The intestinal barrier in multiple sclerosis: implications for pathophysiology and therapeutics

Abstract

Biological barriers are essential for the maintenance of homeostasis in health and disease. Breakdown of the intestinal barrier is an essential aspect of the pathophysiology of gastrointestinal inflammatory diseases, such as inflammatory bowel disease. A wealth of recent studies has shown that the intestinal microbiome, part of the brain-gut axis, could play a role in the pathophysiology of multiple sclerosis. However, an essential component of this axis, the intestinal barrier, has received much less attention. In this review, we describe the intestinal barrier as the physical and functional zone of interaction between the luminal microbiome and the host. Besides its essential role in the regulation of homeostatic processes, the intestinal barrier contains the gut mucosal immune system, a guardian of the integrity of the intestinal tract and the whole organism. Gastrointestinal disorders with intestinal barrier breakdown show evidence of CNS demyelination, and content of the intestinal microbiome entering into the circulation can impact the functions of CNS microglia. We highlight currently available studies suggesting that there is intestinal barrier dysfunction in multiple sclerosis. Finally, we address the mechanisms by which commonly used disease-modifying drugs in multiple sclerosis could alter the intestinal barrier and the microbiome, and we discuss the potential of barrier-stabilizing strategies, including probiotics and stabilization of tight junctions, as novel therapeutic avenues in multiple sclerosis.

Introduction

Biological barriers separate the internal milieu from the external environment and are essential components of maintaining homeostasis. A compromised intestinal barrier function is a prominent feature of many diseases, such as inflammatory bowel disease (Choi et al., 2017; Martini et al., 2017; Mu et al., 2017), graft versus host disease (Nalle and Turner, 2015) and coeliac disease (Schumann et al., 2017), but other biological barriers also fail in a myriad of pathological conditions (e.g. renal tubules in glomerulonephritis and lung alveoli in acute respiratory distress syndrome). The CNS is highly sensitive to homeostatic changes, and as such requires its own specialized barrier, the blood–brain barrier, for appropriate functioning. Breakdown of the blood–brain barrier is an essential hallmark of multiple sclerosis pathophysiology. Immune mediated dysregulation of the blood–brain barrier allows for migration of activated inflammatory cells into the brain, which in turn induces demyelination, axonal loss and other tissue damage (Ortiz et al., 2014; Kamphuis et al., 2015). Interestingly, many of the tight junction molecules in endothelial cells of the brain–blood barrier are identical to those in intestinal tissues, such as occludin, claudins and zona occludens-1 (Reinhold and Rittner, 2017). In this review, we examine the multiple lines of evidence, albeit mostly indirect, linking the intestinal barrier function and multiple sclerosis pathophysiology. We also discuss the possible effect of multiple sclerosis disease-modifying therapies and their association with the gut microbiome.

The intestinal barrier

The intestinal barrier maintains homeostasis by preventing the unwanted movement of antigenic molecules and microbes from the lumen of the gastrointestinal tract, while allowing the products of digestion and water to enter the body. The intestinal barrier consists of a physical barrier provided by the inter-epithelial tight junctions, a secretory barrier that includes antimicrobial peptides, mucus and fluid and an immunological barrier, including cells and molecules of the innate and adaptive immune system. The secretory component of the epithelial barrier is regulated by neural mechanisms that integrate this component of barrier function with digestive processes in the gut. Intestinal barrier function refers to ability of the intestinal mucosa and extracellular barrier components (e.g. mucus, antimicrobial peptides) to modulate epithelial permeability and act as a physical and functional limiting step for organism-luminal interactions.

The intestinal lumen and its contents are separated from the rest of the gastrointestinal tissue (and the body) by a single layer of epithelial cells along the length of the gastrointestinal tract. These cells are being constantly renewed and thus require constant proliferation (Delgado et al., 2016). Intestinal stem cells, present in the crypts of the intestinal mucosa, differentiate into both enterocytes, and specialized secretory (Paneth cells and goblet cells) and sensory cells (enteroendocrine cells and tuft cells), a process regulated by complex transcriptional and epigenetic mechanisms (Smith et al., 2017). The intestinal barrier is permeable to water and other small molecules, a property modulated by tight junctions, located around the apical surface of adjacent epithelial cells. Tight junctions consist of a heterogeneous group of transmembrane proteins such as occludins, claudins, junctional adhesion molecules and zona occludens-1, each with specific roles (Gasbarrini and Montalto, 1999; Sturgeon and Fasano, 2016; Volynets et al., 2016; Capaldo et al., 2017; France and Turner, 2017).

The intestinal barrier (Fig. 1) is continuously exposed to a number of immunological and microbiological factors. When the permeability of the intestinal barrier is breached, undesired large molecules, and commensal bacteria, may enter the lamina propria with pathological consequences (Odenwald and Turner, 2017). One of the main causes of increased permeability of the intestinal barrier is inflammation, an event thought to be essential in the pathophysiology of inflammatory bowel disease (IBD) (de Souza et al., 2017; Martini et al., 2017), coeliac disease and sepsis (Yoseph et al., 2016; Schumann et al., 2017). Inflammatory cytokines including interferons, interleukin (IL)-17 and tumour necrosis factor alpha (TNFα), as well as calcium-dependent oxidative stress, have been shown to alter the expression of tight junction proteins and lead to increased intestinal permeability (Reynolds et al., 2012; Yang et al., 2014; Al-Sadi et al., 2016; Gangwar et al., 2017).

