SUMMARY
Proteus species, members of the Enterobacteriaceae family, are usually considered commensals in the gut and are most commonly recognized clinically as a cause of urinary tract infections. However, the recent identification of Proteus spp. as potential pathogens in Crohn's disease recurrence after intestinal resection serves as a stimulus to examine their potential role as gut pathogens. Proteus species possess many virulence factors potentially relevant to gastrointestinal pathogenicity, including motility; adherence; the production of urease, hemolysins, and IgA proteases; and the ability to acquire antibiotic resistance. Gastrointestinal conditions that have been linked to Proteus include gastroenteritis (spontaneous and foodborne), nosocomial infections, appendicitis, colonization of devices such as nasogastric tubes, and Crohn's disease. The association of Proteus species with Crohn's disease was particularly strong. Proteus species are low-abundance commensals of the human gut that harbor significant pathogenic potential; further investigation is needed.
INTRODUCTION
Proteus species are members of the Enterobacteriaceae family of bacteria. Most commonly, they are recognized clinically as a cause of urinary tract infections. Although Proteus spp. are typically considered commensals in the gastrointestinal (GI) tract, their abundance as a proportion of the microbial community is very low (<0.05%) (1). As a result, their detection in disease states using 16S profiling, and possibly metagenomics, may have rendered Proteus spp. undetectable due to bioinformatic abundance thresholds.
The recent identification of Proteus spp. as potential pathogens in Crohn's disease recurrence after intestinal resection (2, 3) serves as a stimulus to examine their potential role as gut pathogens. This review aims to provide an overview of the genus Proteus in terms of its known virulence factors as well as to collate the evidence surrounding the role of Proteus spp. in the pathophysiology of gastrointestinal diseases.
CHARACTERISTICS OF THE PROTEUS GENUS
Proteus spp. are Gram-negative bacteria belonging the Enterobacteriaceae family and are common commensals of the gastrointestinal microbiota (4). The first isolates were reported and characterized by Hauser in the late 19th century (5). The genus is currently comprised of Proteus mirabilis, P. vulgaris, P. penneri, P. hauseri, P. terrae, and P. cibarius, along with the unnamed genomospecies 4, 5, and 6 (6–10). In humans, all current members of the genus, except for P. cibarius and P. terrae, have been isolated from clinical specimens (4, 5, 7, 9, 11). Typically, the human gut is colonized by various combinations of P. vulgaris, P. mirabilis, and P. penneri, but they comprise less than 0.05% of the gut microbiota of healthy subjects (Fig. 1) (1, 12). While Proteus spp. are widely recognized as pathobionts and the gut is the reservoir of these bacteria, the research focus on this genus has been on their role in urinary tract infections rather than intestinal manifestations (13–15).
Phylogenetic tree showing the species from the Enterobacteriaceae family that colonize the human gastrointestinal tract. GenBank accession numbers of the16S rRNA gene sequences are provided for each species, and the family names are indicated. E. coli is highlighted in green, and the Proteus genus is shown in blue. (Reproduced from reference 133.)
Many recent studies of the gut “microbiome” in health and disease have revealed that there are gross alterations in the relative proportions of key bacterial taxa associated with active disease, which is generically referred to as “dysbiosis.” One of the hallmarks of dysbiosis in the inflammatory bowel diseases (IBDs) is the population expansion of the phylum Proteobacteria, specifically the Enterobacteriaceae (16). Other genera within the Enterobacteriaceae family, such as Escherichia, Shigella, Salmonella, and Klebsiella, have received due attention in this regard, while Proteus has not been comprehensively investigated.
Pathogenic FeaturesProteus species are short (1.5- to 2-μm) straight rods that demonstrate dimorphism as “swimming” and “swarming” forms, as do some other members of the Enterobacteriaceae family (17). Swimmer cells predominate in liquid environments as single cells with 4 to 10 peritrichous flagella (Fig. 2, bottom right) (5, 14). The swarming behavior of Proteus species results in a characteristic bull's-eye pattern on a plate culture, as a result of a cyclic process of swarming and consolidation phases (18). When Proteus cells are placed in a viscous environment or on a solid surface, they undergo differentiation to filamentous, multinucleated, highly flagellated swarmer cells (Fig. 2, top and bottom left). Following this differentiation, a consolidation phase occurs, where the cells revert to a shorter morphotype, and metabolic preparation occurs prior to the next swarming cycle (18).
(Top) Strain of P. mirabilis inoculated twice, 1 h apart, demonstrating the macroscopic characteristic bull's-eye pattern produced by periodic swarming. (Reproduced with permission from reference 134.) (Bottom left) Interacting P. mirabilis swarmer cells; (bottom right) combination of swimmer and swarmer cells within a biofilm. (Both panels reproduced from reference 135 with permission from Elsevier.)
