The interaction of gut microbes with host ABC transporters

Abstract

ATP binding cassette (ABC) transporters are increasingly recognized for their ability to modulate the absorption, distribution, metabolism, secretion and toxicity of xenobiotics. In addition to their essential function in drug resistance, there is also emerging evidence documenting the important role ABC transporters play in tissue defense. In this respect, the gastrointestinal tract represents a critical vanguard of defense against oral exposure of drugs while at the same time functions as a physical barrier between the lumenal contents (including bacteria) and the intestinal epithelium. Given emerging evidence suggesting that multidrug resistance protein (MDR) plays an important role in host-bacterial interactions in the gastrointestinal tract, this review will discuss the interplay between MDR of the intestinal epithelial cell barrier and gut microbes in health and disease. In particular, we will explore host-microbe interactions involving three apically restricted ABC transporters of the intestinal epithelium; P-glycoprotein (P-gp), multidrug resistance-associated protein 2 (MRP2) and cystic fibrosis transmembrane regulator (CFTR).

Introduction

The ABC transporter superfamily is among the largest and most broadly expressed protein superfamilies known. This family of transporters is found in all living organisms and is involved in almost every cellular, biological and physiologic system. Indeed, several thousand distinct ABC genes have been identified across numerous species. The vast majority of the ABC superfamily members play an essential role in the active transport of a wide variety of substrates across biological membranes. Remarkably, such substrates are chemically diverse and range from bulky lipophilic drugs and toxins, to organic anions, carbohydrates, bile acids and other xenobiotics.¹ For most ABC transporters, the binding and subsequent hydrolysis of ATP at their nucleotide binding domains (NBD) is required to provide energy for the movement of their substrates across membranes. In addition, two sequence motifs located 100–200 amino acids apart in each NBD, designated Walker A and Walker B, are conserved among all ABC transporter superfamily members.² ABC transporters apically positioned at the epithelial mucosal surface not only facilitate secretion of various molecules and ions such as chloride (Cl⁻) during physiological processes, but are also responsible for the efflux of harmful foreign substances including many drugs, toxins and food components encountered in the environment.³ Less appreciated, however, is recent evidence suggesting that multiple ABC transporters play a direct role in host interactions with bacterial pathogens, potentially expanding its credentials as a key player involved in the protection of the host.

The gastrointestinal tract, in particular, represents the body's first line of defense against oral exposure to toxins, drugs and pathogenic insult. The intestinal epithelium acts as a physical barrier separating the luminal environment and subepithelial tissues. In fact, the interface between luminal contents and the intestinal epithelium represents the largest mucosal surface directly confronting the external environment.⁴ Thus, central to the homeostatic balance of the gastrointestinal tract, is the regulation of the intestinal barrier function. The intestinal barrier function regulates transport and host-defense mechanisms at the mucosal interface and plays a critical role in the pathophysiology of gastrointestinal diseases.⁵ Several lines of evidence support the concept that impaired barrier function is the trigger to altered tolerance to xenobiotics and exogenous antigens, leading to the establishment of chronic inflammation.^5–8 The ABC efflux transporters P-gp, MRP2 and CFTR are all expressed at the apical surface of the epithelial cells lining the intestine. Under normal conditions, the expression of P-gp increases from the duodenum to the colon, whereas MRP2 expression is highest in the duodenum and subsequently decreases to become undetectable towards the terminal ileum and colon. Breast cell receptor protein (BCRPABCG2), another apically expressed ABC transporter, is found throughout the small intestine and colon.^9–11 CFTR, on the other hand, is a Cl⁻ channel that is the final rate limiting step for intestinal Cl⁻ secretion (i.e., fluid secretion in cholera and other enterotoxin-mediated secretory diarrheas). In the intestine, decreasing gradients of CFTR gene expression are observed along both the crypt-villus and proximal-distal axes.¹² This expression is consistent with CFTR being responsible for secretion of Cl⁻ in the intestinal crypts.