Together with intestinal epithelial cells as the first layer of the intestinal barrier are Paneth cells (Fig. 1), which are specialized secretory cells derived from intestinal stem cells. Paneth cells produce antimicrobial peptides, the defensins, which are secreted into the mucus layer (Dupont et al., 2014; Yu et al., 2016; Capaldo et al., 2017). Mucus, secreted from goblet cells, is composed of heavily glycosylated oligomeric mucin proteins, water, ions and secretory IgA. This layer modulates bacterial growth in the intestinal lumen adjacent to the intestinal barrier, prevents bacterial adherence and acts as part of the innate immune response of the organism against microbial pathogens (Dupont et al., 2014).

After the mucus and the epithelial lining of the gastrointestinal tract, the next layer of the intestinal barrier is mostly immunological. Innate lymphoid cells, located in the epithelial layer, can be activated to produce a variety of inflammatory mediators, which play a defensive or a pathogenic role in mammal gut homeostasis (Bostick and Zhou, 2016). Found in close proximity to the single layer of enterocytes, intraepithelial lymphocytes are a heterogeneous population of cells that provide immune protection against pathogens and also regulate immune responses that, if unchecked, could jeopardize the integrity of the barrier (Cheroutre et al., 2011; Olivares-Villagómez and Van Kaer, 2018). The lamina propria (Fig. 1) is populated by B, T and dendritic cells that can initiate and modulate a host of immunological responses (Persson et al., 2013; Gronke et al., 2017). Peyer’s patches are secondary lymphoid tissues present in the intestinal mucosa. They are continuously exposed to a variety of antigens, presented to Peyer’s patches by microfold epithelial cells and resident dendritic cells (Rochereau et al., 2011; Hashiguchi et al., 2015).

CNS demyelination and intestinal barrier breakdown in gastrointestinal disorders: an important link?

An association between multiple sclerosis and IBD has been suggested because of common epidemiological, immunological and genetic patterns (Barcellos et al., 2006). IBD patients have an increased risk for cerebrovascular disease, peripheral neuropathy and demyelinating disease (Casella et al., 2014; Ferro et al., 2014; Morís, 2014), and anti-TNF therapies that are widely used in IBD have also been associated with CNS demyelination (Katsanos and Katsanos, 2014). Indeed, a recent meta-analysis of 10 case-control studies including over 1 million patients found a risk ratio of 1.54 for multiple sclerosis/IBD comorbidity, with no difference between Crohn’s disease and ulcerative colitis (Kosmidou et al., 2017). Certain authors propose that IBD can be conceived as a disorder of the intestinal epithelial barrier, and barrier breakdown is known to be an essential step in the pathophysiology of both Crohn’s and ulcerative colitis (for reviews see Jäger et al., 2013; Antoni et al., 2014; Goll and van Beelen Granlund, 2015). Evidence showing white matter involvement in IBD could also provide a link between intestinal barrier breakdown and CNS demyelination.

In an early study, investigators found a 3-fold increase of white matter hyperintensities in the MRIs of patients with IBD (Geissler et al., 1995). A recent estimate suggested that over half of all IBD patients will have white matter hyperintensities in a routine MRI (Ferro et al., 2014). Other findings in IBD include decreased grey matter volume and decreased axial diffusivity in major white matter tracts (Zikou et al., 2014). The aetiology of the white matter lesions found in patients with IBD is uncertain, and some authors suggest that ischaemia and vasculitis might be responsible (Zikou et al., 2014). However, in a report of five cases of patients with Crohn’s disease with symptomatic acute white matter lesions suggestive of demyelination, systemic infection, coagulation disorders, or vasculitis were ruled out (de Lau et al., 2009). Additionally, other studies have attempted to describe white matter lesions in patients with Crohn’s disease with more detail. White matter lesions suggestive of demyelination were found in 72% of 54 patients compared to 34% in age- and sex-matched controls (Chen et al., 2012). The role of anti-TNF therapy is also debated, and some observational studies have not found an association between therapy and presence of white matter hyperintensities (Chen et al., 2012). In a retrospective analysis of 9095 patients with IBD, anti-TNF therapy was not found to increase the risk of confirmed inflammatory demyelinating CNS lesions (de Felice et al., 2015).

Other gastrointestinal diseases where the intestinal barrier is impaired have also been associated with CNS demyelination. In patients with multiple sclerosis, serological and histological markers of coeliac disease are more frequent than in healthy controls (Rodrigo et al., 2011), although other studies have found inconsistent results (Salvatore et al., 2004). Cases of comorbid coeliac disease and multiple sclerosis are abundant in the literature (Batur-Caglayan et al., 2013; Casella et al., 2016), as are cases of coeliac disease with white matter lesions mimicking multiple sclerosis or other CNS demyelinating diseases (Mirabella et al., 2006; Finsterer and Leutmezer, 2014; Krom et al., 2017). MRI studies in patients with coeliac disease have also shown higher proportion of white matter lesions and grey matter atrophy (Bilgic et al., 2013).

Although a causal link between intestinal barrier breakdown and CNS demyelination cannot be concluded with certainty in these cases, there appears to be an association not solely explained by their shared epidemiological and immunological characteristics. The association between these entities is certainly complex and in need of further study.

Intestinal barrier homeostasis, the microbiome and neuroinflammation: possible mechanisms linking these entities

The interactions between the microbiome and the intestinal barrier, particularly the contribution of the microbiome in maintaining barrier homeostasis, could be central in accounting for its regulation of neuroinflammation (Fig. 2). Several studies have established that there are alterations in the gut microbiome of patients with multiple sclerosis, which has further fuelled the interest in the brain-gut-microbiome connection in multiple sclerosis research.

