Gut Barrier Dysfunction and Type 2 Immunity: Implications for Compulsive Behavior

Abstract

To date, much of the focus of gut-brain axis research has been on gut microbiota regulation of anxiety and stress-related behaviors. Much less attention has been directed to potential connections between gut microbiota and compulsive behavior. Here, we discuss a potential link between gut barrier dysfunction and compulsive behavior that is mediated through “type 2” rather than “type 1” inflammation. We examine connections between compulsive behavior and type 2 inflammation in Tourette syndrome, obsessive-compulsive disorder, autism, addiction, and post-traumatic stress disorder. Next, we discuss potential connections between gut barrier dysfunction, type 2 inflammation, and compulsive behavior. We posit a potential mechanism whereby gut barrier dysfunction-associated type 2 inflammation may drive compulsive behavior through histamine regulation of dopamine neurotransmission. Finally, we discuss the possibility of exploiting the greater accessibility of the gut relative to the brain in identifying targets to treat compulsive behavior disorders.

1. Introduction

1.1. Compulsive behavior is a core component of several neuropsychiatric disorders.

Compulsive behaviors are defined as repetitive actions driven by obsessive thoughts and urges [1]. They are a defining feature of obsessive-compulsive disorder (OCD) [1], but similar habitual behaviors are also hallmarks of many other neuropsychiatric disorders—from tic disorders such as Tourette syndrome (TS) [2], to behavioral addictions such as compulsive eating [3]. Addictions, particularly those involving substance abuse, have been classified as compulsive behavior disorders, as the rewarding attributes of addictive behaviors diminish over time, but the behavior persists [4]. Restricted and repetitive behaviors also constitute one of the three core diagnostic criteria for autism spectrum disorder (ASD) [1]. Stress-related disorders, such as post-traumatic stress disorder (PTSD), may also be highly linked to compulsive behavior disorders: as many as 22% of subjects with PTSD have been diagnosed with persistent OCD [5], and such subjects have worse PTSD outcomes [6]. Vice versa, as many as 82% of OCD patients report a history of trauma [7].

The Research Domain Criteria (RDoC) initiative of the National Institutes of Health, aims to identify mechanisms underlying behavior, from normal to disordered, by categorizing behavior into fundamental bio-behavioral constructs that share neural mechanisms and circuitry [8]. Given the shared cognitive, neurochemical, and neuroanatomical correlates underlying compulsive behaviors found in OCD and other behavioral addictions, compulsivity has been proposed as a fundamental construct for RDoC, as highlighted by the high co-morbidity of OCD with various addiction and tic disorders described in Table 1 [9]. For example, the dorsal striatum and posterior hypothalamus, which contains histaminergic neurons that may serve as a link between peripheral inflammation and compulsive behavior (see Section 5), demonstrate pathology in TS [10, 11], ASD [12–14], obesity [15–17], alcoholism [18–20], and PTSD [21–23]. Currently, compulsive behavior is an element of the “Habit” construct in the RDoC classification scheme, which specifies potentially pathological, non-goal-directed habitual behavior [8, 24]. Habitual behaviors can be both adaptive and maladaptive, and the proclivity to encode and express habitual behaviors may have evolved as a part of several host defense mechanisms. For example, stress has been shown to inhibit goal-directed behavior in favor of habit learning in rodents as well as humans [25, 26]. Stressful situations may promote the development of adaptive avoidance against omnipresent threats, and the habitual employment of these avoidance behaviors preserves energy and saves time in avoiding aversive situations. However, in compulsive behavior disorders, such as OCD, these sorts of aversive responses persist even in the absence of aversive stimuli. For example, persons with OCD who were trained to make a physical response to an aversive shock stimulus paired with a cue were unable to suppress responding despite removal of the shock stimulus, even when the inability to receive continued shock was apparent to the participant [27]. These maladaptive responses may be observed across a number of stress-related disorders. Since compulsive behaviors are core diagnostic criteria or complicating factors across a large number of psychiatric disorders, treatment approaches that seek to ameliorate compulsive behavior may have a significant impact on a wide spectrum of mental health conditions.

Table 1:

Co-morbidity of obsessive-compulsive disorder (OCD) and other neuropsychiatric disorders

Condition	Subject	N of sample	Age	Outcome	Reference
Tourette’s Syndrome (TS)	Global sample across 80 “Tourette Syndrome International Database Consortium” sites.	6,805; Analyses split between individuals with TS with and without co-morbid ADHD	All ages, (Mean age not reported)	21% of children with OCD and 37% of adults with OCD also diagnosed with TS with ADHD:. 16% of children with OCD, 29% of adults with OCD also diagnosed with TS without ADHD:	[28]
Autism Spectrum Disorder (ASD)	Meta-analysis of 31 studies	2,121 ASD subjects over 31 studies	Mean age ranged from 4.2 to 16.3	17.4% of ASD subjects also diagnosed with OCD	[29]
Obesity	Subjects with OCD	104 OCD subjects	Mean age 39.5	36.5% of OCD subjects also diagnosed with abdominal obesity	[30]
Alcoholism	Alcoholic patients in Brazilian outpatient clinic	54 alcoholic patients	Mean age 49.5	20.4% of alcoholic patients presented with “probable OCD”	[31]
Post-Traumatic Stress Disorder (PTSD)	Cross-section of Australia	10,641 individuals in general population, of which 0.5% diagnosed with OCD and 0.9% with PTSD	18+, (Mean age not reported)	19% of OCD subjects OCD were diagnosed with PTSD, and 34% of subjects with OCD were diagnosed with PTSD	[32]

Representative studies demonstrating significant co-morbidity of obsessive-compulsive disorder (OCD) in various tic and addiction disorders.

