Enterococcus faecalis: an overlooked cell invader

SUMMARY

Enterococcus faecalis and Enterococcus faecium are human pathobionts that exhibit a dual lifestyle as commensal and pathogenic bacteria. The pathogenic lifestyle is associated with specific conditions involving host susceptibility and intestinal overgrowth or the use of a medical device. Although the virulence of E. faecium appears to benefit from its antimicrobial resistance, E. faecalis is recognized for its higher pathogenic potential. E. faecalis has long been considered a predominantly extracellular pathogen; it adheres to and is taken up by a wide range of mammalian cells, albeit with less efficiency than classical intracellular enteropathogens. Carbohydrate structures, rather than proteinaceous moieties, are likely to be primarily involved in the adhesion of E. faecalis to epithelial cells. Consistently, few adhesins have been implicated in the adhesion of E. faecalis to epithelial cells. On the host side, very little is known about cognate receptors, except for the role of glycosaminoglycans during macrophage infection. Several lines of evidence indicate that E. faecalis internalization may involve a zipper-like mechanism as well as a macropinocytosis pathway. Conversely, E. faecalis can use several strategies to prevent engulfment in phagocytes. However, the bacterial and host mechanisms underlying cell infection by E. faecalis are still in their infancy. The most recent striking finding is the existence of an intracellular lifestyle where E. faecalis can replicate within a variety of host cells. In this review, we summarize and discuss the current knowledge of E. faecalis–host cell interactions and argue on the need for further mechanistic studies to prevent or reduce infections.

INTRODUCTION

Enterococci are common colonizers of the human gastrointestinal and genitourinary tracts and have become a major cause of healthcare-associated infections (1, 2). They are among the most important nosocomial bacterial pathogens in North America and Europe (3). Enterococci are the major Gram-positive cause of urinary tract infections (UTI). They can cause intra-abdominal and pelvic infections, bacteremia, which accounts for approximately 10% of bacteremia worldwide, and endocarditis (4, 5). Moreover, the constant increase of antibiotic resistance poses a serious therapeutic challenge, particularly for Enterococcus faecium. The emergence of isolates resistant to both vancomycin (VRE) and aminoglycosides raises questions about the use of this combination as a standard therapy and whether new therapy guidelines are necessary (6, 7). Enterococcus faecalis and E. faecium account for the majority of human infection isolates, with E. faecalis responsible for the highest percentage of enterococcal infections and E. faecium representing a growing burden due to multiple mechanisms of antibiotic resistance (4, 8, 9). Recently, a role for E. faecalis in exacerbating liver injury has been demonstrated (10, 11). Furthermore, the presence of E. faecalis in tumor microbiota, its involvement in the promotion of liver carcinogenesis and in the migration of murine CT26 and human colon cancer HCT116 cells support its possible role in cancer (12–15).

How E. faecalis and E. faecium evolve from commensal colonizers to pathogens is a significant public health concern. They are resistant to bile salts, oxidative stress, and survive within macrophages, which facilitates their dissemination in patients. They exhibit high genomic diversity and genomic plasticity through recombination or horizontal gene transfer, favoring both diversity and adaptation. This diversity is responsible for phenotypic differences between isolates (16, 17). Entry into the bloodstream is a critical step in the development of enterococcal infections in immunocompromised and critically ill patients (18–20). The main portals of entry are the gastrointestinal tract after surgery or translocation through the intestinal mucosa, the urinary tract after catheterization or by the ascending route, and the bloodstream through the use of central lines (21). The intestinal epithelium is a polarized, semipermeable monolayer of epithelial enterocytes connected by tight junctions and interspersed with goblet cells, Paneth cells, M cells, and enteroendocrine cells. The increase in intestinal permeability induced by inflammatory or dysbiotic conditions is exploited by E. faecalis to enter the bloodstream (22, 23). In this process, the gelatinase GelE can exacerbate the disruption of the colonic barrier by degrading the junctional protein E-cadherin (24). In addition to the ability of enterococci to cross the intestinal barrier through breaches in the epithelial layer, E. faecalis is able to enter and cross the intestinal epithelial cells. In Wells’ pioneering work in 1990, bacteria resembling E. faecalis always appeared within the epithelial cells of the mouse gastrointestinal tract, never between them, indicating that the intestinal epithelial cells are a main portal of entry for E. faecalis (25). Since then, enterococcal transcytosis across intestinal epithelial cells has been supported by the ability of different strains of E. faecalis to cross a monolayer of polarized human intestinal epithelial T84 cells while preserving the integrity of the junctions without any apparent increase in permeability (26–29). The enterococcal polysaccharide antigen EPA, the gelatinase GelE, and the fsr genes, which encode an Agr-type quorum-sensing system that positively regulates the expression of gelatinase, are important for the passage of E. faecalis strain OG1RF through the intestinal epithelial cells (26, 29). Similarly, glycolipids and polyGlcNAc-containing exopolymers allow E. faecalis translocation across intestinal epithelial cells (27, 28). The relative contribution of paracellular and transcellular transport in the process remains to be clarified. On the other hand, E. faecalis can persist in macrophages, suggesting that macrophages may serve as a vehicle for E. faecalis to disseminate (30). Recent studies have shown the ability of some E. faecalis strains to survive and divide inside epithelial cells (31, 32).