Early studies showed that, when compared to controls, patients with relapsing-remitting multiple sclerosis have an abundance of Anaerostipes, Faecalibacterium, Pseudomonas, Mycoplasma, Haemophilus, Blautia, and Dorea and a relative decrease of Bacteroides, Prevotella, Parabacteroides and Adlercreutzia (Cantarel et al., 2015; Miyake et al., 2015; Chen et al., 2016). In paediatric multiple sclerosis, patients have higher levels of members of Desulfovibrionaceae and depletion in Lachnospiraceae and Ruminococcaceae (Tremlett et al., 2016a). However, a clear and consistent ‘multiple sclerosis microbiome phenotype’ has not been described, and a myriad of different species have been implicated. For example, studies have found a significant depletion in Clostridial species (Rumah et al., 2013; Miyake et al., 2015), Butyricimonas (Jangi et al., 2016), Roseburia (Swidsinski et al., 2017) and increases in Streptococcus (Cosorich et al., 2017), Methanobrevibacter, Akkermansia and Coprococcus (Cantarel et al., 2015; Jangi et al., 2016). Multicentre studies aiming at defining a ‘core microbiome’ are underway (Pröbstel and Baranzini, 2018). Furthermore, some of these changes in the microbiome have been associated with immunological derangements, such as differences in the expression of genes involved in interferon and nuclear factor kappa-B (NF-κB) signalling (Jangi et al., 2016), and numbers of pro-inflammatory T helper 17 (Th17) cells in the intestine (Cosorich et al., 2017). At least one study found that differences in the microbiota could predict relapse risk in paediatric multiple sclerosis patients (Tremlett et al., 2016b).

Insights into how the microbiome could alter neuroinflammatory responses (reviewed in Colpitts and Kasper, 2017; Wekerle, 2017) have been illuminated by studies in germ-free mice where the microbiome regulates the shift back-and-forth of immune cells from pro- to anti-inflammatory phenotypes (Berer et al., 2011). Mice maintained under germ-free conditions have an attenuated form of experimental autoimmune encephalomyelitis (EAE), an inflammatory model of multiple sclerosis, and show lower levels of IL-17 in both the gut and the CNS, while also showing an increase in regulatory T cells (Tregs) peripherally (Lee et al., 2011). Colonization with segmented filamentous bacteria in germ-free mice leads to increased production of IL-17 and development of severe EAE. In contrast, other gut commensals such as P. histicola are able to suppress EAE severity, by decreasing pro-inflammatory Th1 and Th17 cells, and increasing Tregs and suppressive macrophages (Mangalam et al., 2017). B. fragilis, another common commensal strain, can also suppress EAE by expanding Tregs expressing the ectonucleotidase CD39, allowing for increased migration of this regulatory cell type into the CNS (Wang et al., 2014). Microbiota abundant in patients with multiple sclerosis induce the differentiation in vitro of human peripheral blood mononuclear cells into Th1 cells while reducing Treg numbers; conversely, microbiota that are decreased in patients with multiple sclerosis stimulate anti-inflammatory IL-10-expressing T cells and FoxP3+ Tregs (Cekanaviciute et al., 2017). Microbiota from patients with multiple sclerosis transplanted to mice prone to develop spontaneous EAE increases their susceptibility to EAE (Berer et al., 2017). Interestingly, multiple sclerosis patient-derived microbiota transplantation did not lead to changes in tight junction protein expression in the mouse recipient gut, but splenic lymphocytes had impaired IL-10 production (Berer et al., 2017).

An altered microbiome also leads to changes in some bacteria-associated products known to influence neuroimmune responses. Short chain fatty acids (SCFAs) such as butyrate, propionate and acetate are produced by bacterial fermentation of dietary carbohydrate and fibre. They play important roles in maintaining intestinal homeostasis, such as mediating sodium transport, serving as the principal energy source of intestinal epithelial cells and modulating gene transcription via inhibition of histone deacetylase activity (Kiela and Ghishan, 2016). Although not focusing on the concentration of SCFAs, CSF metabolomics studies from patients with multiple sclerosis have shown significant differences when compared to controls. SCFAs such as acetate are reduced (Simone et al., 1996; Kim et al., 2017), while others such as formate (Kim et al., 2017) have been found to be elevated in patients CSF. In studies evaluating metabolites in urine, propionate metabolism has also been found to be altered in patients with multiple sclerosis (Gebregiworgis et al., 2016).

In experimental models, eradication of the gut microbiota, or even just limiting the intestinal microbiome diversity, leads to impaired microglia structure and immune function, a process regulated by SCFAs (Erny et al., 2015, 2017). Astrocytes may also be influenced by SCFAs and the microbiome. Dietary tryptophan is metabolized by the gut microbiota into aryl hydrocarbon receptor agonists such as indoxyl-3-sulfate and indole-3-propionic acid, which can modulate astrocyte inflammatory function trough limiting NF-κB activation in a suppressor of cytokine signalling 2-dependent manner (Rothhammer et al., 2016). SCFAs also reduce T cell proliferation and cytokine production in the gut (D'Souza et al., 2017; Wan Saudi and Sjöblom, 2017). In EAE models, the administration of SCFAs led to amelioration of disease severity in association with a reduction of Th1 cells and an increase in Tregs (Mizuno et al., 2017). Interestingly, an altered microbiota may also alter innate immune responses in the gut favourable for systemic autoimmunity. For example, some types of intraepithelial lymphocytes may act as Tregs that suppress the pathogenic response to the immunizing antigen in EAE (Tang et al., 2007). CD4(+) intraepithelial lymphocytes obtained from transgenic mice prone to develop spontaneous EAE can infiltrate the CNS and ameliorate EAE severity in wild-type mice on transfer, showing regulatory properties (Kadowaki et al., 2016). These same cells proliferate in response to gut-derived antigens, aryl hydrocarbon receptor ligands and microbiota.