Proteus mirabilis undergoes swarming differentiation at much higher concentrations of agar (1.5 to 2%) than other swarming bacteria (19). When Proteus spp. swarm, there is a dramatic increase in the production of secreted proteins, including virulence factors such as the protease ZapA (17, 20, 21). In vivo, swarmer cells have been demonstrated in mouse models of ascending urinary tract infection only infrequently, with the occasional swarmer cell being isolated from the kidneys and bladder stones of infected mice (22, 23). The swarming phenotype can occur under both aerobic and anaerobic conditions (24) and can be induced by the concentration of amino acids, in particular glutamine (25, 26). It has also been demonstrated that a more acidic pH, as might be expected to occur in the proximal small bowel and cecum, dramatically increased swarming behavior (27–29). Swarming has been shown to be an important factor in intracellular invasion and persistence, with 15- to 20-fold more swarmer cells than swimmer cells being capable of the intracellular invasion of uroepithelial cells (30). There is some evidence that swarming Proteus strains are more invasive in urinary tract mouse models than are swarm-defective mutant strains (22). Furthermore, a number of metabolites present in the intestinal tract have been shown to promote swarming, including choline, glutamine, and the most abundant polyamine in the gut, putrescine (25, 31–33). While we cannot yet conclude definitively that swarming behavior occurs in the gut in vivo, the combination of a viscous surface (such as the gut mucosa), the high availability of glutamine and polyamines such as putrescine (26, 31, 33), and electron acceptors for anaerobic respiration such as choline (32) makes it likely that the gut environment may be permissive for swarming.
Adhesion and mucosal attachment.Adhesion to epithelial surfaces is essential for the pathogenesis of Proteus infections in both the urinary and gastrointestinal tracts. Sequencing of Proteus mirabilis strain HI4320 revealed 17 fimbrial gene sets (operons), more than any other bacterial genome currently characterized (34–36). Six of the fimbrial types that Proteus mirabilis can produce have been characterized (Table 1), including mannose-resistant Proteus-like fimbriae (MR/P fimbriae), mannose-resistant Klebsiella-like fimbriae (MR/K fimbriae), nonagglutinating fimbriae (NAF) (also known as uroepithelial cell adhesin [UCA]), ambient-temperature fimbriae (ATF), P. mirabilis P-like pili (PMP), and P. mirabilis fimbriae (PMF) (5, 36). These fimbriae and adhesins play a major role in the formation of bacterial biofilms, a common complication of both urinary and gastrointestinal instrumentation (37). It is likely that the MR/P and NAF/UCA fimbrial types are most important in gastrointestinal pathogenesis, due to their role in epithelial adhesion (38–41).
Fimbriae and pili expressed by Proteus speciesa
A comparison of the 17 individual chaperone-usher fimbrial operons across the 7 sequenced P. mirabilis strains as well as 58 clinical isolates showed 99% conservation in 13 of 17 fimbrial operons, demonstrating that these genes are highly conserved across strains isolated from various clinical sites (36). Of these, it is likely that at least two, and up to six, of the characterized fimbriae can be assembled on the cell surface at any one time (36, 42).
The regulation of motility and the expression of adhesion factors are tightly coupled; of the 17 fimbrial operons, at least 10 gene clusters possess a homolog of the mrpJ gene, a repressor of motility (43). MrpJ downregulates the flagellar master regulator flhDC via binding to the promoter sequence (43). When these cells revert back to the swimming morphology, the expression of MR/P and NAF fimbriae returns (37, 44). Furthermore, the induction of MR/P fimbriae appears coupled to the availability of oxygen (45). MR/P fimbriae are phase variant and can be “switched” on or off by a site-specific DNA recombinase (MrpI) that inverts a promoter region flanked by inverted repeats. This “invertible element” switches on or off the MR/P fimbrial operon, depending on the environmental conditions. There appears to be a growth advantage to MR/P fimbrial expression under low-oxygen conditions, such as those that would be present in the intestinal tract, likely contributing to the adhesiveness and persistence of Proteus species in the gut (45). The adaptation of Proteus species to mucosal surfaces by way of both fimbrial expression (for adherence) and swarming motility could increase the invasiveness, persistence, and pathogenicity of these species in the gut (Table 1 and Fig. 3).
Potentially important Proteus-related virulence factors in relation to anatomical disease location and disease. *, immune evasion includes the production of the ZapA metalloprotease, O-antigens, and flagellin variation.
Urease.The urease enzyme is a microbiological adaptation to metabolize urea, the most abundant nitrogenous waste product of human metabolism (46). Urease generates ammonia and carbonic acid as end products, with this ammonia providing a rich source of nitrogen for microbial metabolism in the gut (47). Urease confers a survival advantage to Proteus by providing nitrate for nonfermentative anaerobic respiration. This in turn promotes the population expansion of the Enterobacteriaceae (including Proteus) (48, 49). Additionally, as with Helicobacter pylori, the presence of this enzyme likely confers a survival advantage through increasing the local pH of the environment, allowing urease-positive organisms to survive in more-acidic environments such as the upper digestive tract. Proteus spp. have a wide pH range at which they are able to grow, from pH 5 to 10, with an optimal pH of 7 to 8 (9). Urease activity has been confirmed for P. mirabilis, P. vulgaris, and P. penneri, and production is regulated by both the environmental concentration of the substrate (urea) and increased chromosomal transcription in swarming cells (50–52).
Colonization of the upper gastrointestinal tract by urease-producing Proteus species may cause occasional false-positive results for the [13C]urea breath test (UBT), especially if the patient has already been treated with proton pump inhibitors and antibiotics for H. pylori eradication (53). Given that urease is a known contributor to the pathogenesis of H. pylori-mediated damage in the upper gastrointestinal tract, it may also contribute to the pathogenicity of Proteus spp. in the gut in the setting of environmental perturbations, such as changes in pH.