Despite the apical distribution on enterocytes little is known regarding how ABC transporters, such as Pgp, MRP2 and CFTR are influenced by gut microbes (pathogenic or otherwise), and the net physiologic consequences of such interactions. Therefore, this review will specifically highlight the emerging concept regarding the potential interplay between ABC transporters of the intestinal epithelial cell barrier and gut microbes in health and disease. A particular emphasis will be placed on the apically positioned efflux transporters P-gp, MRP2 and CFTR, all of which play a fundamental role in gastrointestinal homeostasis.^13,14

P-glycoprotein (P-gp)

P-glycoprotein (P-gp, ABCB1/MDR1) is a 170-kDa cell membrane-associated protein that is encoded by the human mdr1 gene, which maps to chromosome 7q21.1. P-gp was the first human ABC transporter characterized through its ability to confer a multi-drug resistance phenotype to cancer cells.¹⁵ This ATP-dependent efflux pump is highly involved in cell resistance to a variety of completely unrelated substrates, including metabolic products and drugs. Among the drugs most commonly effluxed by P-gp are colchicine, doxorubicin, etoposide, hydroxydaunorubicin, vinblastine and paclitaxel.¹⁵ The net result of such drug efflux is a decrease in intracellular drug concentration coupled to a reduction in cell cytotoxicity.

There is a regional variation of P-gp expression in the epithelial cells of the ileum with a gradual decline proximally to the jejunum, duodenum and stomach.^1,2 Classically, the role of P-gp in the gastrointestinal tract is to detoxify cells by actively effluxing toxins out of cells, but the precise physiological role of intestinal P-gp is unknown. Given its apically restricted distribution in the intestine, P-gp is ideally positioned to limit the absorption of substances that the cell perceives as harmful by polarized efflux into the intestinal lumen. Apical localization of P-gp also affords the opportunity for interactions with enteric bacteria. To this end, recent investigations have revealed an intriguing interaction between P-gp expressed on intestinal epithelial cells and enteric pathogens Listeria monocytogenes,¹⁶ Salmonella enterica serovar Typhimurium,¹⁷ and also perhaps unidentified members of the residential microbial community. For example, L. monocytogenes is a ubiquitous Gram-positive food-borne pathogen that causes listeriosis, a disease characterized by severe gastroenteritis, encephalitis, meningitis or septicemia, as well as miscarriage and is responsible for considerable morbidity and mortality.^16,18 L. monocytogenes is capable of infecting its host by actively invading non-phagocytotic cells, such as the intestinal epithelium. Consequently, the intestinal epithelium comes in contact with this pathogen along with its arsenal of secreted and surface-attached proteins. For instance, nearly 5% of the coding capacity of the L. monocytogenes genome is dedicated to surface proteins.¹⁶ Because L. monocytogenes interacts intimately with the intestinal epithelium, and given the inferred role of P-gp in acting as a cytotoxic defense mechanism, Neudeck and colleagues posed the question of whether the intestinal epithelium resists invasion by food-borne pathogens through the action of ABC efflux transporters. To test the hypothesis that intestinal P-gp is involved in host defense against L. monocytogenes, Neudeck et al. used complementary in vitro (Caco-2/MDR) and in vivo (mdr1a^-/- mice) models¹⁶ and found that overexpression of P-gp led to increased resistance to L. monocytogenes invasion, whereas P-gp protein inhibition led to increased invasion. These results imply that expression and function of P-gp is an important determinant in resistance to early invasion of L. monocytogenes.¹⁶ Their findings also invite speculation that that P-gp plays a novel role in host defense by its ability to efflux bacterial virulence proteins out from the host cell. Further studies will be required to validate this fascinating concept.