SCFAs could also modulate blood–brain barrier permeability. It is well known that SCFAs enhance intestinal epithelial cell barrier function by increasing the expression of tight junction proteins (D'Souza et al., 2017; Wan Saudi and Sjöblom, 2017). Butyrate has also been shown to increase the expression of occludin and zona occludens-1, thus restoring blood–brain barrier permeability in models of traumatic brain injury (Li et al., 2016). In germ-free mice exhibiting an altered blood–brain barrier, butyrate administration led to increased occludin expression and preserved blood–brain barrier permeability (Braniste et al., 2014). Overall, changes in SCFA-producing bacteria in the gut, and the influx of SCFAs into the blood stream, could thus have a distal effect in microglia and astrocyte functions, as well as in modifying blood–brain barrier permeability and the entrance of immune cells into the CNS (Fig. 2).

Besides the above-discussed mechanisms suggesting bystander activation, another possible immunopathogenic link between multiple sclerosis and the gut microbiota is that of molecular mimicry. CNS-specific, self-reactive lymphocytes might be cross-activated by both gut microbiota antigens and myelin (Berer and Krishnamoorthy, 2014). Although there is no conclusive evidence for these mechanisms, commonly found pathogenic and non-pathogenic gut bacteria such as Bacteroides spp. and Enterococcus faecalis possess potential myelin basic protein encephalitogenic mimics (Westall, 2006).

The intestinal barrier in multiple sclerosis: consequences of a leaky gut

Recent attention in the brain-gut connection in multiple sclerosis research has been focused on the role of the commensal gut microbiome while largely ignoring the interface of the microbiome with the organism, i.e. the intestinal barrier. Therefore, actual evidence for an alteration of the intestinal barrier in multiple sclerosis is limited. In a study of 12 jejunal biopsies from multiple sclerosis patients, Lange and Shiner (1976) found subtle histological changes, such as two cases of villous atrophy, as well as some cases of intestinal inflammatory cell infiltration. A later study found similar infiltrates, and also evidence of intestinal malabsorption in close to 20 of 52 patients with multiple sclerosis (Gupta et al., 1977).

In 1996, Yacyshyn et al. (1996) showed that 5 of 20 patients with multiple sclerosis had an altered lactulose/mannitol permeability test, suggesting increased intestinal permeability, a finding also associated with peripheral expression of CD45RO on CD20+ B cells. In the most recent study to date, the lactulose/mannitol permeability test was again used to evaluate intestinal permeability in 22 patients with multiple sclerosis and compared with age- and sex-matched controls (Buscarinu et al., 2017). Investigators found abnormal permeability in 73% of cases versus 28% in controls, but no association between permeability and brain MRI lesion load.

Similar findings have been recently described in the EAE model, the prototypic inflammatory animal model of multiple sclerosis. Investigators have found altered intestinal permeability, reduced submucosa thickness and altered tight junction expression in intestinal epithelial cells (Nouri et al., 2014). These alterations could also be induced in mice by adoptive transfer of pathogenic T cells. Furthermore, a recent study showed that the degree of intestinal permeability disturbance is closely associated with EAE severity (Secher et al., 2017). Treatment with Escherichia coli strain Nissle 1917, a probiotic known to improve intestinal barrier function, preserved tight junction expression and decreased intestinal permeability, leading to reduced EAE severity and decreased secretion of pro-inflammatory cytokines and an increased production of the anti-inflammatory cytokine IL-10 (Secher et al., 2017). This reduction of intestinal permeability led to a reduction of the migration of inflammatory T cells to the CNS, suggesting an impact on blood–brain barrier permeability as well (Secher et al., 2017).

The above studies suggest that there is indeed an alteration in the intestinal barrier in patients with multiple sclerosis and that these changes are at least partly due to an altered intestinal immune response (Buscarinu et al., 2017). The clinical relevance of these findings is unclear, but several possibilities arise. Intestinal barrier dysfunction has been associated with susceptibility to systemic infections (König et al., 2016), and both CNS and systemic infections are a common complication in patients with multiple sclerosis (Venkatesan, 2015). Another possibility is that the intestinal barrier’s interplay with commensal microbiota could modulate the immune response pathologically. Finally, alterations in intestinal permeability may modulate or perpetuate neuroimmune dysregulation by increased transmucosal passage of injurious or immunogenic antigens.

The essential role of the commensal microbiome in the regulation of intestinal immunity is beginning to be recognized, and several recent reviews have been published on this subject (Haak and Wiersinga, 2017; Shi et al., 2017). Commensal bacteria are able to strengthen the gut barrier and regulate intestinal permeability (Lin and Zhang, 2017). A healthy microbiota also preserves intestinal epithelial cell integrity through the production of SCFAs that increase tight junction expression and through toll-like receptor activation (Wells et al., 2017). Intestinal commensal bacteria are recognized by toll-like receptors, a process leading to protection of intestinal epithelium against injury and barrier disruption (Rakoff-Nahoum et al., 2004). Toll-like receptor signalling also promotes epithelial cell proliferation, IgA secretion and expression of antimicrobial peptides in Paneth cells (Abreu, 2010; Wells et al., 2011).