Hemolysins.The Proteus genus produces two distinct cytotoxic hemolysins, HpmA and HlyA (51, 54, 55). P. mirabilis and most P. vulgaris strains produce only HpmA, most P. penneri strains produce HlyA, and a few isolated P. vulgaris strains produce both HpmA and HlyA (54, 56, 57). HpmA has been shown to lyse erythrocytes, bladder epithelial cells, B-cell lymphoma cells, and monocytes, while HlyA can lyse erythrocytes, fibroblasts, and neutrophils (55, 56). HpmA is a cell-associated hemolysin, encoded on the hpm locus along with HpmB (an activator and chaperone of HpmA). The expression of these hemolysin proteins is tightly coupled to the swimming-swarming cycle, with swarming cells being 18-fold more cytotoxic than swimmer cells (30). HpmA has also been shown to lyse erythrocytes under anaerobic conditions and at multiple temperatures (58). The contribution of hemolysins to gastrointestinal pathogenesis may occur through the lysis of innate immune cells, the induction of the NOD-like receptor protein 3 (NLRP3) inflammasome, and downstream interleukin-1β (IL-1β) release (59).
Intracellular invasion and persistence.Intracellular invasion by Proteus mirabilis has been assessed mainly by using cellular invasion assays in cell lines ranging from uroepithelial cells to colonic cell types.
In a cell-based urinary tract model, swarming cells were 15-fold (0.18% intracellular entry) more invasive to uroepithelial cells than swimmer cells (0.012% intracellular entry) (30). After the invasion of uroepithelial cells, swarmer cells start to divide, develop septums, and differentiate back to an average of 50 to 300 swimmer cells within the cytoplasm (30). P. mirabilis has been identified as being intracellularly invasive in a number of cell lines (summarized in Table 2). There are differences in the intracellular invasion and uptake pathways depending on the cell type, with intracellular vacuoles containing P. mirabilis having a single membrane within urothelial cells and a double membrane within intestinal cells (60, 61). These mechanisms may contribute to effective intracellular colonization (cytoplasmic colonies), evasion of the host immune system, and resistance to antibiotics (60).
Cell culture lines capable of intracellular uptake of Proteus mirabilis
Translocation of Gram-negative Enterobacteriaceae (including P. mirabilis) has been demonstrated in vitro, in terms of both uptake and intracellular persistence. By using Caco-2 and HT-29 cells (enterocyte cell lines), it was demonstrated that P. mirabilis not only is invasive but also can survive intracellularly for ≥20 h without affecting the viability of enterocytes (62). The origin of the P. mirabilis isolate influences the invasiveness of Proteus species in laboratory studies, with fecal isolates showing higher invasion efficiencies (61).
Proteus mirabilis is a common cause of pathogenic infection of bladders augmented with bowel segments (enterocystoplasty). A study investigated the colonization of colonic cell lines by P. mirabilis as a model for stone formation in enterocystoplasty. The HT29-18N2 cell line (a mucus-secreting intestinal goblet cell subclone of HT-29 [63]) underwent widespread cellular destruction, with extensive intracellular bacterial colonies. Furthermore, upon confocal microscopy, there was colocalization of P. mirabilis and both human colonic mucin MUC2 and human gastric mucin (MUC5AC) (63). There is experimental evidence that Proteus spp. can invade at least the outer mucous layer of the gut and the inner layer to the epithelial surface if there are certain immune defects. These defects include the absence of the Lypd8 protein or defects in innate immunity (the T-bet−/− × Rag2−/− colitis [TRUC] mouse model) (64–66). Proteus mirabilis has many invasive characteristics; however, they remain to be directly characterized in the context of gastrointestinal disease.
Immune evasion.In addition to intracellular invasion, P. mirabilis also secretes an extracellular metalloprotease (ZapA) that has been shown to degrade a wide range of substrates (Table 3). This enzyme has a key role in the evasion of innate immune destruction by Proteus species by the proteolytic digestion of secretory IgA, IgG, and other cellular components in the urinary and gastrointestinal tracts (20, 21). Kerr et al. showed ZapA cleavage products (IgA and IgG) in the urine of patients with P. mirabilis, reflecting protease activity in vivo (67). The expression of this protein has been shown to be correlated tightly with cellular differentiation from swimmer to swarming cells by P. mirabilis, with increased transcription of the ZapA locus (20). ZapA hydrolyzes human β-defensin 1, a constitutively expressed innate immune antimicrobial peptide that is expressed in the colonic epithelium, as well as secretory IgA (20, 21). The expression of ZapA in the gut may provide a survival advantage to Proteus spp. by perturbing host mucosal immune responses, increasing the likelihood of persistent colonization. The genomic rearrangements encoding the flagellin proteins of Proteus mirabilis and the antigenic variability of Proteus species lipopolysaccharide (LPS) (both reviewed below) also contribute to evasion of the host immune system (Fig. 3).
Substrates degraded by Proteus mirabilis ZapA metalloprotease
Endotoxin and flagellins.As a Gram-negative pathogen, Proteus species possess intrinsic characteristics similar to those of other Enterobacteriaceae, such as Escherichia coli and Salmonella enterica serovar Typhimurium, including the production of flagellin and the proinflammatory cell wall component LPS (68, 69). A constituent of LPS, endotoxin (lipid A), is highly immunostimulatory (70). Endotoxin is sensed by the innate immune system (specifically Toll-like receptor 4) and activates the downstream signaling of NF-κB (71). This, in turn, triggers a proinflammatory cascade (mediated by tumor necrosis factor alpha [TNF-α]) that can lead to acute sepsis (72, 73).