1.2. Hypothesis

One potential site of intervention for compulsive behavior disorders is the gut microbiome. To date, a smattering of reviews have spotlighted connections between dysregulation of gut microbiota and compulsive behavior [33–35]. Here, we hypothesize a mechanistic link, reviewing studies that show that gut barrier dysfunction, driven by dysregulated gut microbiota, may contribute to compulsive behaviors. We explore several conditions that exhibit comorbidities of gut barrier dysfunction and compulsive behaviors, such as compulsive eating in obesity [36, 37], compulsive drinking in alcoholism [38, 39], OCD in PTSD [5, 40–42], and repetitive behaviors characteristic of ASDs [43, 44]. Chronic low-grade inflammation triggered by gut barrier dysfunction observed in each of these conditions may contribute to the expression of compulsive behaviors. Gut barrier dysfunction is associated with “type 2” (Th2) inflammation [45], as opposed to “type 1” (Th1) inflammation. Briefly, type 2 inflammation is the allergic immune response, primarily against environmental toxins and large eukaryotic parasites, whereas type 1 inflammation is the canonical immune response to most bacterial and viral pathogens [46]. The inflammatory profiles that define Th1 and Th2 inflammatory states are discussed in Section 2. We propose that an increase in Th2 inflammation, particularly that driven by gut barrier dysfunction, may stimulate compulsive behaviors.

Most of the literature surrounding the gut-brain axis concerns stress, anxiety behavior, and cognition [47]. Given the high prevalence of compulsive behavior across several psychiatric conditions and the new categorical distinction of obsessive-compulsive and related disorders as separate from anxiety disorders [1], a focus on how the gut-brain axis affects compulsive behaviors is needed. This review aims to spark further research into whether dysbiosis of the gut microbiota and concomitant gut barrier dysfunction affects the incidence and severity of compulsive behaviors. Specifically, we discuss connections between type 2 inflammation and compulsive behavior in TS, OCD, ASD, obesity, alcoholism, and PTSD. We explore a potential mechanism whereby gut dysbiosis-induced type 2 inflammation may drive compulsive behavior by altering brain histamine receptor activation in key brain regions postulated to drive compulsive behavior. Finally, we discuss the potential for developing gut-based therapies to mitigate compulsive behavior in these disorders.

2. Behavioral effects of immune polarization

2.1. Compulsive Behavior As A Component of the Allergic Immune Response

Compulsive behaviors may be a component of the behavioral arm of the allergic immune response, which likely evolved as a defense strategy against environments harboring noxious substances and ectoparasites. Allergens are non-infectious, noxious substances that trigger the type 2 immune response, generally recognized to provide protection against multi-cellular parasites, transmitted by, for example, insects such as tics. Until recently, it was believed that these allergens, which often provoke an immune response at very low doses, simply triggered a mis-firing of an immune response. However, these allergic immune responses may assist with environmental surveillance [48]. Exposure to allergens have been shown to increase anxiety behavior in rodents, promoting avoidance of anxiogenic environments that may contain more of the noxious substance [48, 49]. There are other behavioral patterns that also serve to limit exposure to allergens. Repetitive behaviors such as repetitive digging, indexed by the marble burying test, may also limit mammalian exposure to noxious foods [50]. This defensive burying response may be provoked by delayed toxicosis, such as when a rat is administered a poisonous dose of lithium chloride following self-administration with sweetened milk, or by immediate exposure to noxious stimuli, such as self-administration of pepper tobacco sauce [51]. Repetitive locomotor patterns, such as repeated entry into the same arm of a T-maze or repetitive circling behaviors, have also been induced by an allergic response to whey protein or by suppression of type-1 immune signaling, respectively [52, 53]. Absent additional exposure to noxious stimuli, these behaviors limit exploration into contiguous environments that may contain noxious stimuli that may further exacerbate the allergic immune response.

Grooming behavior is an innate behavior with several fixed-action patterns that also serves as to limit exposure to ectoparasites. Ectoparasites transmit microbial parasites that trigger type 2 inflammation [48]. As grooming behaviors are observed in rodent models of allergy/asthma, this behavior may have evolved as a part of the behavioral arm of the allergic immune response as well [52, 54]. Grooming behavior is observed to a higher degree in species inhabiting environments with a higher tic burden [55]. Grooming behavior is also seen at higher rates in females than in males, correlating with sex differences in the adult mammalian immune system, which is biased toward higher levels of allergic inflammation in females relative to males [55]. Thus, grooming behavior may have evolved as a behavioral defense strategy against allergenic parasites.

One of the primary obsessions of obsessive-compulsive disorder is fear of contamination, which also links compulsive behavior disorders with protection against allergens and allergenic microbes. Up to 50% of people with OCD present with concerns surrounding germs and contamination [56]. This corresponds with higher ratings of feelings of disgust, as opposed to anxiety, in OCD patients in response to symptom-provoking stimuli [57]. This is supported by inconsistent findings of brain regions associated with anxiety, such as the amygdala, showing activation during obsessive-compulsive episodes [58]. Disgust is a universal emotion believed to have evolved to provide protection against contamination and disease and is associated with the rejection of unpleasant tastes that serve to signal avoidance of pathogenic microbes and food toxins [57]. If compulsive behaviors evolved to provide protection against allergenic compounds and microbes, it stands to reason that compulsive behavior disorders would be more strongly associated with the emotion of disgust than anxiety. Indeed, this unique psychopathology of obsessive-compulsive disorder was recently in the 5^th version of the Diagnostic and Statistical Manual of Mental Disorders, which removed obsessive-compulsive disorder from the general category of Anxiety Disorders and placed it in its own category [1].

2.2. Type 2 inflammation in compulsive behavior disorders

The host immune system is capable of mounting polarized immune responses to either intracellular or extracellular pathogens, foreign particles, or cellular damage (collectively referred to as “foreign threats”) [59, 60]. This is characterized by activation of helper T lymphocytes (helper T cell 1 and 2, or “Th1” and “Th2”) that produce different cytokine profiles, which control distinct immune responses. These immune responses recruit components of both innate and adaptive immunity, and are thus more broadly labelled as “type 1” vs “type 2” immune responses [59]. Th1 inflammation targets intracellular pathogens and drives acute immune responses, whereas Th2-associated inflammation drives allergic and chronic inflammation [59]. Activation of the Th1 immune response tends to downregulate Th2 immunity, and vice versa [59, 61]. Therefore, while inflammatory diseases typically recruit both arms of the immune system, in most cases either Th1 or Th2 immune responses predominate. Other players outside of the Th1 and Th2 immune responses drive immunity, however cytokines associated with Th2 immunity contribute to non-Th2 cell responses that support allergic inflammation. For example, IL-6, a cytokine that drives Th2 responses, also promotes Th17 immune responses [62]. However, Th17 cells contribute to allergic/atopic inflammation (predominantly Th2-driven) as well and promote mast cell proliferation [63]. Core differences in cell recruitment, cytokine signaling, and pathology between Th1 and Th2 immune reactions are well-explained by Saini et al., 2011 and Kidd, 2003 [59, 64] and the balance between Th1 and Th2 inflammation is diagrammed in Figure 1.