In the light of the recent evidence that E. faecalis, generally recognized as an extracellular pathogen, has an intracellular lifestyle, it now appears appropriate to summarize these findings. Therefore, here, we review the current knowledge of how E. faecalis interacts with mammalian host cells during the infectious process. We summarize how E. faecalis strains adhere to, enter, and survive within host cells. This review also discusses the fate of E. faecalis once internalized in host cells and its strategy to hijack the host immune response (31–34). Despite the differentiation of commensal, clinical, and probiotic strains, there is no established correlation between the pathogenic potential or the presence of virulence genes and the origin of the strains (16, 17, 35). Consequently, various strains have been utilized to investigate the interaction between E. faecalis and the host cells. For the sake of clarity and with the exception of model strains, the specific names of the strains will not be subsequently identified. Table 1 recapitulates E. faecalis–cell interaction studies performed on common clinical strains. Figure 1 provides an overview of some of the aspects of E. faecalis transit in the host. For a comprehensive view of the effects of enterococcal infection on the host cells, including cytotoxicity, antimicrobial and inflammatory responses, programmed cell death pathways, stress and bystander effects, the interested reader is referred to other reviews and reports (36–40).

TABLE 1.

E. faecalis–cell interaction studies conducted on common strains

Strain	Description	Cellular models
12030	Clinical strain originally isolated in Cleveland, USA (41)	Adhesion to intestinal epithelial Caco2 cells, epithelial Hep-2 cells (HeLa derivative), and bladder epithelial T24 cells (42–46); opsonization and survival in PBMCs and in PMNs (41, 45, 47–49).
ATCC29212	Clinical strain isolated from a human urine sample collected in Portland, USA (50)	Survival dendritic cells from murine bone marrow-derived stem cells (51).
FA2-2	Laboratory strain, Rif/Fus resistant mutant derived from plasmid-free strain JH2 (52)	Adhesion to intestinal epithelial Caco2 cells and bladder epithelial T24 cells (42, 53); survival in RAW264.7 macrophages and JAWS II dendritic cells (54).
E99	Clinical strain isolated from the urine of a patient at the Veterans Administration Hospital in Arkansas, USA (55).	Survival in RAW264.7 macrophages and JAWS II dendritic cells (54, 56–59).
JH2-2	Rif/Fus-resistant mutant from the non-hemolytic clinical JH2 strain (60)	Transcytosis to intestinal epithelial T84 cells (26, 29); survival in zebrafish macrophages and in peritoneal macrophages, according the in vivo–in vitro macrophage infection model of Gentry-Weeks et al. (30, 61–64).
MMH594	Epidemic strain, isolated from the blood of a patient with bacteremia in Wisconsin, USA (65)	Survival in RAW264.7 macrophages (56)
OG1	Human oral cavity isolate, subsequently shown to cause dental caries in rats (66)	Adhesion to intestinal epithelial Caco2 cells (42)
OG1RF	Laboratory strain (ATCC47077), Rif/Fus-resistant mutant of OG1 (67)	Adhesion to vaginal epithelial VK2 cells, and endocervical epithelial End1 cells (68); adhesion and translocation in colonic epithelial Ptk6 cells (69); internalization in placental epithelial Jeg-3 cells, hepatic HepG2 cells, cervical epithelial HeLa cells, intestinal epithelial Caco-2, HT-29, and HCT-8 cells, kidney epithelial A-704 cells, fibroblastic ACHN cells, and HaCaT keratinocytes (32, 70, 71); transcytosis in intestinal epithelial T84 cells (26, 29), intracellular multiplication in Huh7 hepatocytes, HaCaT keratinocytes, and RAW264.7 macrophages (31, 32); opsonization, evasion of phagocytosis and survival in PMNs, BMDMs, J774, and RAW264.7 macrophages, and zebrafish macrophages (32, 34, 56, 61, 70, 72–77).
OG1X	A gelatinase-defective mutant originally derived by mutagenesis of OG110, a streptomycin-resistant derivative of OG1 (78)	Adhesion to bladder epithelial T24 cells (53); survival in PMNs and PBMCs (77, 79).
V583	Vancomycin-resistant clinical strain (ATCC 700802), from the hospital in St Louis, USA (80)	Adhesion to intestinal epithelial Caco2 cells, epithelial Hep-2 cells (HeLa derivative), vaginal epithelial VK2 cells, and endocervical epithelial End1 cells (42, 45, 68); adhesion and translocation in colonic epithelial Ptk6 cells (69); autophagy in epithelial cells of the mouse small intestine (81); opsonization and survival in PMNs, THP-1 monocytes, HBDMs, and BMDMs, RAW264.7 macrophages, and zebrafish macrophages (61, 72, 82–84).

Fig 1.

Overview of the E. faecalis transit in the host. Extracellularly, high levels of E. faecalis (in blue) increase the risk of translocation from the intestinal lumen into tissues and dissemination. In some individuals, gastrointestinal leakage provides a route for bacteria to cross the epithelial barriers. In addition, gelatinase E and cytolysin produced by E. faecalis, for example, can mediate epithelial permeability by acting on cell junctions or by lysing the host cells. E. faecalis can also directly penetrate epithelial monolayers and transit cytosolically or within intracellular compartments. Proteins, such as AS, assist in adhesion to host cells and likely mediate uptake. When inside an intracellular compartment, E. faecalis is resistant to bactericidal lysosomal proteins, such as the lysozyme, and affects the levels of proteins important for lysosomal fusion, such as Rab5 and Rab7. In addition, E. faecalis can prevent the activation of NF-κB and modulate the host response to infection to promote its survival.

