmans et al. (101), 50% Wnt3 conditioned medium (ATCC L-Wnt3 cell line CRL-2647), 1× N2 supplement, 1× B27 supplement, 1 mM n-acetylcysteine, 10 nM gastrin, 10 mM nicotinamide, 500 nM A83 (transforming growth factor beta 1 [TGF-β1] type-I receptor inhibitor), and 10 μM SB202190 (P38 inhibitor). The protocol for generating HJIE monolayers (HJIEMs) was modified from VanDussen et al. (102) and established by Estes's laboratory (103). Briefly, a 2.5% solution of Matrigel in cold phosphate-buffered saline (PBS) was overlaid into the wells of a 96-well plate. Four-day-old undifferentiated HJIEs were mixed with 0.05% trypsin–EDTA to promote the disruption of the enteroid into individual cells, the mixture was filtered through a 40-μm cell strainer to remove large aggregated cell structures, and the cell suspension in CMGF+ with 10 μM Y-27632 (ROCK inhibitor; Sigma-Aldrich, St. Louis, MO, USA) was added to the matrigel-coated wells to promote the formation of an enteroid monolayer. After 1 day, the medium was changed to differentiating medium (CMGF+ without R-spondin, Wnt3, nicotinamide, and SB202190) to promote differentiation and formation of the intact and polarized monolayer.
Bacterial adhesion and invasion assays.For adhesion (modified from reference 32) and invasion (modified from reference 33) assays, Caco-2 cells were seeded in 24-well plates at a density of 5 × 104 cells/well. Cells were grown until visible dome formation, which indicates differentiation based on the unidirectional movement of water and ions (104). Monolayers were washed once with PBS to remove remnants of antibiotic from culture medium, and then DMEM with 10% FBS with or without 1% d-mannose, methyl α-d-mannopyranoside (mannoside), or methyl α-d-glucopyranoside (glucoside) (Sigma-Aldrich, St. Louis, MO, USA) was added. Diluted overnight bacterial cultures were washed once with PBS, resuspended in DMEM plus 10% FBS, and added to Caco-2 monolayers at a multiplicity of infection (MOI) of 1, 10, or 50, followed by incubation for 1 h at 37°C in the presence of 5% CO2 in a humidified incubator. To quantify adherent bacteria, monolayers were washed 2 times with PBS and lysed with cold PBS containing 0.1% Triton X-100, and bacteria were enumerated by plating on LB/agar plus 10 μg/ml CM. For invasion, monolayers were washed as previously described and subsequently incubated in DMEM plus 10% FBS containing 50 μg/ml gentamicin (Santa Cruz Biotechnology, Inc., Dallas, TX, USA) for an additional 2 h to kill extracellular bacteria. Caco-2 monolayers were washed and lysed, and intracellular bacteria were enumerated as mentioned above. The HJIEMs also were used to determine bacterial internalization. The HJIEMs were seeded as noted above in the protocol and infected with an MOI of 100.
Transepithelial bacterial translocation.Bacterial translocation (modified from reference 34) was determined by seeding Caco-2 cells at 1 × 105 cells/Transwell insert (filter pore size, 3.0 μm in diameter; BD Biosciences, San Jose, CA, USA) in 24-well plates and allowing differentiation for 21 days to form a tight polarized monolayer. To check the integrity of the monolayer, transepithelial electrical resistance (TER) was measured (Voltmeter; Millipore, Billerica, MA, USA), and the monolayers with a TER of at least 200 Ω cm2 were used (47, 48). The wells and inserts were washed once with PBS, and then DMEM with 10% FBS was added to the top and bottom chambers. Diluted overnight bacterial cultures were washed once with PBS, resuspended in DMEM–10% FBS, and added apically to the Caco-2 monolayers at an MOI of 10 or 50. To determine the number of translocated bacteria after 1, 2, 4, and 6 h, medium from the bottom chamber was removed, and the bacteria were enumerated by plating on LB/agar with 10 μg/ml CM. During the infection, the effect of bacteria on monolayer integrity was monitored by TER measurements at times 0, 3, and 6 h.