In addition to lipid A, Proteus species LPS includes O-antigens (polysaccharides that are repeating oligosaccharide units, each made up of 2 to 8 sugar residues) that are highly structurally diverse. Approximately 80 O-antigen serogroups have been reported, derived from a total of 60 O-antigen gene clusters (11, 73). Antibodies to O-antigens are not uncommon in human sera, for example, 25% of blood donors have anti-P. mirabilis antibodies to the O36 serogroup, which is just one of the many serogroups (74). The expression levels of virulence factors, such as urease, proteases (ZapA), and hemolysin, can vary significantly between the O-antigen serogroups, with negatively charged O-polysaccharide serogroups having higher ureolytic, proteolytic, and swarming activities (75). This heterogeneity of surface structures occurs across all Proteus species.
In addition, bacterial flagellins, the repeating protein subunits from which flagella are built, are highly immunogenic due to their three-dimensional structure (76). Bacterial flagellin is sensed by Toll-like receptor 5, which activates a number of downstream inflammatory pathways, including MyD88 (72, 76, 77). Recombination of the flagellin genes flaA and flaB, leading to hybrid flagellin proteins with significant antigenic variation, also contributes to innate immune evasion by Proteus spp. (78, 79) These features contribute to the overall pathogenicity of Proteus spp. via stimulation of the host innate immune system by bacterial products (Fig. 3).
Conjugation, plasmid acquisition, and antibiotic resistance.Proteus species are inherently antibiotic resistant. Resistance to polymyxins is mediated via covalent modifications of lipid A (80). The substitution of l-arabinoso-4-amine for either the Kdo residue or the ester-linked lipid A phosphate moiety increases the overall charge of the usually negative LPS to zero, reducing the binding of cationic polymyxins (69, 80, 81). They also possess intrinsic resistance to colistin, tigecycline, and tetracycline (82). Sequencing of P. mirabilis strain HI4320 demonstrated the presence of genes for a conjugal transfer pilus, which allows the horizontal genetic transfer of plasmids encoding antibiotic resistance (35).
While most species of Proteus remain sensitive to a range of antibiotics, increasing rates of acquired antibiotic resistance in the Enterobacteriaceae are a growing problem (83). In 2007 to 2008, observations of P. mirabilis strains acquiring Salmonella genomic island 1 (SGI1) were reported by groups in China and Palestine (84). SGI1 is a mobile genomic element first identified in Salmonella Typhimurium that integrates into the recipient chromosome and carries multiple genes encoding resistance to streptomycin, trimethoprim, tetracycline, sulfonamides, chloramphenicol, fluoroquinolones, and a broad spectrum of β-lactam antibiotics (84). A further SGI1-positive P. mirabilis strain (NKU) has also been identified in Europe (84). Recently, an SGI1-positive P. mirabilis strain acquired a plasmid containing the New Delhi metallo-β-lactamase 1 gene, leading to the identification of extensively drug-resistant (XDR) P. mirabilis strain PM58, which is resistant to all antibiotics used for Enterobacteriaceae except aztreonam (82). A recently sequenced P. mirabilis isolate (NO-051/03) from a patient with a soft tissue infection in Europe had acquired the resistance genes for trimethoprim, β-lactams, phenicols, sulfonamides, and aminoglycosides (85).
In gastrointestinal disease, the antibiotic sensitivity profile of Proteus species is relevant to pathogenicity under conditions that may be exacerbated by antibiotic perturbation of the gut microbiome. Given the potential for a “bloom” of Enterobacteriaceae (including Proteus species) both in the presence of an inflammatory process (86) and with a perturbation of the enteric environment via surgery (87), the use of antibiotics should be considered a possible potentiating factor.
Vaccine CandidatesA full review of the treatment of Proteus infection is outside the scope of this review.
Purified MR/P fimbrial proteins have been tested for their antigenic potential as vaccine candidates (88, 89). There has been success by using the intranasal delivery of MrpH, the fimbrial tip adhesin of P. mirabilis (88). A fusion protein comprised of MrpH and mannose-binding protein delivered intranasally provided 75% protection from P. mirabilis ascending urinary tract infection in a mouse model (88). A fusion protein comprised of MrpH from P. mirabilis and FimH from uropathogenic E. coli delivered intranasally with monophosphoryl lipid A (as an adjuvant) induced robust IgG and IgA responses in mice (89). A clinical study of an inactivated bacterial cell suspension of four bacterial species, including a strain of P. vulgaris, was trialed in 159 patients with a history of recurrent urinary tract infections, compared with 160 patients maintained on prophylactic sulfamethoxazole-trimethoprim (90). The group of patients who received the vaccine had a mean number of 0.36 urinary tract infections in 3 months, versus 1.60 for those receiving sulfamethoxazole-trimethoprim (P < 0.0001) (90).
PROTEUS SPECIES AS GASTROINTESTINAL PATHOGENS
Colonization by Proteus SpeciesProteus species are known human digestive tract commensal organisms with abundances varying according to location (Fig. 3). Colonization occurs early. Infants from Sweden and Pakistan were assessed for Enterobacteriaceae based on mode of delivery (vaginal versus cesarean) and breastfeeding behavior (91). Cesarean births in Pakistan were associated with Proteus species colonization within 3 days, with 11 of 21 cesarean-delivered and 1 of 9 vaginally delivered infants being positive for Proteus spp. (P = 0.049) (91).
Zilberstein et al. cultured mucosal samples from the upper (n = 20) and lower (n = 24) digestive tracts of healthy controls. Proteus species were present in 8% of gastric samples, 46% of duodenal and jejunal samples, 19% of ileal samples, 13% of cecal samples, and 38% of samples from the transverse colon (12).