Figure 1: Balance between type 1 and type 2 immunity in the gut.

Factors that promote gut barrier dysfunction promote pro-allergic (Th2 or “type 2”) inflammation, such as pathobionts, which invade the gut barrier and downregulate intestinal epithelial tight junction expression. Type 2 cytokines IL-4 and IL-33, and type 2 mediator histamine, promote both intestinal and systemic type 2 inflammation. Type 1 inflammation (characterized by cytokines IFN-gamma and IL-12) counter-regulates type 2 inflammation, and probiotic bacteria, such as Lactobacillus, both increase tight junction expression and increase immune bias toward type 1 inflammation.

Interestingly, allergic diseases exhibit high rates of co-morbidity with compulsive behavior disorders. Patients with TS and OCD often exhibit increased levels of Th2 inflammation. A nationwide study in Taiwan revealed that subjects with allergic rhinitis have double the risk of TS [65]. A similar study in Germany revealed a 60% higher incidence of allergies in OCD subjects relative to healthy controls [66]. Smaller clinical studies confirm links of allergy with OCD and TS. For example, in one such study, OCD subjects showed a greater frequency of positive allergic skin prick tests compared to TS subjects or healthy controls. In addition, half of the OCD and TS subjects were diagnosed with allergies, compared to only one-fifth of healthy controls [67].

ASD subjects also demonstrate high rates of allergies and other atopic diseases (for review, see [68]). For example, ASD subjects had significantly higher rates of allergies and more than twice the rate of psoriasis compared to control subjects [69]. In addition, a cross-sectional study of data derived from the National Survey of Children’s Health revealed a high prevalence of ASD in children with atopic dermatitis [70]. Atopic diseases, characterized by allergic inflammation, may be a risk factor for ASD, as a longitudinal study found that atopic disease in early childhood (before the age of three) significantly increased the risk of ASD diagnosis later in life. Furthermore, children diagnosed with more than one atopic disease had a greater risk of developing ASD than children with only one atopic disease [71].

Sex differences in rates of allergic diseases parallel sex differences in the incidence and severity of compulsive behavior across the life span. A systemic review found that overall boys have a higher incidence of food allergies than girls (64% boys vs 36% girls). This male/female ratio reverses after puberty, where 35% of men and 65% of women are reported to have food allergies [72]. A similar reversal in disease incidence is observed in OCD, with males largely having a childhood age-of-onset and females an onset in adulthood [73]. While boys represent about 70% of OCD patients in childhood, studies tend to report an equal incidence in men and women or a slight elevation in women [73, 74]. TS is more frequent and severe in boys than girls, but in adulthood TS appears to be more severe in women than in men [75]. Similar sex differences extend to other psychiatric disorders featuring compulsive behavior. The reported ratio in incidence of ASD is 4 to 5 boys to 1 girl [76]. Directly comparing male and female ASD subjects, males tend to exhibit greater frequency of restricted interests and repetitive behaviors than do females [76]. These correlations between sex differences in immunity and sex differences in the incidence of TS, OCD, and ASD are consistent with a role of the immune system in compulsive behavior.

2.3. Type 2 Inflammation Promotes, while Type 1 Inflammation Suppresses, Habitual Behavior in Rodents

Animal studies corroborate correlations between allergic disease and compulsive behavior disorders. For example, different rodent strains exhibit differing levels of compulsive-like behaviors. The following paragraphs will provide some examples.

The C57Bl/6 mouse strain typically exhibits a Th1 immune profile, whereas Balb/c mice typically display a Th2-biased immune profile [77]. C57Bl/6 mice produce more IFN-gamma and activated macrophages in response to both extracellular and intracellular pathogens than do Balb/c mice [77]. In addition to differences in stimulated immunity, there are differences in baseline immunity, given higher intestinal levels of histamine in unchallenged Balb/c mice relative to unchallenged C57Bl/6 mice [78]. These mouse strains also demonstrate behavioral differences: C57Bl/6 mice are more social than Balb/c mice [79], which has been suggested to be due to differences in the immune profile between the two strains [80]. In the marble burying test, there is an apparent trend of higher burying behavior in Balb/c mice relative to C57Bl/6 mice [50]. If this difference is accentuated by allergic or other immune stimuli in Balb/C mice relative to C57Bl/6 mice, this may further implicate Th2 inflammation in burying behavior.

Germ-free mice also represent another example of a Th2-biased mouse model. Resident microbiota are critical to the normal development of the immune system, and germ-free mice maintain exaggerated baseline serum levels of the primary Th2-biasing cytokine, IL-4, throughout adulthood [81]. Germ-free mice also show slightly higher levels of marble burying relative to conventionally colonized mice [82]. Marble burying is a marker of non-goal directed, compulsive-like, repetitive digging, as the presence of marbles does not elicit an increase in digging behavior [50]. Furthermore, germ-free mice exhibit greater levels of repetitive self-grooming during social interactions than do conventionally colonized mice [83]. While germ-free mice exhibit decreased anxiety-like behavior on exploration-based assays, such as the elevated plus maze and light/dark box tests [84, 85], high levels of grooming correlate with low social interaction in what has been posited as a mouse model of social anxiety [86]. As germ-free mice exhibit less time investigating novel conspecifics, the stress induced by social interactions with conventionally colonized conspecifics may elicit grooming behavior. Stressful stimuli may induce the release of histamine from mast cells [87, 88], and as discussed in section 4.2, this may contribute to the presentation of repetitive behaviors in germ-free mice.