E. FAECALIS PHAGOCYTOSIS: RECOGNITION AND SUBVERSION

Long considered extracellular and the exclusive target of phagocytic cells, mechanistic studies of E. faecalis–cell interactions have focused primarily on macrophages and neutrophils. Phagocytes express several classes of receptors, including opsonic receptors, scavenger receptors, C-type lectin receptors, and pattern recognition receptors. Complement and antibody-mediated opsonization are involved in the recognition of E. faecalis (82, 85). The absence of wall teichoic acid (WTA), corresponding to enterococcal polysaccharide antigen (EPA) decorations, was correlated with an increased binding of mannose-binding lectin and a subsequent complement deposition (82, 86). This suggests that WTAs are involved in complement evasion. In the case of antibody-mediated opsonization, Rossmann and collaborators showed that pentaglobin (IgG, IgM, and IgA mixture) has a higher opsonic killing activity against the encapsulated E. faecalis strain 12030 than IgGs alone (47). Antibodies directed against cell wall carbohydrates are probably responsible for this protection (47). However, in the case of complement-mediated opsonization, Thurlow and collaborators showed that encapsulated and unencapsulated strains had similar levels of C3 complement deposition but encapsulated strains could avoid recognition by murine RAW264.7 macrophages (72). These observations indicate that the capsule masks bound C3b to prevent phagocytosis. Aggregation substances are adhesins involved in bacterial–bacterial contact or bacterial clumping occurring during plasmid exchange between enterococci (87). In E. faecalis, the two most studied aggregation substances are Asa1 (also called AS) and Asc10, which are encoded by the pheromone-inducible plasmids pAD1 and pCF10, respectively. The aggregation substance Asc10 is involved in the adhesion of E. faecalis to human neutrophils (88) and human macrophages (79). In addition, adhesins containing D-glucose residues mediate an interaction between clinical strains E. faecalis and neutrophils (89), but the neutrophil receptor involved in D-glucose binding has not been identified.

Antagonists of phagocytic receptors have been used to characterize the receptors involved in E. faecalis engulfment by macrophages (90, 91). Although specific blockade of the CD206 receptor had no effect on E. faecalis phagocytosis, blocking mannose receptors in THP-1 cells with multiple antagonists, such as mannose, mannose-6-P, laminarin, and mannan before E. faecalis infection inhibited phagocytosis by approximately 50% (90). These results suggest that mannose receptors play a role in E. faecalis recognition and internalization by macrophages (90). The role of mannose receptors was further confirmed in murine peritoneal macrophages, bone marrow-derived macrophages, and dendritic cells (91). Blockade of the dectin-1 receptor did not abolish phagocytosis, indicating that engulfment is either dectin-1 independent or that other receptors compensate for dectin-1 blockade (91). Inhibition of E. faecalis adherence to THP-1 macrophages by heparin and heparan sulfate suggested that glycosaminoglycans are host receptors for E. faecalis (90). In this study, E. faecalis was surrounded by small ruffles in macrophages 24 h post-infection (90). As described for other Gram-positive bacteria, lipoteichoic acid (LTA) from E. faecalis is a major ligand for binding to scavenger receptors (92). Consistent with this, overexpression of the scavenger receptor CD36 in kidney epithelial HEK293 cells and in cervical epithelial HeLa cells increases the uptake of E. faecalis (93). In the fruit fly Drosophila melanogaster, the EGF-like repeat containing scavenger receptor Eater expressed by S2 phagocytes mediates E. faecalis phagocytosis (94). Phagocytosis of E. faecalis by immune cells requires actin polymerization and microtubules, as the use of an inhibitor of actin polymerization (cytochalasin D) and microtubule elongation (colchicine) inhibited E. faecalis uptake by THP-1 macrophages (90).

Despite this array of molecular interactions involved in E. faecalis phagocytosis, E. faecalis can also evade macrophage and human neutrophil-mediated phagocytosis (56, 61, 73, 95). E. faecalis OG1RF does not induce NETosis (neutrophil extracellular traps), and it can suppress S. aureus-induced NETosis by attenuating histone citrullination in polymicrobial infections, suggesting that lack of NETosis induction is an active process (96). The capsule confers resistance to complement-mediated opsonization and neutrophil killing (48, 72, 97). ElrA, the enterococcal leucin-rich protein A, is implicated in E. faecalis OG1RF evasion of phagocytosis. Strains expressing ElrA have a reduced adhesion to murine RAW264.7 macrophages (74). Indeed, ElrA acts as an antiphagocytic cloak that allows E. faecalis to evade host detection (70). The gelatinase GelE allows bacteria to evade the immune response by degrading the C3 components and reducing phagocytosis by neutrophils (98). More recently, Ali et al. showed that E. faecalis recruits the complement factor H in the blood to escape the alternative pathway of the complement, one of the three pathways of complement activation (99). In conclusion, E. faecalis can use the capsule and cell wall surface proteins, secrete bacterial effectors, and degrade or sequester host effectors to prevent engulfment by phagocytic cells.

E. FAECALIS ADHERES TO AND INVADES A VARIETY OF MAMMALIAN CELLS

The first descriptive studies of enterococcal adhesion and internalization by host cells other than phagocytes was mainly performed with E. faecalis using electronic microscopy or counting adherent or internalized bacteria after short interaction periods ranging from 30 min to several hours (53, 100–102). Most of these studies were conducted to (i) determine whether enterococcal isolates exhibit a cell type tropism that correlates with their site of isolation and (ii) correlate specific virulence factors with adhesion, rather than to decipher the mechanisms that promote intracellular uptake of E. faecalis. Compared with traditional enteropathogens, the internalization of E. faecalis into mammalian cells is approximately 3-logs lower than that of Listeria monocytogenes or Salmonella typhimurium into enterocytes (103). Adhesion and internalization assays showed that E. faecalis adheres to the microvilli of intestinal epithelial HT-29 and Caco-2 cells and that bacteria were found intracellularly after 1 h of contact (101–103). More recently, a comparison of E. faecalis OG1RF internalization in seven human cell lines, all derived from tissues potentially targeted by E. faecalis during human infection, showed that the percentage of internalized E. faecalis compared with the initial inoculum varied by cell type (70). The placental epithelial Jeg-3 cells exhibited the higher percentage (0.4%) of intracellular bacteria. By contrast, E. faecalis infection appeared 10-fold lower in hepatic HepG2 cells and in cervical epithelial HeLa cells than in renal fibroblastic ACHN cells, intestinal epithelial Caco-2 cells, ileocecal epithelial HCT-8 cells, or renal epithelial A-704 cells (70).