Murine model of translocation.Our lab has previously established a murine model of chemotherapy-induced ExPEC translocation (49), which we used here to evaluate the role of FimH in this process. Briefly, 8-week-old male and female BALB/c mice were infected with 1 × 109 CFU ExPEC CP9 wild-type, ΔfimH, and pUC19:ΔfimH strains. Mice infected with pUC19:ΔfimH were also given water containing 10 mM IPTG (changed on the 4th day postinfection) to induce and maintain FimH expression (105). The day of the experiment (day 0), the cultures were centrifuged at 15,000 rpm, washed twice, and suspended in PBS. Each mouse (n = 14 to 16 per group) was subjected to gavage with a sterile (20-gauge, 38-mm) flexible needle that contained 1 × 109 CFU/100 μl. Cyclophosphamide (Baxter Healthcare Corporation, Deerfield, IL, USA) was dissolved in sterile water at a concentration of 10 mg/ml, and the mice were given a total dose of 450 mg/kg of body weight (3 × 150-mg/kg doses on days 1, 3, and 5) by intraperitoneal injection. Colonization was measured by plating fecal homogenates obtained on days 1, 3, 5, and 7. On day 8, mice were euthanized with CO2 (2 liters/min, 5 min), and organs (lungs, spleen, kidneys, liver, small intestine, and large intestine) were weighed, homogenized, and serially diluted. For the intraperitoneal infection, each mouse was given 107 CFU/100 μl of wild-type or fimH mutant strains, and the extent of dissemination was assessed 16 h postinfection as described above. All experiments were done with approval by Baylor College of Medicine’s Institutional Animal Care and Use Committee.
Statistical analysis.Adhesion, invasion, and translocation experiments data are means ± standard errors of the means (SEM) of assays (n = 3 to 8) performed in triplicate and over several passages of cells. For the chemotherapy-induced murine model of translocation, each group (3 groups) had 14 to 16 animals, and data are means ± SEM. The intraperitoneal murine model of infection had 8 animals for both groups, and data are means ± SEM. Statistical significance was determined by the Kruskal-Wallis and two-tailed unpaired t tests; differences between groups were assessed by Dunn's and Mann-Whitney test employing GraphPad Prism version 7.03 Windows (GraphPad Software, San Diego, CA; www.graphpad.com ). Differences in means were considered significant at P levels of ≤0.05.
ACKNOWLEDGMENTS
We thank James R. Johnson from University of Minnesota Medical School for donating the ExPEC strain CP9 and Christophe Herman from Baylor College of Medicine for donating the K-12 strain MG1655. We also thank Mary Girard for her assistance in developing mutant strains using recombineering and P1 phage transduction. Our sincere gratitude to Joseph Hyser, Sarah Blutt, and Noah Shroyer from Baylor College of Medicine for their help in generating monolayers from human enteroids.
FOOTNOTES
- Received 10 August 2017.
- Accepted 11 August 2017.
- Accepted manuscript posted online 14 August 2017.
Supplemental material for this article may be found at https://doi.org/10.1128/IAI.00581-17 .
- Copyright © 2017 American Society for Microbiology.
REFERENCES
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.↵
- 14.↵
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.↵
- 46.↵
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.↵
- 56.↵
- 57.↵
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.↵
- 78.↵
- 79.↵
- 80.↵
- 81.↵
- 82.↵
- 83.↵
- 84.↵
- 85.↵
- 86.↵
- 87.↵
- 88.↵
- 89.↵
- 90.↵
- 91.↵
- 92.↵
- 93.↵
- 94.↵
- 95.↵
- 96.↵
- 97.↵
- 98.↵
- 99.↵
- 100.↵
- 101.↵
- 102.↵
- 103.↵
- 104.↵
- 105.↵