Alterations in the gut homeostatic mechanisms in multiple sclerosis could have as one of its consequences increased bacterial translocation through an impaired intestinal barrier. One recent study found elevated levels of endotoxin [lipopolysaccharide (LPS)] in plasma of patients with multiple sclerosis, and endotoxin concentrations were related to in vivo IL-6 production and increased in vitro T-helper 17 (Th17)-like responses (Teixeira et al., 2013). Circulating endotoxin was also correlated with the Expanded Disability Status Scale, a measure of clinical disability in multiple sclerosis. In another study, LPS and LPS-binding protein were found to be elevated in the serum of EAE-induced mice; investigators also found increased LPS-binding protein levels in the serum of multiple sclerosis patients compared to healthy controls (Escribano et al., 2017). These studies are evidence of a low-grade endotoxaemia that could be present in patients with multiple sclerosis, possibly due to bacterial translocation in the setting of an altered intestinal barrier.

Besides LPS, enteric bacteria also produce microbial-associated molecular patterns (MAMPs) such as bacterial lipoproteins and double-stranded RNA that can enter the systemic circulation and act through toll-like receptors to modulate the immune system (Patten and Collett, 2013). Toll-like receptors are known to be expressed in microglia and to modulate initiation and severity of EAE in experimental models (Miranda-Hernandez and Baxter, 2013). LPS is a well-known stimulant of microglial responses and is able to disrupt the blood–brain barrier by increasing microglial production of matrix metalloproteinases (Frister et al., 2014). LPS and other MAMPs could constitute another pathway by which an altered intestinal barrier could affect neuroimmune responses in multiple sclerosis.

Finally, the use of oral disease-modifying therapies and/or symptomatic drugs in multiple sclerosis also constitute a concern, as the intestinal barrier is essential in drug absorption (Sánchez-Navarro et al., 2016). On the other hand, there are no currently marketed therapies to improve intestinal barrier function; nutritional, microbial-derived and probiotic agents are being investigated. In the next section, we will discuss the possible effects of currently used disease-modifying therapies on the intestinal barrier as well as other pathophysiological considerations.

Disease-modifying therapies and the intestinal barrier

An interesting aspect of the above mentioned findings is that the microbiome can also be altered by whatever immunomodulatory therapy the multiple sclerosis patient is receiving (Cantarel et al., 2015; Tremlett et al., 2016b). The question of whether gut dysbiosis precedes the development of multiple sclerosis or follows the immune alterations (innate, acquired or drug-induced) is also a matter of debate (Ochoa-Repáraz et al., 2017). Disease-modifying therapies are medications that have improved the clinical course of relapsing-remitting multiple sclerosis. While their principal mechanisms are thought to be immune-modulating, their possible effects over the intestinal barrier that may contribute to therapeutic efficacy have not been explicitly evaluated. Below we summarize evidence suggesting that disease-modifying therapies could modulate the intestinal barrier, the gut microbiome and the interaction between the two (Fig. 3). However, the evidence is indirect, and whether this actually plays a meaningful role in clinical response remains to be established.

Interferons

There is evidence suggesting that endogenous interferons could affect the intestinal barrier. Type I interferons, including IFNα and IFNβ, are an integral part of the innate host immune response to gut microbiota, and they modulate bilateral interactions between epithelial cells and commensal flora (Giles and Stagg, 2017). For example, IFNβ has shown stabilizing properties in biological barriers (such as the intestinal, blood–brain and blood–lung barriers), partly through the upregulation of tight junction proteins in endothelial cell layers (Kraus et al., 2004; LeMessurier et al., 2013; Long et al., 2014). The commensal microbiota also stimulates dendritic cell IFNβ production, which increases the proliferation of Tregs in the intestine, a process itself inhibited by intestinal epithelial cell apoptosis (Nakahashi-Oda et al., 2016). Type I interferons also inhibit the continuous proliferation of the intestinal epithelium by activating the p53 pathway and inducing epithelial cell apoptosis (Katlinskaya et al., 2016), and mice lacking type I interferon receptor on Paneth cells show an altered microbiota (Tschurtschenthaler et al., 2014).

Glatiramer acetate

Various studies have shown that glatiramer acetate reduces colonic injury in animal models of colitis, through reduction of TNFα signalling, elevation of regulatory T cells and increase in anti-inflammatory mediators such as IL-10 and TGFβ (Aharoni et al., 2005, 2007). In one such study, glatiramer acetate attenuated colitis severity and prevented the destabilization of the intestinal epithelial barrier (Yablecovitch et al., 2011). There is also evidence suggesting that patients with multiple sclerosis treated with glatiramer acetate have different microbiota composition. In a small study, glatiramer-treated patients had stool taxonomic units (evaluated by hybridization of 16S rRNA to a DNA microarray) of Bacteroidaceae, Faecalibacterium, Ruminococcus, Lactobacillaceae, Clostridium, and other Clostridiales that were significantly different than those of untreated patients (Cantarel et al., 2015).

Natalizumab

Dysregulated recruitment of leucocytes into the intestine is one of the components of the immune response responsible for barrier breakdown in IBD (Danese et al., 2005; Fiorino et al., 2010). Integrins are expressed on intestinal lymphocytes and are essential in their homing to intestinal lymphoid tissues and trafficking through the intestinal mucosa (Hamann et al., 1994; Tanaka et al., 1995; Miura et al., 1996; Farstad et al., 1997; Bradley et al., 1998; Fujimori et al., 2002). Natalizumab, which blocks the activity of integrins (both α4β1 and α4β7), has shown effectiveness in reducing the severity of IBD (Fiorino et al., 2010; Bamias et al., 2013). However, its association with JC virus-related CNS complications has led to the development of specific α4β7-antibodies such as vedolizumab, now routinely used in the treatment of IBD (Zundler et al., 2017).