Brattleboro rats exhibit an impaired immune system that may contribute to their low levels of compulsive-like behavior. Brattleboro rats are Long Evans rats with a missense mutation in the arginine-vasopressin gene, resulting in the development of diabetes insipidus [89]. In addition, these rats exhibit low numbers of blood lymphocytes, reduced macrophage activity, and low antibody production [90]. This suggests that Brattleboro rats may demonstrate attenuations of both Th1 and Th2 immune responses. It is plausible that Brattleboro rats exhibit low systemic levels of histamine and low levels of intestinal permeability. Brattleboro rats exhibit heightened levels of Lactobacillus, which have been demonstrated to decrease intestinal permeability [91]. Heightened plasma levels of vasopressin promote mortality and intestinal permeability in a model of colitis [92]. These effects are blocked by a mast cell stabilizer, suggesting that vasopressin may act directly on mast cells to induce degranulation [92]. Brattleboro rats have low levels of brain histamine relative to WT Long Evans rats [93]. This may reflect basal levels of histamine constitutively released from brain mast cells in WT Long Evans rats. These immune changes may be related to behavioral differences observed between Brattleboro and WT Long Evans rats. Brattleboro rats exhibit abnormal social and behavioral development [94, 95]. In addition, they exhibit low levels of anxiety-like behavior as measured on the elevated plus maze and low levels of compulsive-like behavior as measured on the marble burying test [96, 97].

Conditions that exhibit a bias toward type 1 immunity are correlated with suppression of repetitive behaviors. Experimental Autoimmune Encephalitis (EAE) is a rodent model of multiple sclerosis with a strong involvement of Th1 and Th17 cells [98, 99]. For example, while transfer of Th2 cells from EAE animals into a non-EAE host does not transfer autoimmune pathology to the host, transfer of Th1 cells does produce EAE pathology in the host [100–102]. EAE mice demonstrate reductions in nestlet shredding, a marker of perseverative behavior [103]. Of note, even mild induction of EAE, which does not induce significant motor deficits, suppresses nest building activity [104]. Another treatment that enhances type 1 inflammation in rodents is administration of the drug leuprolide: a gonadotropin-releasing hormone (GnRH) agonist. GnRH is a potent Th1 inducer [105, 106]. Leuprolide both prevents and decreases enhanced marble burying behavior observed during ethanol withdrawal in mice [107]. Taken together, these rodent studies suggest that type 1 and type 2 inflammation may have opposing effects on compulsive behavior. These and related studies are outlined in Table 2.

Table 2:

Side by Side Comparison of Th1 and Th2 Inflammation and Corresponding Effects on Compulsive Behavior.

Experimental Endpoints	Subjects with Th1 bias	Subjects with Th2 bias	References
	C57Bl/6J mice	Balb/C mice
Inflammation Markers	Lower intestinal histamine levels	Higher intestinal histamine levels	[78]
Compulsive Behavior	Lower marble burying behavior	Higher marble burying behavior	[50]
	Conventionally-colonized mice	Germ-free mice
Inflammation Markers	Lower IL-4 and higher IFN-gamma serum levels	Higher IL-4 and lower IFN-gamma serum levels	[81]
Compulsive Behavior	Higher marble burying behavior	Lower marble burying behavior	[82, 83]
	Long Evans rats	Brattleboro rats
Inflammation Markers	Lower brain histamine levels and lower gut levels of Lactobacillus	Higher brain histamine levels and higher gut levels of Lactobacillus	[93, 108]
Compulsive Behavior	Lower marble burying behavior	Higher marble burying behavior	[96, 97]
	Sham-sensitized mice	Mice induced to have allergic response to whey protein
Inflammation Markers	No change in serum mMCP (marker of mucosal mast cell degranulation) 16 hours after whey challenge	Elevated serum mMCP 16 hours after whey challenge	[52]
Compulsive Behavior	Decreased grooming and elevated spontaneous alterations on T-maze	Increased grooming and reduced spontaneous alterations on T-maze	[52]
	Subjects injected with saline	Subjects injected with G-CSF
Inflammation Markers	-	Case studies: Clinical administration of G-CSF (in humans) induces anaphylactoid response	[109]
Compulsive Behavior	No changes in cocaine-induced locomotion or cocaine-seeking behavior	Enhanced cocaine-induced locomotion and cocaine-seeking behavior	[110]
	Experimental autoimmune encephalitis rodent model of multiple sclerosis	Control mice
Inflammation Markers	Injection of Th1 clones reactive to MBP into healthy mice required for transfer of autoimmune pathology, whereas injection of Th2 clones reactive to MBP does not transfer pathology	Caveat: Injection of unreactive T cells not used as a negative control.	[100]
Compulsive Behavior	Suppressed nestlet shredding and nest building behavior, without affecting locomotion	Normal levels of nestlet shredding and nest building behavior.	[103]
	Mice treated with gonadotropin-releasing hormone (GnRH) or drugs that promote GnRH release	Control mice
Inflammation Markers	Elevated blood IFN-gamma levels in mice treated with GnRH	Lower blood IFN-gamma levels	[105]
Compulsive Behavior	Decreased marble burying behavior in mice treated with leuprolide, a drug that promotes GnRH release	Higher marble burying behavior in control mice	[107]

Abbreviations: G-CSF, granulocyte colony stimulating factor; GnRH, gonadotropin releasing hormone; IFN-gamma, interferon-gamma; IL, interleukin; MBP, myelin basic protein; mMCP, mucosal mast cell protease; Th1, helper T cell type 1; Th2, helper T cell type 2.