A negative association between invasion of intestinal epithelial cells and P-gp has also been reported for S. typhimurium.¹⁷ However, S. typhimurium appears to take direct charge of its fate by specifically dampening P-gp function, thereby promoting its ability to invade. S. typhimurium is a facultative intracellular pathogen that interacts intimately with the intestinal epithelium and causes a variety of diseases in humans and animals, ranging from gastroenteritis to systemic infection.¹⁹ The outflow of electrolytes and water characterizes the pathophysiology of such localized enteritis. In addition, large numbers of polymorphonuclear leukocytes (PMNs) migrate into the intestinal mucosa and lumen from the underlying microvasculature.¹⁹ Since S. typhimurium can co-opt numerous host cell biochemical pathways in an effort to subvert host functions, it was reasoned that this enteric pathogen might have evolved mechanisms that interfere with the function of multi-drug resistance transporters. Using an in vitro model of S. typhimurium infection of several polarized human cancer intestinal cell lines, Siccardi and colleagues found that S. typhimurium was able to functionally downregulate the efflux capabilities of P-gp.¹⁷ Moreover, cells suppressed in mdr1 expression were significantly more susceptible to the cytotoxic effects of S. typhimurium infection while on the other hand, this enteric pathogen was significantly less able to invade cells engineered to overexpress mdr1. It remains to be determined which factor(s) of S. Typhimurium are involved in reducing P-gp function in epithelial cells.¹⁷ To date, Cif, a secreted toxin from Pseudomonas aeruginosa, is the only bacterial component identified to selectively reduce the apical plasma membrane expression of P-gp in a variety of epithelial cells, including airway, intestinal and renal epithelial cells.²⁰ These results are consistent with a role for P-gp in the maintenance of homeostasis in the gastrointestinal tract and support the emerging concept of a protective role for this transporter in shielding the host not only from harmful xenobiotics, but also from invading bacterial pathogens. Of note, attempts are currently being made to harness the ability to S. typhimurium and Cif to impair P-gp for the purposes of enhancing the utility of chemotherapeutics towards their target. The development of novel P-gp inhibitors could be highly impactful, as the ability of cancer cells to develop resistance to multiple structurally and functionally non-related cytotoxic drugs, is a major barrier to successful chemotherapy.^21,22 Therefore, products from S. typhimurium, Cif analogs or the molecular target(s) of S. typhimurium or Cif may be candidates for developing a new class of specific P-gp inhibitors that will ultimately enhance the ability of chemotherapeutic agents to not only kill tumor cells but also to improve the bioavailability of other P-gp transport substrates as well (i.e., antibiotics and steroids).

P-gp is encoded by multidrug resistance genes located in a region of the human genome (7q21.1) where a gene involved in susceptibility to inflammatory bowel disease (IBD) may also be present.^23–25 IBD integrates two major conditions, ulcerative colitis and Crohn disease, which are generally assumed to result from abnormal mucosal immune responses to microbial antigens that would otherwise be well tolerated. Considerable progress with regards to understanding the pathophysiology of IBD has been garnered through numerous murine models of mucosal inflammation in the gastrointestinal tract. In one model, approximately 25% of mdr1a-deficient (mdr1a^-/-) mice were shown to be susceptible to developing severe, spontaneous colitis when maintained under specific pathogen-free conditions.²⁶ Interestingly, challenge of these animals with Helicobacter strains accelerated the occurrence of chronic colitis.²⁷ In mdr1a^-/- mice it is likely that the epithelium is no longer able to expel xenobiotics, and as a consequence of this defect, the mucosa is burdened with an overwhelming exposure to bacterial and other antigens. Thus, when this barrier function is impaired due to a failure to process antigens appropriately, colitis can ensue.⁴ Typically, the intestinal inflammation observed in this model is characterized by dysregulated epithelial cell growth and leukocytic inflammation into the lamina propria of the colon; pathological changes that are similar to that of human IBD. Furthermore, the IBD-like syndrome induced in mdr1a^-/- mice is completely reversed upon oral treatment with broad-spectrum antibiotics.²⁶ Therefore, this model not only demonstrates the physiological importance of P-gp in gastrointestinal tract homeostasis but also underscores the involvement of residential microbes in the induction of colitis related to loss of P-gp.