The difference in uptake efficiency may be related to the ability of the strain to adhere to the host cell surface. For instance, clinical E. faecalis isolates from vaginal swabs showed higher adhesion to renal epithelial Vero cells than to intestinal epithelial Caco-2 cells (104). E. faecalis OG1RF and V583 also adhere to human vaginal epithelial VK2 cells and human endocervical epithelial End1 cells (68). Invasion in the human urothelium has been documented in human bladder T24 cells, urothelial R24 cells, and shed urothelial cells from patients’ bladders and leads to the formation of intracellular colonies (105–108). E. faecalis strains isolated from UTI have a greater ability to adhere to human urinary tract epithelial cells and embryonic kidney cells than strains isolated from endocarditis (109). By contrast, E. faecalis strains isolated from endocarditis preferentially adhere to heart Girardi cells (109), suggesting the existence of a cellular tropism of enterococcal isolates according to their tissue of origin. This hypothesis was supported by a study that identified five highly adherent E. faecalis strains among 30 strains isolated from the urine of UTI patients to bladder T24 cells, whereas strains from fecal isolates (n = 30) of healthy volunteers did not adhere (53). However, this cellular tropism was challenged by Archimbaud et al. who found that E. faecalis isolates from endocarditis adhered more efficiently to intestinal Int-407 cells than to Girardi heart cells (110). This discrepancy may be related to the fact that the Girardi cell line used in these two studies was found to be genetically indistinguishable from the cervical epithelial HeLa cells by short tandem repeat (STR)-PCR profiling (111).

Host environmental conditions can induce the expression of adhesion factors, possibly allowing E. faecalis to differentially colonize different niches. For example, E. faecalis isolates from UTI that were grown in human serum before infection adhered more strongly to Girardi cells than E. faecalis not grown in serum (109). E. faecalis can be a cause of bovine mastitis. Similarly, the growth of subclinical mastitis E. faecalis in milk increases adherence and subsequent entry into bovine mammary epithelial MAC-T cells (112, 113). Interestingly, this increase correlated with the biofilm-forming ability of the strains, which also increased when grown in bovine milk. This suggests that E. faecalis aggregates can be internalized into epithelial cells, as has been observed for both macrophages and dendritic cells that phagocytize biofilm cells (54, 90).

In summary, E. faecalis is capable of adhering to and being internalized by fibroblasts, endothelial cells, and epithelial cells, including intestinal, renal, and bladder cells, as well as hepatocytes. To date, it remains difficult to establish a correlation between the strain origin and its propensity to infect cells (69, 110, 114).

HOW DOES E. FAECALIS ADHERE TO AND ENTER EPITHELIAL CELLS?

Surface proteins and glycopolymers of the bacterial envelope play important roles in the interactions of Gram-positive bacteria with the host (115–117). The cell wall of E. faecalis contains LPxTG-type surface proteins, including pili, as well as surface proteins containing a WxL domain (87, 118). Enterococcal glycopolymers include lipoteichoic acid (LTA), wall teichoic acid (WTA), enterococcal polysaccharide antigen (EPA), and capsular polysaccharide (86, 87, 116).

Cell-wall polysaccharides are important mediators of cell adhesion

The initial adhesion of E. faecalis to epithelial cells has been correlated with bacterial surface saccharide residues. Guzmàn et al. showed that adherence to Girardi epithelial cells and urinary tract epithelial LoVo cells is mediated by bacterial residues containing D-mannose, D-glucose, L-fucose, and D-galactose (119). Moreover, the uptake of endocarditis isolates into intestinal epithelial Int-407 cells is enhanced after a proteolytic treatment of the bacterial surface with trypsin, suggesting that some non-protein surface-exposed components adhere to intestinal cells (110). Analogous bacterial surface modification with proteinases and carbohydrate oxidation confirmed that carbohydrate structures rather than proteinaceous moieties drive E. faecalis adherence to intestinal epithelial Caco2 cells, as observed for the strains 12030 and FA2-2 (42). Monoglucosyl-diacylglycerol (MGlcDAG) and diglucosyl-diacylglycerol (DGlcDAG) are the major glycolipids of the E. faecalis cell envelope. Sava et al. showed that DGlcDAG, the membrane glycolipid anchor of LTA, is involved in the binding of E. faecalis to intestinal epithelial Caco-2 cells, most likely through interaction with host heparin and heparin sulfate (42). The latter are glycosaminoglycans (GAGs) consisting of repeating disaccharide units modified with sulfate groups and covalently bound to core proteins. However, the localization of DGlcDAG in the envelope raises the question of a direct interaction with GAGs. Sava et al. hypothesized that LTA shed from planktonic bacteria inserts into the host cell membrane and interacts with GAGs through its kojibiose residue of the glycolipid anchor. This interaction would allow contact with a yet-to-be identified enterococcal receptor (42). However, LTA–GAG interactions may not be common to all E. faecalis–epithelial cell interactions, as GAGs are not involved in E. faecalis 12030 binding to bladder epithelial T24 cells (43). In addition, loss of the glycolipid DGlcDAG by gene inactivation increases adherence to the T24 cells, whereas it modestly reduces attachment to intestinal epithelial Caco-2 cells (43, 44). D-alanylation of the LTA is a well-known modification involved in the adherence of pathogens, such as L. monocytogenes and S. aureus (120). Consistent with this, alanylation of LTA enhances adhesion of E. faecalis 12030 to epithelial Hep-2 cells (HeLa derivative), most likely by increasing the positive charge of the bacteria and thus promoting electrostatic interactions with the host cells (45). However, in E. faecalis, the role of D-alanylation of LTA may be more complex, as it can also decrease adherence to specific host cells, such as uroepithelial T24 cells (46).