Müller compared the recovery of Proteus species from the stool specimens of 1,422 healthy subjects. P. mirabilis was identified in 2.7% of healthy subjects, which is a probable underestimate as the epithelial preference of P. mirabilis means that it is likely undersampled in stool specimens (92). Proteus penneri and P. vulgaris were isolated from 0.9% and 4.2%, respectively, of the same population (92). A smaller culture-based study of 60 patients with gastrointestinal symptoms who tested negative for parasites demonstrated a colonization rate of 33% (20/60 patients) for P. vulgaris (93).
Proteus species, especially P. mirabilis, are often antibiotic resistant, conferring a survival advantage when colonizing the gastrointestinal tract. In a study of multiple-drug-resistant Gram-negative bacteria in the rectum of long-term-care patients, 52 drug-resistant strains were identified, 15 of which were P. mirabilis strains (94). A total of 87% of patients colonized with P. mirabilis were also cocolonized with at least one other resistant Gram-negative bacterium (range, 1 to 4 species; median, 2 species). Of the 15 patients with a resistant P. mirabilis strain, only 1 patient spontaneously cleared the organism, compared to 30 to 75% clearance for other bacterial species (P = 0.007 for clearance of other species versus P. mirabilis, as determined by a log rank test), demonstrating the ability of P. mirabilis to cause more-persistent colonization than other Gram-negative species (94). Of the 15 P. mirabilis strains recovered, 13 were genetically distinct, demonstrating the heterogeneity of P. mirabilis populations between patients within the same health care facility (94).
Proteus spp. in GastroenteritisProteus species have been associated with infectious gastroenteritis. In a comparison of 1,271 patients with diarrhea, P. mirabilis was more prevalent in patients with diarrhea (10.8% in affected cases versus 2.7% in healthy subjects; P < 0.001) (92). However, that study did not take into account the possible administration of antibiotics to affected patients or the potential bystander or overgrowth effect (92). P. mirabilis has also been associated with foodborne gastroenteritis in an outbreak in Beijing, China, associated with the consumption of stewed pork (95). Shi et al. investigated genetic adaptation to the digestive tract of P. mirabilis species isolated from the vomit and feces of patients, obtained at the time of a foodborne outbreak (96). When three clinical isolates were compared to four local and reference strains of P. mirabilis, obtained from food, a healthy subject, and two patients with urinary tract infections (including reference strains HI4320 and BB2000), there was evidence of strain-level genetic adaptation to the digestive tract (96). All seven isolates harbored drug resistance genes, but only the three isolates (one from vomit, one from stool, and one from infected food) contained digestive tract toxicity genes, including one with a complete type 4 secretion system (T4SS) not identified previously in P. mirabilis. Active horizontal gene acquisition has been demonstrated for Proteus mirabilis, including a protein from Yersinia enterocolitica, a known GI pathogen and member of the Enterobacteriaceae family (96). In summary, Proteus species can be linked to diarrheal states, but their primary pathogenic role has not been confirmed.
Proteus spp. in the Upper Gastrointestinal TractColonization of the upper gastrointestinal tract, including the esophagus and stomach, by Proteus species in infants and older adults has been reported, often associated with instrumentation of the oropharynx (97–101). In 13 infants with feeding tubes without gastrointestinal symptoms, Proteus was isolated from the throat in 8% of patients, from the gastric juice in 15% of patients, and from the duodenal fluid in 8% of patients (98). In elderly patients with nasogastric feeding tubes (NGTs), Proteus species were isolated from the oropharynx in 24% of patients and from gastric fluids in 26% of patients (100). In another study, colonization of the oropharynx with Proteus species was present in 13% of patients with percutaneous endoscopic gastrostomy (PEG) and 21% of patients with NGTs, and colonization of the stomach was present in 4% and 23% of the same patients, respectively (99).
Association with Hepatobiliary DiseaseEarly culture-based surveys of patients undergoing biliary surgery showed the occasional isolation of Proteus species from the biliary tract (13% of bile samples) (102). P. mirabilis has also been recovered from bile obtained during endoscopic retrograde cholangiopancreatography (ERCP) in 6% of cultured samples (103).
In a metagenomic analysis using multitag pyrosequencing, the rectosigmoid mucosal community of healthy individuals was compared with that from patients with cirrhosis. Patients with cirrhosis had an elevated proportion of Proteus species compared with controls; the relative abundance in healthy controls was 0.0%, versus 0.1% in patients with cirrhosis (P < 0.00001) (104). Urease-producing microbes, such as Proteus spp., in the gut are known to contribute to the pathogenesis of hepatic encephalopathy through the breakdown of urea to ammonia and carbonic acid (104, 105).
In a series of patients undergoing liver resection, Proteus vulgaris bacteremia was identified in two patients, and polymicrobial infections were identified in eight patients (106). In the hepatobiliary tract, Proteus spp. are an uncommon cause of infection and are usually related to surgical interventions, such as ERCP or abdominal surgery. Implantable devices, such as stents, are also at risk of colonization and biofilm formation (107).
Pancreatic DiseaseThere are isolated reports of Proteus species infections of the pancreas, including a patient with a large infected pancreatic pseudocyst compressing the common bile duct. The cystic contents were polymicrobial, including Proteus vulgaris, Morganella morganii, Stenotrophomonas maltophilia, and Pseudomonas aeruginosa (108).