Several reports have also highlighted a potential role of the apical efflux pump P-gp in the specific IBD pathophysiology of human ulcerative colitis. As P-gp is considered to be a determinant of homeostatic interactions between bacteria and host in humans, loss of P-gp expression or a reduction in its function can result in a critical imbalance in this relationship that may eventuate in inflammatory disorders of the gastrointestinal tract.²⁸ For example, in inflamed intestinal epithelia of patients with gastrointestinal disorders, P-gp expression is strongly decreased, and in fact, such low levels may aggravate intestinal inflammation.²⁹ Furthermore, a single nucleotide polymorphisms (SNP) in the mdr1 gene (C3435T) is correlated with lower intestinal P-gp expression and higher incidence of ulcerative colitis, corroborating findings from the mdr1 knockout mouse model.^24,30 Although loss or altered function of P-gp seems to fundamentally correlate with ulcerative colitis,³¹ there are examples where such a correlation does not hold true.^32,33 Also, while genetic alteration of some response elements involved in sensing luminal pathogens such as nod2 may be associated with a higher incidence of Crohn disease, these mutations seemingly do not correlate with occurrence of ulcerative colitis.³⁴ The only genetic element observed to date for ulcerative colitis is decreased function of P-gp,²⁴ which, incidentally, was not correlated for individuals with Crohn disease.³¹

MRP2

MRP2 (ABCC2) is a 180-kDa cell membrane-associated protein and is encoded by the human mrp2 gene, which maps to chromosome 10q24t. MRP2 was first characterized functionally as a multispecific organic anion transporter in the bile canalicular membrane of hepatocytes, and is now known to be an important transport protein mediating the efflux of organic anions across the apical domain of hepatocytes, enterocytes and renal proximal tubule cells.^1,2 In the intestine, MRP2 expression is highest in the duodenum and subsequently decreases to become undetectable towards the terminal ileum and colon.³ MRP2 has also been demonstrated in the small intestine, where it is exclusively localized to the apical brush border membrane of villi. At this location, MRP2 is thought to play an important role in drug disposition. MRP2 expression also decreases in intensity from the villus tip to the crypts, where no expression has been observed.³⁵ Other parts of the duodenum, such as the submucosa and muscle layers, are negative for MRP2 expression.³⁵

Interestingly, MRP2 expression on intestinal epithelial cells appears to be enhanced rather than reduced by S. typhimurium, differing from the fate of P-gp.³⁶ In particular, during active states of intestinal inflammation induced by the enteric pathogen S. typhimurium, MRP2 is functionally upregulated and its expression is directly linked to active states of intestinal inflammation.³⁶ Furthermore, a profound upregulation of MRP2 is observed at the apical surface of the colonic epithelium in murine models of chronic intestinal inflammation, as well as in intestinal biopsies from patients presenting with active Crohn disease and ulcerative colitis.³⁶ This intriguing finding raises the question of why MRP2 is upregulated during active states of intestinal disease. One plausible explanation is that MRP2 facilitates the apical efflux of the eicosanoid hepoxilin A₃ (HXA₃), a potent neutrophil chemoatractant.³⁷ Neutrophil transmigration across mucosal surfaces contributes to dysfunction of epithelial barrier properties, a characteristic underlying many mucosal inflammatory diseases. Thus, HXA₃ an endogenous product of 12-lipoxygenase activity, is secreted from the apical surface of the epithelial barrier via MRP2 to establish a chemotatic gradient across the intestinal, thereby directing the transepithelial migration of neutrophils (the final step in neutrophil recruitment) associated with active states of intestinal inflammation.³⁷ Consequently, MRP2 is uniformly upregulated at the villus tips of the apical surface of epithelial cells during active states of inflammation and plays a pivotal role in promoting the inflammatory response.³⁶