No common proteinaceous adhesin mediates E. faecalis entry to all mammalian cell types

Few enterococcal protein adhesins have been reported to mediate adhesion to epithelial cells or fibroblasts. Microbial surface components recognizing adhesive matrix molecules (MSCRAMMs) represent a subfamily of bacterial adhesins that recognize and bind to extracellular matrix components. Several Gram-positive pathogens promote their uptake by epithelial and endothelial cells by expressing MSCRAMMs to bind to host integrins using matrix molecules as a molecular bridge (121). E. faecalis expresses a wide variety of extracellular matrix-binding proteins (122–125). Adherence to uroepithelial T24 cells is associated with the ability of E. faecalis isolates from urinary tract infections to bind to extracellular matrix proteins, such as fibronectin, laminin, and collagen types I, II, IV, and V (53, 125). Yet, whether and how MSCRAMMs function as a bridge to allow enterococcal adhesion and internalization by host cells remain to be elucidated.

Aggregation substances also mediate bacterial adhesion to epithelial cells. Asa1 was the first described LPxTG-anchored protein described in enterococci. Asa1 mediates adhesion to mammalian cells, including porcine renal tubular LLCPK1 cells (100). Overall, aggregation substances increase adhesion and entry of OG1-derived E. faecalis strains in intestinal epithelial cell lines derived from ileum (HCT-8 cells), duodenum (Hutu 80), and colon (HT29, T84, Caco-2) (102, 126, 127). Aggregation substances possess an Arg-Gly-Asp (RGD) motif found in some eukaryotic proteins, such as fibronectin, where it mediates binding between cells with integrin receptors. However, the RGD motif in Asc10 is not critical for enterococcal cell adhesion, whereas the aggregation domain of Asc10_(156-358) is required for efficient internalization of E. faecalis into HT-29 enterocytes (71). The adhesion of E. faecalis has been linked to other unanticipated factors. Adhesion of V583 to human platelets was proposed to be mediated by platelet-like proteins encoded by prophages PP1, PP4, and PP6 (128). Similarly, transduction of the V583 prophages PP1, PP5, and the phage-inducible chromosomal island EfCIV583 into the strain Symbioflor doubled adhesion to Caco-2 cells (129). Deletion of the cytoplasmic polynucleotide phosphorylase (PNPase), an exoribonuclease involved in RNA metabolism, in an E. faecalis strain isolated from meconium increases adhesion to intestinal epithelial Caco-2 cells (130). Alterations in the cell envelope observed in this mutant may contribute to this effect.

On the host side, the phospholipid receptor PAFr, whose expression is induced on intestinal epithelial cells during mucosal inflammation, is involved in E. faecalis translocation across intestinal epithelial Caco-2 cells (131). PAFr is also an innate immune recognition receptor for the phosphorylcholine moiety of LTA; however, whether E. faecalis LTA–PAFr interaction mediates E. faecalis internalization remains to be investigated.

To date, only a few enterococcal determinants involved in adhesion to and internalization into host cells have been reported. Only LTA, aggregation substance Asa1, and unidentified fibronectin-binding proteins are involved in E. faecalis adhesion to host cells. However, their role is not necessarily common to all strains and depends on the cell type. To our knowledge, with the exception of GAGs that have been proposed as receptors for E. faecalis during macrophage infection, no specific host receptors for E. faecalis have been formally identified and characterized. Given the heterogeneity among isolates in terms of gene distribution and structural diversity, enterococcal adhesins remain to be discovered according to the origin of the strain (urinary tract, wound infection, or endocarditis) and the host cell type used to model infection.