Proteus species were found in biofilms that form on biliary and pancreatic stents placed via ERCP in 14 of 100 stents (107).
Intestinal DiseaseThere are a number of reported links between Proteus spp. and various intestinal conditions, including small bowel intestinal overgrowth, Crohn's disease, and ulcerative colitis.
A small study of the downstream effects of small bowel ulceration caused by nonsteroidal anti-inflammatory drugs in rats identified a mixed population of E. coli and P. mirabilis. When treated with metronidazole, rats were protected from ulcer development (109). Proteus species were recovered from jejunal fluid in 11% of patients with small intestinal bacterial overgrowth syndrome (SIBOS), which often follows the expansion of facultative anaerobic bacterial communities (110). Viable bacterial translocation of Proteus mirabilis across the intact intestinal barrier has been demonstrated from the cecal and colonic mucosa of a monoassociated mouse model, with bacteria being additionally isolated from the mesenteric lymph nodes and the liver (111).
Crohn's disease.Recent research has implicated Proteus spp. in inflammatory bowel diseases, with evidence being derived both from patient-based microbiome surveys and mechanistic research. Ambrose et al. used culture-based techniques to compare the recovery of pathogenic gut bacteria from the ileal serosa and mesenteric lymph nodes in 45 Crohn's disease patients and 43 patients having surgery for other indications (112). Overall, Crohn's disease cases were more likely to have pathogenic bacteria recovered from the small bowel serosa (12/45 [27%] patients versus 6/41 [15%] controls). Of the 12 patients with positive serosal cultures, 4/12 patients were positive for Proteus spp. (33%). Additionally, involved and uninvolved mesenteric lymph nodes were assessed: 15/45 patients (33%) with involved nodes had a positive culture, and of these samples, 1/15 (7%) grew Proteus spp. Eleven of 45 samples of uninvolved nodes harbored bacteria, with 1/11 (9%) being positive for Proteus species (112). Although these data are not significant due to small numbers, they demonstrate the recovery of Proteus species from patients with Crohn's disease, while the recovery of Proteus species from lymph nodes establishes bacterial translocation, in line with the other members of the Enterobacteriaceae family (113).
Two microbiome studies have linked an overabundance of Enterobacteriaceae and Proteus spp. to Crohn's disease. In a pediatric study, Proteus species comprised 3/18 (16.7%) Gram-negative bacterial strains recovered from 12 Crohn's disease patients, compared to the total absence of Proteus species recovered from patients with ulcerative colitis, indeterminate colitis, and lymphonodular hyperplasia and from controls (114). A microarray-based study comparing patients with active Crohn's disease with matched healthy controls identified an overrepresentation of Proteus species overall, and Proteus vulgaris in particular, and other members of the Proteobacteria phylum in patients with active ileal disease requiring surgery (115).
A number of studies have addressed the mechanisms by which Proteus and other Enterobacteriaceae may contribute to the development of inflammatory bowel diseases. The TRUC mouse model of ulcerative colitis has been used to demonstrate that Proteus mirabilis and Klebsiella pneumoniae can elicit colitis and that this propensity for the development of colitis can be transmitted to wild-type mice via microbiome transfer (65).
Crohn's disease has been associated with changes in nitrogen metabolism. Ammonia produced from the breakdown of urea by bacterial urease provides a source of nitrogen for respiration and amino acid synthesis by pathogenic facultative anaerobes from the Proteobacteria phylum (49).
Interactions between the Enterobacteriaceae, which include Proteus species, and fungi (Candida tropicalis) have recently been implicated in the dysbiosis that characterizes Crohn's disease (116). When Crohn's disease patients were compared with first-degree relatives of Crohn's disease patients and healthy controls, bacterial dysbiosis was identified in affected patients and unaffected first-degree relatives. Crohn's disease patients had an increased presence of Candida tropicalis. Complex interactions between Enterobacteriaceae species (Serratia marcescens and E. coli) and Candida tropicalis were confirmed in laboratory studies, showing that flagellated and/or fimbriated bacteria combined with fungal hyphae to form a robust biofilm, with these three species combined forming the thickest biofilm (P < 0.0001). It was postulated that biofilms enriched for immunomodulatory microbial components (lipopolysaccharides and oligomannans, etc.) may perpetuate inflammation in dysbiotic patients through the induction of proinflammatory cytokine responses and of apoptosis. Proteus spp. were also strongly positively correlated with the abundance of Candida in patients with familial Crohn's disease (r = 0.709; P < 0.005), raising the possibility that Proteus spp. are capable of the same interactions, although this was not demonstrated, possibly due to their low abundance (116).
Seo et al. demonstrated that in the presence of intestinal injury and colonization with P. mirabilis, a marked proinflammatory IL-1β response occurs, via the activation of the NOD-like receptor protein 3 (NLRP3) inflammasome (59). This occurs only in the presence of the P. mirabilis hemolysin HpmA, which appears to induce host macrophage induction of NLRP3. Preexisting injury or inflammation was necessary for the induction of NLRP3 activity and IL-1β by P. mirabilis (dextran sodium sulfate [DSS]-induced colitis), as inflammatory monocytes were required, suggesting that P. mirabilis may act to perpetuate and accelerate preexisting inflammation rather than induce it. Proteus mirabilis was as efficient at inducing IL-1β as pathogenic Salmonella spp., but the presence of the HpmA hemolysin was essential for the induction of IL-1β. IL-1β has been shown to be associated with disease activity in IBD patients (117, 118); however, in that study, most other anaerobic or facultative anaerobic commensals induced TNF-α expression but not IL-1β (59). When hemolysin expression levels were compared across P. mirabilis strains from multiple clinical sources (pyelonephritis, catheter-associated, and fecal isolates), fecal isolates had the highest hemolytic activity and had significantly higher hemolytic titers than those of the catheter-associated strain (P = 0.001) but not those of the pyelonephritis strain (P = 0.065) (119).