On the balance of these observations, an important consideration that remains to be determined is the significance of the reciprocal expression of MRP2 and P-gp during active states of intestinal inflammation. Speculatively, in the case of S. typhimurium-induced colitis, P-gp might play a crucial role in counter-balancing MRP2 efflux. We hypothesize that P-gp mediates efflux of a molecule (or molecules) that is capable of blocking HXA₃/MRP2-driven PMN transmigration. In this way, a steady-state set point is established that limits inappropriate inflammatory responses in the face of the extensive bacterial burden present in the lumen of the colon, but which could also be poised to respond to the presence of a pathogenic insult. Thus, HXA₃ being released from MRP2 would override this steady-state, non-inflammatory set-point, and drive PMN transmigration in response to a pathogen or in association with acute intestinal flares experienced by UC patients. This hypothesis extends an earlier suggestion that mdr1-deficient mice develop mucosal inflammation not because of increased intestinal permeability,^38,39 but as a result of an increase in bacterial activation of enterocytes.²⁸ Thus, a dynamic relationship may exist between pathways that suppress responses to commensal bacterial and pathways that activate responses to pathogens (e.g., S. typhimurium). That MRP2 is extensively upregulated in actively inflamed intestinal tissue³⁶ and P-gp levels are downregulated in inflamed intestinal epithelia^17,29 provides a basis for these activating and suppressing responses. Moreover, there is precedence for coordinated function of ABC efflux family members. For example, the expression of CFTR (see next section) and P-gp are coordinately controlled in epithelial cells and induction of P-gp expression leads to a reversible repression of CFTR biosynthesis in the colon.^40–42

CFTR

CFTR is another protein belonging to the ABC superfamily, and it functions to regulate the secretion of ions across epithelial membranes in many tissues, including the lungs and intestine.^43,44 Quite remarkably, this transporter is used by Salmonella enterica serovar Typhi (S. typhi) as a receptor on intestinal epithelial cells.^45,46 S. typhi is common worldwide and is the causative agent of typhoid fever (enteric fever). This disease is characterized by a slow progressive fever reaching as high as 104°F (∼40°C), profuse sweating, some mild gastroenteritis and non-bloody diarrhea. The initial site of bacterial invasion occurs within the ileum and at this location the bacteria penetrate across the intestinal epithelia where they become phagocytosed by macrophages.^47,48 The organism is then spread via the lymphatics, ultimately providing the organism access to the reticuloendothelial system and then to the different organs throughout the body.⁴⁸ Of particular note, during serovar S. typhi infection of enterocytes the cell surface expression of the CFTR protein by intestinal epithelium was found to be markedly increased.⁴⁹ This increase results from the redistribution of pre-formed CFTR protein from intracellular stores to the epithelial cell plasma membrane, and the increased membrane expression of CFTR is correlated with enhanced CFTR-dependent entry of S. typhi into epithelial cells.⁴⁹ It was also determined that the products of certain strains of bacteria representing the intestinal microflora are also able to trigger CFTR in epithelial cells.⁵⁰ Such data provide compelling evidence to suggest that commensal microorganisms present in the intestinal lumen can affect the efficiency of S. typhi invasion of the intestinal submucosa and as a consequence, could be a key determinant influencing host susceptibility to typhoid fever. Given that the intestinal microflora consists of a heterogeneous population of microorganisms and the intestinal microbiome may contribute to the overall health status of its human host, future studies will be critical in determining what role the intestinal microflora plays in modulating the functional activity of apically restricted ABC transporters.

The CFTR gene encodes an adenosine 3′, 5′-cyclic monophosphate-regulated chloride channel, which when mutated can cause the hereditary disease cystic fibrosis.⁴⁴ Because CFTR is expressed strongly in enterocytes of the gastrointestinal tract and plays an important role in intestinal fluid secretion, this transporter has been a target for therapeutic intervention of enterotoxin-mediated secretory diarrheas, such as cholera. Cholera is a disease that causes watery diarrhea as a result of intestinal infection by the Gram-negative bacillus Vibrio cholerae.⁵¹ Cholera remains a major public health problem in Africa, Asia and Latin America and is transmitted through contaminated food and water and, when untreated, leads to severe dehydration and shock.⁵¹ Intestinal fluid loss in cholera is caused primarily by the release of the heteromeric toxin named cholera toxin.^51,52 Cholera toxin (CT) consists of an ‘A’ subunit coupled to a ‘B’ subunit that contains five identical peptides assembled in a pentameric ring. Once released, cholera toxin binds to intestinal enterocytes through interaction of the B subunit with the GM1 ganglioside receptor, and is then internalized through retrograde endocytosis. Within the cell, the ‘A’ subunit causes constitutive activation of adenylyl cyclase resulting in elevated levels of intracellular cAMP.⁵³ Elevation of cAMP then produces active secretion of salt and fluid through activation of the CFTR Cl⁻ channel in the apical plasma membrane of enterocytes.⁵³