3. Gut barrier dysfunction as a central mechanism across compulsive behavior disorders

3.1. Gut barrier dysfunction and compulsive behaviors

Given the strong correlations between allergic inflammation and compulsive-like behaviors, physiological processes that have a strong influence of allergic inflammation may actively influence compulsive behavior. Active regulation and disease-associated breakdown of the gut barrier is one of these processes, as the gut barrier is one of the largest sites whereby allergens may modify mammalian inflammation and behavior. While it is designed to be permeable to allow passage of food particles and important enterically-derived compounds, dysregulation of the gut barrier can pose important systemic threats and influence both intestinal and systemic inflammation that may have downstream effects on compulsive behavior [111]. In support of this hypothesis, gut barrier dysfunction is often observed to be associated with compulsive behavior disorders. Gut barrier dysfunction is a defining feature of inflammatory bowel disease (IBD), which includes Crohn’s disease and ulcerative colitis [112]. If gut barrier dysfunction drives the expression of compulsive behaviors, one would expect compulsive behavior to be elevated in IBD patients. While relatively little work has been done to characterize the comorbidity of IBD with OCD, there are case studies of abrupt onset of OCD-like symptoms in IBD patients coinciding with IBD flare-ups. For example, a boy with Crohn’s disease demonstrated sudden onset of sexually obsessive thoughts during a period when his symptoms were resistant to treatment [113]. In addition, a patient with ulcerative colitis exhibited symptoms of the OCD-like “bowel obsession syndrome”, including obsessional thoughts over bowel habits and excessive and unnecessary visits to the toilet [114]. OCD-like symptoms have been described for children with Crohn’s or ulcerative colitis and one clinical study described a 35% comorbidity of IBD in OCD patients [115]. Given these promising but still preliminary links between gut barrier dysfunction and OCD, more work on the overlap between IBD and OCD is warranted.

In ASD, there is evidence of increased permeability of the lining of the gut, which may promote Th2 inflammation and an associated increase in repetitive/compulsive behaviors. For example, a lactulose/mannitol intestinal permeability test identified increased intestinal permeability in patients with ASD [44], and ileal tissue sampled from children exhibiting regressive autism demonstrated significant levels of erosion and lymphocyte infiltration [116]. These ileal epithelial cells lacked HLA-DR, a cell surface receptor indicative of T cells, suggesting that the immune response was Th2-biased [116]. Treating the maternal immune activation mouse model of autism, which exhibits increased intestinal permeability, with a bacterial strain that reduces intestinal permeability, also normalized social and repetitive behaviors observed in this model [117]. These studies suggest that gut barrier dysfunction may contribute to repetitive behaviors observed in ASD and may represent an effective target for treatment.

Obese patients have increased intestinal permeability and may have increased risk for development of allergy/atopy, which may eventually contribute to compulsive eating. Measures of intestinal permeability via administration of lactulose and mannitol positively correlated with waist and abdominal circumference in a study that included lean and obese women [118]. This correlation is also found in mice, where a high fat diet increased intestinal permeability after just one day [119]. High fat feeding in mice also induced Th2 inflammation, as it promoted a 50- to 70-fold increase in baseline serum IgE levels [120]. While the initial hedonic aspects of highly palatable foods may stimulate addictive eating, the ensuing allergic inflammation, which is associated with high-fat feeding, may convert this addiction into compulsive eating.

Alcohol ingestion increases intestinal permeability as well, which likewise may contribute to compulsive drinking, and endotoxin derived from gut microbiota is implicated in the pathology of alcoholic liver disease [121, 122]. Chronic alcohol-fed mice also exhibited exaggerated Th2 inflammatory responses, showing increases in the production of IgE and Th2-biasing cytokines IL-4 and IL-10 in response to ovalbumin challenge relative to non-alcohol exposed mice [123].

While no studies have directly measured gut barrier function in individuals diagnosed with PTSD, stressful events have been found to increase intestinal permeability and anxiety behavior. For example, combat training increased anxiety and depression, raised serum cortisol levels, and increased intestinal permeability assayed by the mannitol-lactulose test [124]. In addition, acute psychological stress, in the form of public speech, increased both intestinal permeability and serum cortisol levels [41]. In rats, water avoidance stress decreased ex vivo measurements of electrical resistance across intestinal epithelia and reduced intestinal epithelial tight junction protein expression [125].

3.2. Gut barrier dysfunction contributes to Th2 inflammation

The recurrent observation of gut barrier dysfunction in conditions featuring compulsive behaviors begs the question of how a breakdown of the gut barrier influences these behaviors. The intestinal epithelium is a single layer of cells separating billions of microbes residing in the gut from interacting with the large compartment of intestinal immune cells located directly underneath the epithelial layer (the lamina propria). Any conditions that increase the permeability of this layer also increase access of pro-inflammatory microbial antigens to these immune cells as well as to immune cells within the systemic circulation, provoking both intestinal and systemic inflammation [126]. Gut microbiota are predominantly composed of Gram-negative and Gram-positive bacteria, which carry antigens that induce distinct host immune reactions [127].

Gut barrier dysfunction is often associated with chronic low-grade endotoxemia, also called “metabolic endotoxemia,” which is characterized by a small but chronic increase in blood lipopolysaccharide (LPS) levels [128]. LPS, or endotoxin, is a pathogenic component of the cell membrane of Gram-negative bacteria such as E. coli. Systemic injection of LPS in rodents is used to model the effects of acute infection on behavior. These injections induce a syndrome called “sickness behavior,” which manifests itself as decreased social behavior, increased anxiety behavior and anhedonia, and the occurrence of various sensorimotor deficits [129]. Depending on LPS dose, many of these effects persist even in the absence of lethargy. Marble burying, an index of compulsive behavior, is suppressed after a single systemic injection of LPS in mice, even after 24 hours, when locomotion effects of LPS have subsided [130]. However, these single, acute injections of LPS may not model the effects of chronic, systemic exposure to gut-derived LPS. Robust Th1 responses appear rapidly after LPS administration, whereas Th2 responses are delayed and longer-lasting. In one study, peripheral blood mononuclear cells (PBMCs) were obtained from healthy human subjects who had been injected with LPS. Samples collected from subjects directly after LPS injection secreted both Th1 and Th2 cytokines (after in vitro exposure to alphaCD3/alphCD28 for 24 hours), whereas those collected 3, 6 and 24 hours after an LPS injection show suppressed secretion of the Th1 cytokine IFN-gamma, while the levels of Th2 cytokines IL-4 and IL-5 peaked during this timeframe [131]. This finding was supported by an in vitro study showing that macrophages release Th1-influencing cytokines directly after LPS challenge, but mount a robust delayed release of Th2-promoting cytokines 48 hours after continuous LPS stimulation [132]. It is therefore plausible that chronic exposure to LPS, as observed in metabolic endotoxemia driven by gut barrier dysfunction, may promote a shift in innate immune response away from Th1-biasing and toward Th2-biasing immunity, which then in turn may stimulate compulsive behavior.