Proteus species have been associated with the postoperative recurrence of Crohn's disease by two independent groups (2, 3). Metagenomic surveys of patients at the time of surgery and 6 and 18 months postoperatively demonstrated that patients were more likely to have disease recurrence in the presence of detectable Proteus genera (P = 0.008) and the absence of detectable Faecalibacterium (P < 0.001) (3). The combination of detectable Proteus species and absent Faecalibacterium (<0.1%) in postoperative ileal biopsy specimens was associated with an increased risk of recurrence, with an odds ratio (OR) of 14 (95% confidence interval [CI], 1.7 to 110; P = 0.013). Smoking, an independent risk factor for postoperative disease recurrence, was also associated with an increased presence of Proteus spp. (P = 0.0130) (3). In another study by Mondot et al. of 20 Crohn's disease patients undergoing ileocolonic resection, the presence of a Proteus mirabilis operational taxonomic unit (OTU) was predictive of recurrence at 6 months postoperatively (2).
A recent review of consecutive Crohn's disease patients with intra-abdominal abscesses as a result of active disease demonstrated infection with Proteus spp. in 4.8% of cases, which was associated with high rates of quinolone resistance (120). Whether the presence of Proteus spp. in association with postoperative recurrence is a primary pathogenic event or secondary to disease recurrence remains to be elucidated. However, the association in both studies was established prospectively and longitudinally, with predictive association, making a pathogenic role more likely.
Other large intestinal diseases.Kanareykina et al. obtained samples (mouth, stomach, small intestine, and feces) from 65 patients with ulcerative colitis; performed culture-based enumeration; and identified Proteus mirabilis, Proteus vulgaris, or the closely related species Morganella morganii or Providencia rettgeri (Fig. 1) in nearly all cases (121). In 40/65 patients, these species were recovered from more than one anatomic site. A Proteus species protein “vaccine” was then administered and resulted in clinical improvement in moderate to severe cases of ulcerative colitis as well as a decrease in bacterial counts. However, no details of the vaccine composition or any objective disease activity metrics were described, and the study overall was of low quality (121).
A recent study of children with and without appendicitis showed increases in the relative abundances of 12 genera, of which Proteus species were the only representatives of the Enterobacteriaceae family (0.015% versus 0%; P = 0.028) (122).
Proteus bacteria have been implicated in the perpetuation of colonic inflammation in diversion colitis (123). Inflammatory conditions of the bowel increase the local concentrations of inducible nitric oxide synthase, leading to high levels of nitrate that cannot be metabolized, except by the microbiota (124). This favors the expansion of bacteria that are able to metabolize nitrate under anaerobic conditions, leading to a survival advantage and population expansion of Proteus spp. and other nitrate-reducing Enterobacteriaceae (48, 123).
Nosocomial Infections and Proteus Species Complicating Gastrointestinal DiseaseProteus species, especially P. mirabilis and P. vulgaris, are common causes of nosocomial opportunistic infections. Many patients with preexisting gastrointestinal diseases are liable to secondary Proteus infections, often in the context of polymicrobial infections. Proteus species can also cause peritonitis following perforations of the gastrointestinal tract; in one report of 383 patients with peritonitis, Proteus species were identified in 87 (23%) patients (125).
Proteus species can colonize medical devices placed in the gastrointestinal tract, including ventriculoperitoneal shunts (126), nasogastric tubes (99, 100), biliary and pancreatic stents (107), and tracheoesophageal voice prostheses (127). Proteus bacteria have been shown to be contaminants of gastroscopes and colonoscopes after insufficient disinfection (128). Infections can also be acquired in the hospital setting due to environmental contamination, with P. vulgaris persisting on dry, hard surfaces for up to 2 days (129). There are reports of hospital-based and community epidemics of infection with person-to-person spread, with most patients acquiring gastrointestinal carriage prior to infection (13, 130).
CONCLUSIONS
Proteus species are hardy, adaptable, and potentially pathogenic residents of the human gastrointestinal tract and have been underappreciated as a cause of gastrointestinal disease. Host-microbe and microbe-microbe interactions by Proteus spp., and the pathogenicity of this genus that may result from population expansion in response to environmental changes, are emerging as important aspects of disease associated with this genus. The possible contribution of Proteus spp. to intestinal diseases and infections has been somewhat neglected. Research into the virulence of Proteus spp. in the urinary tract using the bacteriology of ileal conduits (131) and intestinal segments (60, 132) for bladder augmentation suggests that Proteus spp. should be examined more closely for their potential as gastrointestinal pathogens.
There is increasing evidence that Proteus species may play a role in inflammatory bowel disease through the direct action of the bacteria, compounded by host immune evasion and perturbation. As Gram-negative organisms, Proteus species are intrinsically proinflammatory as a result of the production of lipopolysaccharide (LPS) and immunostimulatory flagellin proteins. There may be an association between Proteus species and inflammatory bowel disease, especially Crohn's disease, mainly through population expansion and immune activation. Their low population abundance does not preclude a potential large pathogenic effect.