Without medical treatment, mortality associated with cholera infection is 20–50%.⁵⁴ Current management of cholera is based on rehydration therapy with oral rehydration solutions. However, there remains a particular need for pharmacological approaches to treat cholera, especially in pediatric, elderly and immunocompromised individuals where dehydration can be complicated by malnutrition, pneumonia and other factors. Since CFTR is the final rate-limiting step for intestinal Cl⁻ secretion, and thus fluid secretion in cholera intoxication, CFTR expression at the lumen-facing cell membrane in the intestine provides a unique opportunity to treat secretory diarrheas using a host of CFTR inhibitors.⁵¹ For example, small-molecule thiazolidinone inhibitors of CFTR Cl⁻ conductance were discovered by high-throughput screening, and shown to be effective in preventing Cl⁻ and fluid secretion by cholera toxin in human intestinal cells and rodent models.⁵⁵ Additional screening has also revealed a glycine hydrazide class of CFTR inhibitors that also prevents CT-induced fluid secretion in vivo.⁵⁶ The glycine hydrazides appear to occlude the CFTR Cl⁻ channel pore at its external surface, suggesting the possibility of developing an orally administered, non-absorbable anti-secretory drug. Time will tell as to whether such CFTR inhibitors might become a practical treatment option for cholera.⁵¹

Concluding Remarks

The goal of this review was to highlight several fascinating relationships that exist between ABC transporters and bacteria that reside in the intestine (Table 1). In doing so a number of important themes have emerged: (1) the first emphasizes the important role ABC transporters play in host defense against invading pathogens, (2) the second underscores the remarkable ability of enteric bacteria to functionally alter ABC transporters (either up or down), as we speculate, may be part of their host invasion strategy and (3) the third describes how bacteria can use ABC transporters as a receptor for binding. As ABC transporters are important in maintaining gastrointestinal homeostasis, these few examples illustrate how bacteria can influence the regulation of these transporters and impact human health. These studies also provide a bounty of research opportunities of clinical significance. For example, completely unexplored is the involvement of the intestinal microbiota in the maintenance of ABC transporter homeostasis in health or disease. Moreover, there are untapped therapeutic possibilities in drug development for novel MDR inhibitors, in addition to identifying unique intervention targets for controlling mucosal inflammation. Thus, it is likely that research involving bacterial-host ABC transporter interactions will have far-reaching clinical implications as it opens an entirely new era of pharmaceutical targets for cancer, as well as for the potential treatment of mucosal inflammation in the gut.

Table 1.

ABC transporters and their interactions with GI microbes

	Anatomical distribution in the GI	Function	Interaction with GI Microbes	Role and changes in GI pathology
		Cell detoxification	Determinant of homeostatic microflora-host interactions.28
P-gp	Gradual expression increase from duodenum to colon	Multidrug resistance	Negative association between invasion of enterocytes and P-gp expression has been reported in: L. monocytogenes16S. typhimurium17	Single nucleotide polymorphism (SNP) in the mdr1 gene is correlated with lower P-gp expression and higher incidence of ulcerative colitis.24,30
			Functionally downregulated by S. typhimurium.17
MRP2	Highest in the duodenum and subsequently decreases to become undetectable towards the terminal ileum and colon.	Organic anion efflux	Upregulated during active states of inflammation induced by S. typhimutium.36	Plays an important role in neutrophils transmigration via apical efflux of hepoxilin A3.37
CFTR	Present in goblet cells throughout the GI tract, with higher expression in the duodenum and lower expression in the more distal intestine.12	Chloride ion channel.	Used by S. typhi as a receptor on enterocytes.45,46	Markedly increased surface expression during S. typhi infection, due to redistribution of CFTR from intracellular stores to the epithelial cell membrane.49

Financial Support

National Institutes of Health grants DK56754 and DK3306, and a Senior Investigator Award from the Crohn and Colitis Foundation of America (all to Beth A. McCormick). These funding sources were not involved in the writing of this report or the decision to submit the paper for publication.