Other gut-derived microbial antigens may also promote the development of Th2 inflammation. While LPS predominantly activates toll-like receptor 4 (TLR4), which drives predominantly Th1 immunity, TLR2, which drives predominantly Th2 immunity, is activated by a wide range of microbial antigens including peptidoglycans found on Gram-positive bacteria [133]. Upon breach of the intestinal epithelium, TLR2-activating antigens may interact with various innate immune cells, including subepithelial dendritic cells and mast cells. Mast cells may differentially respond to TLR4 and TLR2 activation. For example, LPS from E. coli, which activates TLR4, evokes a Th1 response from dendritic cells, whereas LPS from Gram-negative Porphyromonas gingivalis, which activates TLR2, facilitates Th2 cytokine expression and suppresses the release of Th1 associated cytokines [134]. It is possible that intestinal and systemic TLR2 activation, by enhancing Th2 immunity, promotes the expression of compulsive behaviors in conditions exhibiting gut barrier dysfunction. In line with this hypothesis, Tlr2−/− mice show less marble burying behavior despite exhibiting increased locomotion in the open field test [135]. In addition, genetic deletion of Tlr4 heightens repetitive circling in mice, suggesting that in these mice activation of still intact toll-like receptors, such as TLR2 or other innate immune receptors, bias immune responses toward a Th2 cytokine profile and promote compulsive-like behaviors [53].

A final mechanism to consider by which gut barrier dysfunction can promote Th2 inflammation is via the production of IL-33 by damaged and necrotic intestinal epithelial cells [136]. IL-33 promotes the degranulation of mast cells and orchestrates the initiation of allergic immune responses (reviewed in [137]). IL-33 also promotes the expression of Th2 proinflammatory cytokines IL-6 and IL-4, fostering a systemic Th2 immune response that may reach the brain, promoting the degranulation of brain mast cells and subsequent neuroinflammation [138, 139].

Taken together, these studies demonstrate that gut barrier dysfunction can initiate Th2 inflammation. Figure 2 summarizes the model presented here, wherein pathological gut bacteria that penetrate the mucosal layer compromise the integrity of the intestinal epithelium through destruction of intestinal epithelial cells and downregulation of intestinal epithelial tight junction proteins. This sets off a cascade of inflammation, triggered by food allergens and cytokines produced by damaged intestinal epithelial cells (such as IL-33), that promotes a systemic allergic immune response. This immune response includes histamine, which can cross the blood-brain barrier and interact with histamine receptors expressed by compulsive behavior circuits. Table 3 highlights studies demonstrating connections between gut barrier dysfunction, Th2 inflammation, and compulsive behavior in ASD, obesity, alcoholism, and PTSD.

Figure 2: Model for how gut barrier dysfunction, mediated through type 2 inflammation, dysregulates dopamine circuits implicated in compulsive behavior.

Bacteria from the gut lumen penetrate the mucosal layer to interact with intestinal epithelial cells, downregulate intestinal epithelial tight junctions, and activate immune cells residing along the lamina propria, such as dendritic cells, mast cells, macrophages, T cells, and B cells. Dysbiosis, particularly the presence of bacteria that penetrate the mucosal lining, triggers the production of IL-33 by intestinal epithelial cells which promotes a type 2 bias in immune cells in the lamina propria. Inflammatory mediators produced by these cells travel through the bloodstream and lymphatic system before reaching the blood-brain barrier, while intestinal cytokines may also activate vagal nerves that, in turn, trigger neuroinflammation through direct stimulation of immune cells resident on the brain side of the blood-brain barrier, including brain-resident mast cells. Systemic histamine promotes increased permeability of the blood-brain barrier, and other inflammatory mediators that reach the blood-brain barrier can thereby penetrate and also activate brain-resident mast cells. Histamine directly decreases phasic dopamine release, decreasing reward plasticity and promoting the expression of compulsive behaviors.

Table 3:

Gut barrier dysfunction, type 2 inflammation, and compulsive behavior in rodent models of autism spectrum disorder (ASD), obesity, alcoholism, and post-traumatic stress disorder (PTSD).

Disorder	Animal Model	Test	Results	Reference
Gut Barrier Dysfunction
ASD	Maternal Immune Activation (MIA) in mice	FITC-dextran test	Increased intestinal permeability in offspring	[117]
Obesity	24hr high fat diet in mice	FITC-dextran test	Increased intestinal permeability	[119]
Alcoholism	6hr alcohol-administration in mice	FITC-dextran test	Intestinal permeability increased 3 and 6 hours after ad libitum ethanol exposure	[140]
PTSD	Exposure to water avoidance stress of Fisher-344 rats	Ex vivo transepithelial resistance measurement	Increased intestinal permeability	[141]
Type 2 Inflammation
ASD	BTBR mice vs C57Bl/6J mice	Blood cytokine and immunoglobulin testing	Higher IgG/IgE and IL-33 blood levels in BTBR vs WT mice	[142]
	Valproic acid treatment of pregnant mice	Splenocyte cytokine testing	Decreased splenocyte ratio of IFN-gamma to IL-10 in valproic acid offspring	[143]
Obesity	16-week high fat diet in mice	Blood immunoglobulin measurement in response to oral ovalbumin allergen test	50- to 70-fold increases in blood IgG1 and IgE in response to food allergen	[120]
Alcoholism	24hr alcohol-administration in mice	Blood immunoglobulin and cytokine measurements	Higher IgE, IL-4, and IL-10 blood levels	[123]
PTSD	Exposure to social defeat in mice	Blood cytokine testing	Elevated serum IL-6 levels	[144]
Compulsive Behavior
ASD	Maternal Immune Activation (MIA) in mice	Marble burying test	Increased marble burying	[117]
	BTBR mice vs C57Bl/6J mice	Marble burying test and grooming behavior	Increased stereotyped repetitive, marble burying, and grooming behavior in BTBR mice	[145]
	Valproic acid treatment of pregnant mice	Marble burying test and grooming behavior	Increased marble burying and grooming behavior	[146]
Obesity	Transfer of gut microbiota from high fat diet fed mice into antibiotictreated mice	Marble burying test	Increased marble burying behavior in mice reconstituted with microbiota from high fat diet-fed mice vs microbiota from normal chow-fed mice	[147]
Alcoholism	Four-week chronic intermittent alcohol exposure	Marble burying test	Increased marble burying	[148]
PTSD	Social defeat stress in mice	Observation of grooming behavior	Increased grooming behavior	[149]