E. FAECALIS INTERNALIZATION IN EPITHELIAL CELLS INVOLVES ZIPPER-LIKE AND MACROPINOCYTOSIS MECHANISMS

Internalization of intracellular pathogens in epithelial cells relies on two main mechanisms (117, 132). The “zipper” mechanism used by Listeria, Neisseria, and Yersinia involves the binding to a cellular receptor and modest membrane expansion. The “trigger” mechanism is best described for Shigella and Salmonella and relies on the injection of effector proteins into the host cells, which subsequently interact with the actin cytoskeleton, leading to massive changes in the host membrane. Since the existence of an injectosome, such as type III secretion system, has not been reported in enterococci, enterococcal entry is more likely to be mediated by a zipper mechanism. Accordingly, once bacteria adhere to the epithelial cell surface, clinical E. faecalis isolates have been observed in direct contact with the curved plasma membrane of cervical epithelial HeLa cells consistent with entry by a zipper mechanism (133). Moreover, exposure of the lateral surfaces of enterocytes increased E. faecalis entry by ~1.5 log, suggesting that a cell receptor might be located on the lateral side of enterocytes (134). E. faecalis OG1X has been shown to efficiently adhere to the microvilli, the finger-like membrane protrusions supported by the actin cytoskeleton, in renal tubular cells (porcine cell line LLC-PK1) (100). Treatment with cytochalasin D (inhibitor of G-actin polymerization) or colchicine (inhibitor of microtubule polymerization) reduced the entry of clinical E. faecalis isolates in cervical epithelial HeLa cells and human umbilical vein endothelial cells (HUVECs) compared with untreated cells (133, 135). Bertuccini et al. proposed that E. faecalis internalization into epithelial cells involves more than one invasion pathway, including macropinocytosis and receptor-mediated endocytosis (133). In agreement, da Silva et al. recently showed that E. faecalis OG1RF internalization into keratinocytes occurs via macropinocytosis into single membrane-bound compartments (32). They also showed a strong dependence of E. faecalis internalization on actin polymerization. It has also been reported that actin is a predominant support for enterococci to pass through the colonic epithelial Ptk6 cell layer (69). In addition, internalization of E. faecalis into bovine mammary epithelial MAC-T cells triggers a rearrangement of the actin cytoskeleton to form stress fibers (113). Mechanisms that support the internalization of enterococci into endothelial cells and fibroblasts have not been reported to date. To summarize, E. faecalis can use different processes to promote internalization into host cells including a non-specific internalization by macropinocytosis and a zipper-like mechanism that may involve an affinity binding to a receptor.

WHAT IS THE FATE OF E. FAECALIS ONCE INTERNALIZED INTO HOST CELLS?

Upon recognition and phagocytosis of bacteria, phagocytic cells can induce the production of inflammatory cytokines and the initiation of downstream inflammatory cascades to clear E. faecalis (11, 37, 57, 136–140). However, E. faecalis is able to survive within phagocytic cells. The survival time varies from 24 h up to 72 h depending on the immune cell types, including human neutrophils and monocyte-derived macrophages and THP-1 cells, murine peritoneal and bone marrow-derived macrophages, and murine J774A.1 and RAW264.7 macrophages (30, 75, 83). Some factors, such as the Esp protein or LTA enhance pro-inflammatory cytokines through NF-kB activation (138, 141). In contrast, Tien et al. have shown that E. faecalis infection activates RAW264.7 macrophages only at low multiplicity of infection (MOI 1) and actively prevents NF-kB activation at an MOI of 100, resulting in a global decrease in cytokine, chemokine, and growth factor expression compared with that with a higher MOI or LPS exposure (142). Similarly, some E. faecalis strains isolated from healthy infants can also downregulate the secretion of the chemokine IL-8 after infection of intestinal epithelial Caco-2, HCT116, and HT-29 cells (143, 144). E. faecalis E99 and Symbioflor 1 produce a TIR domain-containing protein (TcpF) that directly interacts with MyD88 and reduces NF-kB activation (58, 145). However, this is not true for strains, such as OG1RF, which does not have a homolog of the same gene but can suppress NF-kB activation. Therefore, the full mechanism for NF-kB inhibition in macrophages by E. faecalis remains to be fully elucidated (145).

E. faecalis is well equipped to survive stressful conditions (146). However, only a few bacterial effectors have been shown to mediate intracellular survival of E. faecalis. The collagen-binding MSCRAMM Ace mediates E. faecalis JH2-2 survival in mouse peritoneal macrophages (62). Once internalized and engulfed in phagosomes, Asc10-expressing E. faecalis exhibits reduced killing by human neutrophils (75). In E. faecalis, the rhamnopolysaccharide EPA promotes resistance to macrophage phagocytosis and neutrophil killing (73, 76). E. faecalis has a manganese-containing superoxide dismutase (SodA), three NADH peroxidases, two alkyl hydroperoxidase reductases (AhpC, AhpF), two methionine sulfoxide reductases (MsrA and MsrB), and a thiol peroxidase (Tpx), which together mediate resistance to oxidative burst and macrophage killing (63, 64, 147, 148). During E. faecalis macrophage infection, nitrogen and oxygen reactive species are increased. Although E. faecalis can overcome oxidative stress to some extent, blocking ROS and NOS production in macrophages increased the survival of E. faecalis E99 (56). In addition, E. faecalis V583 survival in murine J774.A1 macrophages involves high affinity metal ion transport system (84). More recently, E. faecalis OG1RF was shown to decrease its carbohydrate metabolism to evade from clearance by RAW264.7 macrophages, highlighting for the first time a role for carbohydrate metabolism in E. faecalis survival in phagocytes (34). E. faecalis survival in macrophages involves both the subversion of host defenses and the fine-tuning of gene expression related to metabolism and stress response. Consistent with this, transcriptional regulators and two component regulatory systems (TCS) play a role in the bacterial survival within macrophages (149–156). For example, PerA, an AraC-type transcriptional factor, contributes to the survival of E. faecalis MMH594 in Raw264.7 macrophages (156). In addition, four TCS of the OmpR family are involved in the survival of E. faecalis JH2-2 in mouse peritoneal macrophages (153). Zou et al. reported that during early infection (12 h), E. faecalis blocks cell death of murine RAW264.7 macrophages by activating the PI3K/Akt pathway, a well-known cell survival pathway (59). This leads to the subsequent activation of the serine/threonine kinase Akt (protein kinase A), an anti-apoptotic factor. By contrast, at later time points of infection (24 h), activation of the pro-apoptotic caspase-3 involved in cell death suggests that macrophages could be used as a niche by E. faecalis before release and dissemination (59).