Nonetheless, the effects of natalizumab on integrins and lymphocyte trafficking in the gut suggests it could modulate the inflammatory response in this site in multiple sclerosis. A potential role for intestinal lymphocytes and integrins in multiple sclerosis pathophysiology has been suggested by results from mouse EAE models. Th17 cells, prominent drivers of EAE, are controlled and redirected in the small intestine. Th17 cells, which are normally pro-inflammatory, acquire a regulatory phenotype in the intestine and are ultimately eliminated through the intestinal lumen (Esplugues et al., 2011). In EAE, there is infiltration of proinflammatory Th1/Th17 cells and reduction of Tregs in the gut, in association with functional and morphological changes (Nouri et al., 2014). Furthermore, mice lacking integrin α show a loss of Th17 cells in the intestine and resistance against EAE (Acharya et al., 2010; Melton et al., 2010). In spontaneously EAE resistant B10.S mice, blocking α4β7 integrin leads to peripheral availability of Th17 cells and increased severity of EAE (Berer et al., 2014). In patients with multiple sclerosis, natalizumab treatment reduces the populations of integrin α-4-positive Th1, Th17 and Tregs differentially, while affecting the immune function of residual integrin α-4-positive T cells (Kimura et al., 2016). The gut might act as a checking point, a reservoir and an activation site for Th17 and other T cells, a process regulated in part by intestinal integrins. Natalizumab and its non-selective integrin blockade could lead to changes in the way lymphocytes interact with the intestinal tissue. Considering the abovementioned findings, it is possible that natalizumab’s therapeutic properties in multiple sclerosis could depend, at least in part, on these intestinal effects, besides those seen in blood–brain barrier, integrins and lymphocyte trafficking.

Fingolimod

Another drug that acts through the regulation of leucocyte trafficking is fingolimod, a functional antagonist of the sphingosine 1-phosphate receptor (S1P). S1P1 receptors are highly expressed on lymphocyte membranes and are critical for T and B cell egress from secondary lymphoid organs. S1P can affect the intestinal barrier by modulating tight junction proteins (Greenspon et al., 2011; Pászti-Gere et al., 2016), particularly under inflammatory conditions (Dong et al., 2015). For instance, fingolimod reduces endothelial barrier dysfunction in blood vessels and lung epithelium in experimental models of sepsis and haemorrhagic shock (Lundblad et al., 2013; Bonitz et al., 2014). Fingolimod also sequesters and alters the activation of lymphocytes in intestinal tissues (Chiba et al., 1998; Yanagawa et al., 1998; Henning et al., 2001; Halin et al., 2005; Sugito et al., 2005; Daniel et al., 2007), an effect thought to be mechanistically relevant in multiple sclerosis therapeutics. In the mouse EAE model, development of EAE was associated with increased accumulation of T cells in Peyer’s patches, a process increased by fingolimod (Spirin et al., 2014). Fingolimod can also directly affect the microbiota. Both sphingosine and fingolimod inhibit C. perfringens growth and endotoxin production in vitro, suggesting an intrinsic antibacterial property (Rumah et al., 2017).

Dimethyl fumarate

Dimethyl fumarate (DMF) is derived from the simple organic acid fumaric acid, and it acts as an immunomodulator by promoting T cell apoptosis, shifting to a Th2 response and acting as an antioxidant. There is limited but interesting evidence suggesting DMF could beneficially affect both the intestinal barrier and the gut microbiota. DMF alleviates experimentally induced colitis and reduces the Th1 response in mouse models and protects human intestinal epithelial cells against oxidative barrier dysfunction by preserving zona occludens-1 and occludin expression in vitro (Casili et al., 2016). DMF also preserves intestinal mucosa morphology after mycotoxin exposure and decreases intestinal permeability by strengthening tight junctions (Ma et al., 2017). In this model, DMF also led to increased microbiome diversity, with more abundance of bacteria producing SCFAs, such as Gemella, Roseburia, Bacillus and Bacteroides. DMF can also directly reduce C. perfringens growth and exhibits anti-mildew and antibacterial properties (Ma et al., 2017; Rumah et al., 2017).

Alemtuzumab

Alemtuzumab is an anti-CD52 antibody that causes depletion of mainly lymphocytes and is highly effective in the clinical management of multiple sclerosis (Hartung et al., 2015). Despite its specific mechanism of action, there is evidence suggesting it has detrimental effects over the integrity of the intestinal barrier and might alter the gut microbiome.

In mice, anti-CD52 antibodies induce increased intestinal barrier permeability (Qu et al., 2009) and lead to reductions in epithelial cell populations and to altered tight junction ultrastructure (Shen et al., 2013, 2015). In macaques, alemtuzumab-induced intestinal barrier disruption is associated with epithelial cell apoptosis as well as with increased circulating levels of d-lactate and endotoxin, indirect markers of intestinal barrier breakdown and bacterial translocation (Li et al., 2011; Qu et al., 2015). Lymphocyte depletion with alemtuzumab treatment in macaque models also resulted in dramatic changes in the gut microbiota (Li et al., 2010). Lactobacillales, Enterobacteriales, Clostridiales, and the genus Prevotella and Faecalibacterium were primarily responsible for the variations of the gut microbiota after lymphocyte depletion (Li et al., 2013). The diversity of fungal microbiota was similarly affected (Li et al., 2014). Despite this preclinical evidence, alemtuzumab-induced intestinal barrier disruption is infrequent in clinical practice. However, a case of spontaneous pancolitis was described in a patient with multiple sclerosis treated with alemtuzumab recently (Vijiaratnam et al., 2016), and historically, the use of alemtuzumab in haematological malignancies has been associated with the development of diarrhoea and opportunistic intestinal infections (Goteri et al., 2006; Ronchetti et al., 2014).