Abbreviations: FITC-dextran, fluorescein isothiocyanate-conjugated dextran; IgE, immunoglobulin E; IgG, immunoglobulin G; IL, interleukin; MIA, maternal immune activation.

4. Sex differences in immunity correlate with sex differences in OCD, TS, and ASD

Sex differences in rates of allergic diseases overlap with sex differences in the incidence and severity of compulsive behavior across the life span. While most of the literature reporting sex differences in the clinical presentation of food allergies are biased toward the inclusion of either males or females, a systemic review found that studies tend to report a higher incidence food allegies in males than females among children (64.35% male vs 35.65% female). This male/female ratio reverses after puberty, where 34.82% of females and 65.18% of males are reported to have food allergies [72]. This timeline for sex differences in food allergy map onto broader sex differences rerported for prepubertal children and postpubertal children and adults [150]. Anaphylaxis is precipitated by the binding of allergen-specific immunoglobulin E (IgE) to IgE receptors expressed on mast cells, which causes degranulation and release of histamine and other inflammatory mediators. IgEs are produced in B lymphocytes, many of which reside in the gut immune system [151]. Both male fetuses and male infants exhibit more robust inflammatory reactions than do females [150, 152, 153]. In addition, cord blood from male neonates have higher levels of IgE than female neonates [150, 154]. However, the emergence of sex steroids in puberty reverses these differences in adulthood. Adaptive immunity becomes more robust in females. Adult females have higher numbers of lymphocytes and greater expression of adaptive immune genes, such as those that control the immunoglobulin expression [150].

A similar reversal in disease incidence is observed in OCD, where males largely exhibit a childhood age-of-onset, whereas females develop the disorder in adulthood [73]. Furthermore, males tend to exhibit a higher co-morbidity with tic and substance use disorders [73]. While boys represent about 70% of OCD patients in childhood, studies tend to report equal incidence of OCD between men or women or a slight increase in women [73, 74]. TS is more frequent and more severe in boys than girls, but in adulthood TS appears to be more severe in women than men [75]. This sex difference extends to other psychiatric disorders featuring compulsive behavior. As mentioned earlier, the reported sex difference in autism spectrum disorders (ASD), where expression of restricted and repetitive behaviors is one of the core diagnostic criteria, is 4–5:1 [76]. However, given differential application of diagnostic symptoms across genders, there is still uncertainty regarding the true sex difference in the incidence of ASD. Nevertheless, even directly comparing male and female ASD subjects, males tend to exhibit greater frequency of restricted and repetitive behaviors than do girls [76].

5. Histamine, the “missing link”?

The previously discussed studies indicate that Th2 inflammation is important to modulating compulsive behavior, but how does it affect circuits that regulate this behavior? While there may be a number of links between allergic inflammation and compulsive behavior brain circuits, modulation of dopamine neurotransmission by allergy-associated histamine release is a promising one. Dopamine is critical to plasticity in fronto-striatal circuits that are involved in action selection, the process of deciding what action to perform at any given moment [155, 156], and decreased synaptic (or “phasic”) release and increased non-synaptic (“tonic”) release of dopamine have been postulated to drive compulsive behavior [157], irrespective of inflammation. Decreased activation of synaptic dopamine receptors leads to their upregulation, increasing the sensitivity of these to tonic dopamine activation [158, 159]. This lack of synaptic dopamine release may lead to the hard-wiring of compulsive behavior and increased fronto-striatal connectivity observed in TS, OCD, and compulsive drug use, which may result from pathological reinforcement of compulsive behaviors in the absence of encoding novel rewarding actions [9, 160–162]. Thus, any system that decreases synaptic dopamine release in the striatum may contribute to the development of compulsive behavior. The comorbidity of these compulsive behavior disorders with aberrant gut physiology suggests that inflammation in the gut may play a causative role in compulsive behavior.

For example, circumstantial evidence suggests that histamine may modulate dopamine neurotransmission and tic-like behaviors as recently observed in a family with a high incidence of TS. This family, in which nearly all members had TS and/or OCD, carried a dominant missense genetic mutation in L-histidine decarboxylase (HDC), an enzyme required for histamine production. The father had TS and OCD, and all eight children had TS, with two also exhibiting OCD [163]. Resequencing the Hdc allele in hundreds of additional TS subjects did not reveal any additional missense variants, demonstrating that this particular mutation is extremely rare [163]. However, the link between functional variations in HDC and TS may not be unique to this index family, as a subsequent candidate gene study identified two Hdc gene SNPs that were more common in TS than in control subjects in a large sample of 520 families originating from eight separate countries [164].