Autophagy is a cellular degradation and recycling process that occurs ubiquitously in all eukaryotic cells, whereby cytoplasmic components are targeted to lysosomes for degradation. Zou et al. showed that E. faecalis resists phagosome acidification and, presumably, autophagy (56). In this work, they observed two types of E. faecalis populations in RAW264.7 macrophages, some surrounded by single-membrane vacuoles and some that had lost their vacuolar compartment, suggesting that the latter may escape and reside in the cytoplasm. They also showed that E. faecalis can be found in acidic vesicles that co-localize with the LAMP-2 marker at early time points, albeit less acidic than those formed for E. coli. If their results support a fusion of lysosomes with some enterococcal phagosomes, E. faecalis was never found inside autophagosomes. The mechanism by which E. faecalis deacidifies the phagolysosome, allowing it to survive in this otherwise harsh environment, remains to be elucidated. E. faecalis ATCC 29212 in intracellular vacuoles and within the cytoplasm of murine bone marrow-derived dendritic cells and macrophages has also been observed by others (51, 157). In non-phagocytic cells, E. faecalis has been observed within membrane-bound cytosolic vesicles in cervical epithelial HeLa cells, intestinal epithelial HT-29 cells, keratinocyte HaCaT cells, and endothelial HUVEC cells (32, 102, 133, 135). da Silva et al. showed that once internalized into keratinocytes, some intracellular E. faecalis can be detected in early and late endosomes, but these compartments escape lysosomal fusion with late endosomes (32). At the same time, some intracellular E. faecalis within keratinocytes are found in compartments lacking the early endosome marker Rab5 during the first 30 min of infection and lacking both of the late endosome markers Rab7 and LAMP1 at later time points, indicating that intracellular E. faecalis is heterotypically trafficked. Together, these results do not indicate that E. faecalis always escape fusion with lysosomes, but rather that they can manipulate this pathway. These studies suggest the possibility of distinct E. faecalis-containing intracellular compartments that may ultimately lead to distinct intracellular fates (Fig. 2). The mechanism by which E. faecalis-containing compartments escape fusion with lysosomes has not been fully elucidated; however, E. faecalis infection also affects the levels of the endolysosome pathway proteins Rab5 and Rab7, which are required for lysosome fusion (32). Although E. faecalis was not found in autophagosomes of RAW264.7 macrophages (56), E. faecalis V583 can be engulfed in double-membrane autophagosomes upon activation of autophagy in intestinal epithelial cells, suggesting that E. faecalis reaches the cytosol during the infectious process (81). In agreement, Nunez et al. showed that at least one enterococcal OG1RF population is located in the cytosol during infection of human HUH7 hepatocytes (31). In summary, the location of intracellular E. faecalis after its internalization varies depending on the stage of infection and cell type.

Fig 2.

Overview of the intracellular fate of E. faecalis. After internalization by host cells, E. faecalis (in red) can follow an intracellular pathway towards elimination or survival. Proteins, such as Rab5, Rab7, and LAMP1, are key players in signaling lysosomal fusion, which culminates in bacterial elimination. E. faecalis has been reported to evade lysosomal fusion and replicate both within a compartment and in the cytosol. Several bacterial factors involved in adhesion and internalization of E. faecalis have been reported (Table 2), but those involved in survival within an intracellular compartment escape from it and even involved in potential egress from the host cells to cause reinfection remain largely unknown.

TABLE 2.

Enterococcal factors that promote infection

Factor	Relevance	References
Ace	Mediates survival in macrophages	(62)
AhpC, AhpF, MnSOD, MsrA, MsrB, Npr, and Tpx	Resistance to macrophage-mediated killing by resisting oxidative stress	(63, 64, 147, 148)
AS	Adhesion to host cells	(71, 79, 88, 100)
Capsular polysaccharide	Resistance to complement-mediated opsonophagocytosis	(97, 116, 158, 159)
Ebp	Adhesion to platelets; adhesion to fibrinogen and collagen	(160, 161)
ElrA	Resistance to phagocytosis	(70, 74)
EPA/WTA	Intracellular transit, complement evasion, resistance to macrophage phagocytosis and neutrophil killing	(73, 82, 162)
Esp	Adhesion to host cells; activation of NF-κB	(141, 163, 164)
GelE	Degrades complement and the junctional protein E-cadherin	(23, 165)
LTA	Adhesion to host cells	(45)
OatA and PgdA homologs	Resistance to macrophage-mediated killing by resisting lysozyme killing	(166)
PolyGlcNAc	Adhesion to epithelial cells	(27)
TcpF	Suppression of NF-κB activation	(145)

In 2022, it was shown that E. faecalis can not only persist in the intracellular environment but also replicate there. Together with Nunez et al. who demonstrated that E. faecalis can survive and divide within hepatocytes and kidney cells and form intracellular clusters, da Silva et al. showed that intracellular replication of E. faecalis in macrophages and keratinocytes likely occurs in modified intracellular compartments that evade fusion with lysosomes before release into the cytosol (31, 32). In support of this finding, previous studies have observed putative binary fission and increase in E. faecalis in macrophages (33, 34, 157). Taken together, these studies suggest that the intracellular lifestyle of E. faecalis is a widespread process.