Genetic characterization of enteric isolates compared to urinary tract isolates will be important for determining the effect of virulence factors on gastrointestinal pathogenesis.
Research on the gut microbiome as an ecosystem is informing our understanding of Proteus species, yet there are still unanswered questions. These include obtaining confirmation that Proteus species can swarm within the human gut and addressing the effect of individual environmental changes (e.g., surgery, pH, or oxygen concentrations) on the mucosa-associated Proteus population.
ACKNOWLEDGMENTS
A.L.H. was supported via a Dora Lush postgraduate scholarship from the National Health and Medical Research Council (NHMRC). St Vincent's Research Endowment Fund and the Australasian Gastro Intestinal Research Foundation supported M.A.K. The University of Queensland Diamantina Institute and The University of Queensland supported M.M. The Leona M. and Harry B. Helmsley Charitable Trust supported this work.
We have no conflicts of interest.
M.A.K. and A.L.H. devised the concept. A.L.H. acquired data, screened papers, interpreted the data, and wrote the manuscript. M.A.K., S.C.N., and M.M. provided critical revision of the manuscript for important intellectual content.
- Copyright © 2018 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.↵
- 106.↵
- 107.↵
- 108.↵
- 109.↵
- 110.↵
- 111.↵
- 112.↵
- 113.↵
- 114.↵
- 115.↵
- 116.↵
- 117.↵
- 118.↵
- 119.↵
- 120.↵
- 121.↵
- 122.↵
- 123.↵
- 124.↵
- 125.↵
- 126.↵
- 127.↵
- 128.↵
- 129.↵
- 130.↵
- 131.↵
- 132.↵
- 133.↵
- 134.↵
- 135.↵
- 136.
- 137.
- 138.
- 139.
- 140.
- 141.
- 142.
- 143.
- 144.
- 145.
- 146.
- 147.
- 148.
- 149.
- 150.
Author Bios
Amy L. Hamilton is a Clinical Scientist from St Vincent's Hospital and the University of Melbourne. She graduated from Monash University, Melbourne, Australia, with a bachelor of science, with Honors in Microbiology and Immunology. She has worked on a number of investigator-initiated clinical trials in gastroenterology and is passionate about translational research in inflammatory bowel disease as a way of improving long-term outcomes for patients. She has a special interest in the gut microbiome and its role in the etiology of inflammatory bowel diseases, with this research focus allowing her to utilize both basic science and translational skills. She has completed her Ph.D. in the Department of Medicine at the University of Melbourne, addressing the microbiology of Crohn's disease after surgical resection. She has been a member of Professor Michael A. Kamm's research group since 2010.
Michael A. Kamm is Gastroenterologist at St Vincent's Hospital Melbourne and Professorial Fellow at the University of Melbourne, Melbourne, Australia, since 2008. He was previously Chairman of Medicine and Director of the Physiology and Inflammatory Bowel Disease Units at St Mark's Hospital in London, United Kingdom, from 1989 to 2008 and Professor of Gastroenterology at Imperial College London. He treats and researches in the fields of inflammatory bowel disease (IBD) and functional bowel disorders. Professor Kamm has conducted multidisciplinary research, spanning basic science, translational, and clinical aspects of gut disorders. In the field of the gut microbiota and IBD, he has studied mechanisms of inflammation, including recognition of bacterial antigens by dendritic cells and the immunological and clinical therapeutic effects of putative prebiotics and probiotics. Most recently, he conducted investigator-initiated studies into the perioperative management of Crohn's disease (“POCER” study) and fecal microbiota transplantation in ulcerative colitis (“FOCUS” study), both changing outcomes for patients and identifying specific bacteria associated with clinical outcomes.
Siew C. Ng is Professor at the Department of Medicine and Therapeutics, The Chinese University of Hong Kong. She received her medical degree from St Bartholomew's and Royal London School of Medicine and obtained her Ph.D. from Imperial College London. Dr. Ng's main research interests include inflammatory bowel disease, gut microbiota, and colorectal cancer, aiming to further our understanding of the pathogenesis of inflammatory intestinal disorders via clinical trials and translational research. She is President of the Hong Kong IBD Society, Scientific Secretary of the International Organization of IBD, and member of the Management Committee of the International IBD Genetics Consortium. She has published over 150 peer-reviewed papers in international journals, including Nature Genetics, Nature Communications, Lancet, Gastroenterology, and Gut. She is a pioneer of IBD Epidemiologic and Microbiome research in Asia-Pacific. Her work has received over 20 Prestigious National and International Awards.
Mark Morrison is the Chair in Microbial Biology and Metagenomics at The University of Queensland Diamantina Institute (UQDI), since October 2013. An authority in metagenomic and molecular microbiology, he is currently Australia's science representative to the International Human Microbiome Consortium. Professor Morrison obtained his Ph.D. from the University of Illinois, Urbana, IL. After 20 years in the United States, he returned to Australia in 2006 as CSIRO's first appointment via the Science Leader Scheme (leading the Gut Health Flagship Research Program). Since 2007, he has held a Professorship within the School of Chemistry and Molecular Biosciences, The University of Queensland. Professor Morrison aims to translate genomic and metagenomic data sets into sound biological frameworks, helping to produce novel diagnostic, organismal, and enzyme-based technologies. While Professor Morrison's focus was initially on microbial physiology and metabolism, he has since attained international acclaim for his efforts to successfully develop and use genomics and related methods to study anaerobic “commensal” gut bacteria.