Although the association between Hdc mutations and TS suggests a correlation between histamine production and this tic disorder, a null mutation of the Hdc gene in mice increased amphetamine and stress-induced tic-like behaviors [165]. Treatment with the D2 antagonist haloperidol significantly attenuated amphetamine-induced stereotypies in these mice, as did intracranial infusion of histamine. Not surprisingly, brain histamine receptor expression is also altered in this model. H3R is a histamine autoreceptor located almost exclusively within the brain, and its expression is elevated in the striatum and hypothalamus of Hdc−/− mice [166, 167]. H3R has high constitutive activity and regulates the release of not only histamine but also dopamine, serotonin, noradrenaline, acetylcholine, glutamate, GABA, and several neuropeptides including substance P and calcitonin gene-related peptide [168]. Perfusion of an H3R agonist on mouse striatal slices prevents depolarization-evoked dopamine release, demonstrating the inhibitory effect of this receptor on synaptic dopamine release [168, 169]. In addition, systemic injection of H3R, which may activate hypothalamic and/or striatal H3R receptors primarily located on histaminergic neurons of the tuberomamillary nucleus, increases dorsal striatal activity that is correlated with increased grooming behavior in mice that overexpress central H3R, perhaps by increasing tonic release of dopamine [170]. While it may be unexpected that conditions exhibiting excessive histamine release (allergy), and a model exhibiting an absence of histamine may share common effects on dopaminergic neurotransmission, it is likely that mast cell-released histamine may also activate H3R, thereby potentially reducing phasic release of dopamine and reinforcing the expression of repetitive and stereotyped behaviors.

6. Alternative gut immune-to-brain signaling pathways

The immune system may not be the only link between gut microbiota and compulsive behavior of the CNS. For example, bacterial metabolites influence brain circuits and behavior. Butyrate, a short-chain fatty acid, actively modulates microglial activation states [171], which have been shown to directly affect brain development [172], anxiety behavior [173], and social behavior [174]. In addition, vagus nerve activity directly influences brain transcriptomes with behavioral implications for neuropsychiatic disorders such as schizophrenia [175], and resection of the vagus nerve (vagotomy) is often used to demonstrate the central importance of vagal nerve communication to changes in the brain and behavioral output [176]. Relevant to OCD and other anxiety-related disorders, vagotomy prevented increases in anxiety behavior induced by DSS colitis [177]. In addition, germ-free mice also demonstrate changes in central dopamine metabolite and receptor levels that may be driven by changes in vagal tone or the absence of bacterial metabolites [85]. However, given that the gastrointestinal tract contains the highest load of immune cells in the mammalian body [178] and given the robust differences in immune development between germ-free and conventionally colonized mice [81], the observed differences in central dopamine systems are likely driven at least partially by differences in the immune system between the two models.

Further research is needed to substantiate an inflammatory link between gut barrier dysfunction and compulsive behavior disorders. For example, research on the connection between gut microbiota and allergic inflammation is still in its early stages [179, 180], and several questions remain surrounding the time course of gut barrier disruption, cytokine cascades and the core succession of events that follow these initial inflammatory triggers that result in a full allergic response. Furthermore, while the studies discussed here suggest that differing inflammatory responses (driven by predominantly type 1 vs type 2 biasing cytokines) may have differing effects on behavior, a series of studies examining this question directly is still needed. How systemic allergic inflammation drives these behavioral changes and related changes in the CNS also deserves further study. A central role of systemic inflammation in modulating compulsive behavior does not preclude other modes of immune signaling to the brain. The vagus nerve, too, may carry information about the intestinal immune system, as it has a rich network within the lamina propria [181], expresses innate immune receptors [182], and is critical to certain aspects of sickness behavior response induced by intraperitoneal injections of LPS or IL-1 [183]. Further work on how the vagus nerve responds to gut-derived inflammation, what signals are conveyed, and how these signals specifically affect CNS function and behavioral output is still needed.

7. Shared mechanisms hint at shared treatments: Beyond autism, addiction, and PTSD

There are other striatal-related movement or motivation disorders that may be similarly affected by gut barrier dysfunction. For example, in spite of the apparent and wide differences in disease states, Parkinson’s disease may share many of the same whole-body systemic mechanisms that contribute to pathology in ASD, obesity, alcoholism, and PTSD, as all are linked by gut barrier dysfunction, Th2 inflammation, and dysregulation of striatal dopamine circuits. For example, prior to disease onset, Parkinson’s patients are commonly reported to experience gastrointestinal issues, such as IBD, which necessarily involves the breakdown of the gut barrier [184]. In addition, allergic rhinitis, which is driven by Th2 inflammation, is associated with a three-fold risk increase in developing Parkinson’s disease later in life [185]. The alpha-synuclein mouse model of Parkinson’s demonstrates mechanistic links between gut microbiota, inflammation, and dopamine function, as transplantation of fecal samples from Parkinson’s patients (as opposed to those from healthy controls) into germ-free mice worsens dopamine-dependent motor function and alpha-synuclein pathology in the striatum [186]. Thus, gut dysbiosis, colitis, and allergic inflammation are all associated with a higher risk for Parkinson’s. If all of the conditions discussed in this review truly share etiology, then there may be shared treatments that will ameliorate these conditions as well. In this way, the RDoC approach of identifying shared circuits and mechanisms that contribute to behavioral outputs would benefit from extending a search for targets in the CNS to other body systems affected by gut dysbiosis as well. How might shared mechanisms between conditions as disparate as Parkinson’s, addiction, and compulsive behavior disorders be exploited toward broad-spectrum treatments for these disorders?

The gut offers several possibilities for treating striatal dysfunction disorders, particularly treatments meant to restore order to gut microbiota and treat gut barrier dysfunction. For example, FMT therapies from donors that successfully treat allergies may be helpful in Parkinson’s and the neuropsychiatric conditions discussed here, and clinical trials for FMT treatment for allergies is ongoing, supported by a case report of its successful application [187, 188]. One open label trial for fecal microbiota transplantation (FMT) showed significant improvements in sociability and repetitive behaviors in autistic patients, and case reports offer hope for this treatment in TS and Parkinson’s as well [189–191]. In addition, strategies used to target colitis may also be effective in weakening the link between gut dysbiosis and compulsive behavior. For example, nanoparticles containing anti-inflammatory agents have been used to successfully treat ulcerative colitis in rodents, and such an approach appears ideal to treat gut barrier dysfunction-associated compulsive behaviors [192]. The accessibility of the gut offers an exciting frontier for treatments of compulsive behavior and other psychiatric conditions, and further research on this front may provide wide application to a panoply of movement and motivation-associated disorders.