Teriflunomide

Teriflunomide selectively and reversibly inhibits dihydroorotate dehydrogenase, leading to a reduction in the number of activated lymphocytes that enter the CNS (Miller, 2015). Teriflunomide could alter the microbiome and the host response to enteral pathogens. Treatment of porcine intestinal epithelial cells with teriflunomide led to reduced capacity to fight bacterial infection through suppression of STAT-6 signalling (Yi et al., 2016). Teriflunomide could also directly inhibit C. perfringens growth in vitro (Rumah et al., 2017). Animals treated with teriflunomide in a mouse model of EAE had fewer antigen-presenting cells in Peyer’s patches as well as an increase in gut-specific CD39(+) Treg cells that could protect against EAE when used in an adoptive transfer regimen (Ochoa-Repáraz et al., 2016).

Minocycline

Minocycline is a second-generation tetracycline that was first introduced over half a century ago. Besides its antibiotic effects, it also has anti-inflammatory, immune-modulating and anti-apoptotic properties, all of which have been proposed as possible pathways towards neuroprotection (Yong et al., 2004; Giuliani et al., 2005). A recent randomized, double-blind, placebo controlled trial showed that oral minocycline could delay the appearance of a new demyelinating events in patients with clinically isolated syndrome, as well as reduce the appearance of T₂ lesions in the brain (Metz et al., 2017).

Minocycline’s immune-modulating and anti-inflammatory properties have also been observed in intestinal tissues. In a chemically-induced colitis model in mice, minocycline reduced intestinal inflammation, mucosal injury, restored microbiota and preserved tight junction protein expression (Huang et al., 2009; Garrido-Mesa et al., 2011a, b). As an antibiotic, minocycline also alters the gut microbiome. A recent study evaluated the effects of various commonly used antibiotics, including minocycline, on the salivary and gut microbiome in 66 healthy adults. Antibiotic exposure led to reductions in health-associated butyrate-producing species as well as proliferation of potentially resistant strains in the gut microbiome, although the changes were more robust after amoxicillin and ciprofloxacin administration (Zaura et al., 2015). Other studies have shown that some gut commensals such as Bifidobacteria and E. coli are susceptible to minocycline (Moubareck et al., 2005; Kirchner et al., 2014). Minocycline thus presents an intriguing option in dual modulation of the intestinal barrier function. It could have protective anti-inflammatory properties while also altering the composition of the gut microbiome.

Treating the diseased intestinal barrier

Current treatments for a diseased intestinal barrier are limited, but there are various interesting avenues of research. One of the main therapeutic targets are tight junctions. Larazotide acetate, also known as AT-1001, is a synthetic octapeptide related to the zonula occludens toxin produced by Vibrio cholera, developed as a treatment for coeliac disease. It acts locally to decrease tight junction permeability by blocking zonulin receptors and thus preventing actin rearrangement in response to stimuli, and in vitro it can stabilize tight junctions and decrease intestinal permeability (Paterson et al., 2007; Gopalakrishnan et al., 2012; Khaleghi et al., 2016). However, clinical trials in coeliac disease have yielded conflicting results, despite showing a beneficial effect over intestinal permeability (Kelly et al., 2013; Leffler et al., 2012, 2015).

Another approach in improving intestinal barrier function is enrichment of the mucus layer, a strategy being explored in IBD (Stange, 2017). Lecithin, or phosphatidylcholine, accounts for the majority of the phospholipids in the intestinal mucus layer, and is available as a delayed release oral formulation. In randomized phase II controlled studies, delayed-release lecithin was proven to be clinically and endoscopically effective in ulcerative colitis, and phase III studies are underway (Stremmel and Gauss, 2013; Stange, 2017). Recent interest has also been placed on stem cell-based therapies to regenerate the intestinal epithelium, through luminal transplantation (Holmberg et al., 2018), but these approaches are still in an experimental phase.

There has also been recent interest in the effects of vitamin D over intestinal barrier function and immune homeostasis (Dimitrov and White, 2017). In a model of experimental colitis, mice overexpressing vitamin D receptor in the intestinal epithelium show preserved intestinal permeability, reduced caspase expression and less induction of apoptosis (Liu et al., 2013). Vitamin D also attenuates TNFα-induced apoptosis in human colonic cells through reduction of NF-κB activation and mucosal IKK kinase activity, thereby preserving barrier function (see Li et al., 2015 for a review). Vitamin D signalling also preserves the mucosal barrier integrity by abrogating myosin light chain kinase dependent tight junction dysregulation during colonic inflammation through suppression of NF-κB in vitro (Du et al., 2015). Cultured colonic samples from patients with ulcerative colitis have altered expression of the tight junction claudin as well as increased pro-inflammatory cytokine expression; these changes were reversed by incubation with vitamin D (Stio et al., 2016). A recent small, randomized and placebo controlled study reported improvements in intestinal permeability [assessed by excretion of oral sugars (lactulose and mannitol were used as markers of small intestine permeability, sucrose as a marker of gastro-duodenal permeability, and sucralose as marker of combined small- and large-bowel permeability)] as well as serum immune markers in patients with IBD after vitamin D treatment (Raftery et al., 2015). Vitamin D appears to be important in the regulation of the intestinal barrier function, a mechanism not yet thoroughly evaluated in multiple sclerosis research.