INTERACTIONS BETWEEN E. FAECIUM AND MAMMALIAN CELLS DESERVE EXPLORATION

Similar to E. faecalis, E. faecium colonizes humans, withstands multiple stresses, resists antibiotic treatment, and persists in clinical settings, but studies of E. faecium host–cell interactions remain limited (167). Regarding interactions with phagocytic cells, E. faecium shares several similarities with E. faecalis: (i) complement and antibody-mediated opsonization play a role in recognizing E. faecium, (ii) the capsule likely provides resistance to complement-mediated opsonization and neutrophil killing, (iii) E. faecium evades phagocytosis by macrophages and neutrophils, iv) glycopolymers contribute to the resistance of E. faecium to phagocytosis and killing by leukocytes, and v) E. faecium LTA is a key ligand for binding to scavenger receptors (77, 85, 92, 95, 97, 158, 168). By contrast, information on the entry and internalization of E. faecium into epithelial cells is scarce. Paracytosis transport of E. faecium in intestinal epithelial Caco-2 cells has been reported (169). Apart from the clinical strain E. faecium E1162, which adheres to intestinal epithelial Caco-2 cells, and renal MDCK cells, and bladder T24 cells (163, 164), antibiotic-resistant E. faecium isolates from clinical and food samples were unable to adhere to Caco-2 cells (104, 170). At the molecular level, only the enterococcal surface protein (Esp), a multidomain LPXTG-anchored protein, has been reported to promote E. faecium adhesion to bladder T24 and kidney epithelial MDCK cells (163, 164). Similar to E. faecalis, E. faecium expresses LPxTG-type and WxL domain surface proteins (171, 172), and a wide variety of MSCRAMMs (173–175), but their role in adhesion to host cells remains to be investigated. In light of recent classification refinements, it cannot be ruled out that some of the E. faecium isolates used in the interaction studies were Enterococcus lactis (176). This gap in knowledge between E. faecalis and E. faecium is likely due to the fact that the primary risk of E. faecium is its high antibiotic resistance rather than virulence. This feature also complicates the traditional approach based on the antibiotic protection assay to studying E. faecium–cell interactions.

CONCLUSIONS AND FUTURE DIRECTIONS

Increasing evidence shows that opportunistic pathogens can be present both inside and outside cells, highlighting the need for a more flexible classification between “intracellular” and “extracellular” pathogens (177). Over the past decade, there has been an increase in reports of E. faecalis host–cell interactions. E. faecalis can be internalized by and persist in a wide range of mammalian cells, including vaginal, urinary, and intestinal epithelial cells, keratinocytes, hepatocytes, endothelial cells, fibroblasts, and immune cells, such as macrophages and neutrophils. E. faecalis is not only able to evade the host’s cellular defenses but also able to grow intracellularly. However, our understanding of the molecular mechanisms of E. faecalis cell invasion and their consequences for infection is still limited. It is important to understand what determines the specificity of isolates and the range of permissive cell types. Rational analysis with a set of representative isolates and related mutant collections would help to define proper cellular models of infection and investigate the molecular mechanisms that support E. faecalis internalization and replication, and that of the host membrane trafficking pathway. Technological advancements are also needed for precise detection of enterococci within human tissues to better understand their location and progression of infection. E. faecalis infects a variety of cell types, corresponding to the variety of organs and host sites it can infect. Whether the intracellular lifestyle of enterococci promotes persistence, asymptomatic low-grade infection, or resistance to treatment remains to be investigated. Consistent with the incidence of E. faecalis in wounds, endocarditis, and urinary tract infections, as well as its recent contribution to liver damage in alcoholic liver disease and to carcinogenesis, E. faecalis–host cell interactions deserve further investigation to fully understand E. faecalis-associated diseases.

Biographies

Cristel Archambaud is a microbiologist specialized in the study of bacterial pathogens and host interactions. She is currently a researcher in the CPE "Commensalism and Pathogenesis of Enterococci" team (Micalis Institute, INRAE, France) led by Pascale Serror, where she leads the thematic axis on "E. faecalis-host interactions". Over the last decade, her work has been instrumental in understanding the complexities of intracellular E. faecalis infection.

Natalia Nunez is a microbiology scientist specialized in the interplay of host-pathogen-microbiome. She currently is a research Lead at Life & Soft and works on vaginal microbiome and its impact in women’s health. Her objectives are the identification of biomarkers (genetic, metabolomic and bacterial) in different fertility-associated pathologies and the production of live biotherapeutic products for women health.

Ronni A. G. da Silva is a Research Scientist who leads projects that aim to develop strategies to harness the immune system's power to tackle infections and to investigate the pathogenicity of gram-positive bacteria like E. faecalis. His research interests lie in the intersection of host-pathogen interactions, antibiotic resistance and developing new therapies designed to target the host. Currently, he is part of the Antibiotic Resistance Interdisciplinary Group of SMART, a partnership between MIT and the Singaporean Government, and he is also a visiting Research Fellow at the Singapore Centre for Environmental Life Science Engineering (SCELSE) of Nanyang Technological University (NTU) - Singapore.

Kimberly A. Kline is a Professor in the Department of Microbiology and Molecular Medicine at the University of Geneva, and a visiting scientist at the Singapore Centre for Environmental Life Science Engineering (SCELSE) in Singapore. Her lab focuses on the pathogenesis of biofilm-associated infections, with a special focus on enterococci. Her team is working to understand the host-polymicrobial interactions that frequently characterize enterococcal biofilm infections, and how these interactions contribute to virulence.

Pascale Serror is Director of Research at the French National Research Institute for Agriculture, Food, and Environment (INRAE) and Leader of the "Commensalism and Pathogenesis of Enterococci" (CPE) team in the Microbiology of Human Health Laboratory (MICALIS - Univ. Paris-Saclay, INRAE, AgroParisTech). Her team focuses on the transition of enterococci from commensalism to pathogenesis. Her lab is currently working to understand microbiota-mediated mechanisms that prevent on enterococcal colonization of the gut and the interaction of E. faecalis with host cells to identify non-antibiotic strategies to prevent pathobionts from becoming pathogenic.