Probiotics have emerged as an interesting option in regulating intestinal barrier function, fuelled by research in both in vitro and in vivo models that show that some microbiota can stabilize the intestinal barrier (Bron et al., 2017). However, small clinical studies in necrotizing enterocolitis, irritable bowel syndrome and IBD have shown only modest effects. There are no large randomized, placebo controlled studies and there is no obvious standardization of the quantities and composition of a given therapeutic probiotic ‘agent’, making trials difficult (Bron et al., 2017). There has been growing interest in the use of faecal microbiota transplantation (the ultimate microbiome modification) for the treatment of patients with chronic gastrointestinal infections and IBD (Smits et al., 2013), with excellent results observed in C. difficile colitis. It is also a safe procedure. Its effectiveness in autoimmune diseases and multiple sclerosis is unknown at this time.

There are other sources of interest in probiotics in multiple sclerosis. Probiotic administration is known to modulate the immune response in the mouse EAE model. Different formulations have been shown to reduce EAE duration (Ezendam et al., 2008), inhibit the pro-inflammatory Th1/Th17 polarization (Kwon et al., 2013), induce IL-10 producing Treg cells (Ochoa-Repáraz et al., 2010a, b; Takata et al., 2011) and enhance CD103 expression in dendritic cells (Ochoa-Repáraz et al., 2010b), all while preventing, delaying or attenuating EAE. E. coli strain Nissle 1917 has been shown to reduce EAE-induced intestinal barrier dysfunction, while also reducing disease severity and beneficially modifying T cell functions (Secher et al., 2017).

Despite these encouraging studies, few clinical trials have been performed using probiotics in multiple sclerosis. In one early trial, investigators used the non-pathogenic helminth Trichuris suis (Fleming et al., 2011). Five newly diagnosed patients with relapsing-remitting multiple sclerosis were given T. suis orally for 3 months, and favourable trends were seen in MRI outcomes (reduction in enhancing lesions from baseline) and immunological assessments (increased IL-10). A recent double-blind, placebo-controlled trial randomized 60 multiple sclerosis patients to receive a probiotic capsule or placebo for 12 weeks (Kouchaki et al., 2017). Probiotic treatment mildly improved Expanded Disability Status Scale (an absolute 0.4-point difference) and depression and anxiety symptoms, reduced high-sensitivity C-reactive protein and improved other metabolic measures such as insulin sensitivity and high-density lipoprotein-cholesterol levels. Probiotics also downregulated the gene expression of some pro-inflammatory cytokines in patients’ peripheral blood-derived mononuclear cells (Tamtaji et al., 2017). In these studies, the treatment was safe and tolerable, but follow-up was too short to show any meaningful benefit in radiological or clinical outcome measures. Nonetheless, the encouraging results seen in the EAE model will surely promote further clinical trial development.

SCFAs are bacterial fermentation products from indigestible diet components. The most common SCFAs are acetate, propionate and butyrate. SCFAs could have a beneficial effect over the intestinal barrier. Butyrate was shown to be able to accelerate tight junction protein assembly and preserve permeability in a single enterocyte layer in vitro model, a process mediated by AMP-activated protein kinase activity (Peng et al., 2009). SCFAs could also increase prostaglandin-dependent mucin expression in intestinal epithelial cells, enhancing their mucoprotective properties (Willemsen et al., 2003). In an EAE model, dietary SCFA ameliorated the course of EAE through expanded Treg cell populations in the lamina propria, through suppression of the JNK1 and p38 pathway (Haghikia et al., 2015). CD44 knockout mice that show attenuated EAE also have increased microbiota diversity and SCFA production in the gut (Chitrala et al., 2017).

Dietary interventions that increase the availability of SCFAs and reduce other types of fatty acids could be an interesting therapy in improving the intestinal barrier function in multiple sclerosis, with the additional possibility of beneficial immunological effects. However, evidence showing a benefit for any kind of dietary interventions in multiple sclerosis is scarce, despite widespread acceptance that a ‘healthy’ diet is probably best (Altowaijri et al., 2017; Esposito et al., 2017). Some probiotic species are also rich sources of SCFAs, suggesting the possibility of a combination approach.

Concluding remarks

The recent interest in the role of the gut microbiota in multiple sclerosis has not been accompanied by a similar interest in the intestinal barrier. The intestinal barrier is the physical and functional zone of interaction between the luminal microbiome and the organism, and it is also responsible for modulating multiple biochemical processes and immune modulation of the mucosa. It appears that besides dysbiotic changes in the gut microbiome, the intestinal barrier function is also altered both in EAE models and in patients with multiple sclerosis, but the precise consequences of this alteration are unclear. Evidence of CNS demyelination in gastrointestinal disorders where there is barrier breakdown and basic studies showing how the intestinal barrier homeostasis can directly influence microglia and neuroinflammation provide some insights. Furthermore, most disease-modifying therapies appear to also impact on the intestinal barrier and the gut microbiome. To advance the understanding of this complex interaction, future studies will have to take into consideration the microbiome, the intestinal barrier and the downstream neuroimmunological changes to accommodate for them in a single integrative model. Both the precise mechanisms involved in the breakdown of the intestinal barrier, and the value, if any, of therapeutic modulation of the intestinal barrier in multiple sclerosis, also require further study.

Funding

C.R.C.L. holds the Lejoie-Lake Fellowship awarded by the Hotchkiss Brain Institute, K.A.S. is the Crohn’s and Colitis Canada Chair in IBD Research at the University of Calgary, while V.W.Y. is a Canada Research Chair (Tier 1). The authors acknowledge operating grant support from the Canadian Institutes of Health Research, the Multiple Sclerosis Society of Canada, and the Alberta Innovates – Health Solutions CRIO Team program.

Abbreviations

Abbreviations
 
• EAE

  experimental autoimmune encephalomyelitis

 
• IBD

  inflammatory bowel disease

 
• LPS

  lipopolysaccharide

 
• SCFA

  short chain fatty acid

References