Contribution Of Intestinal Microbiota To Depression

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

Approximately 18.1 million American adults suffer from depression, predominantly major depressive disorder (MDD), often with co-occurring anxiety disorders(Kessler, Chiu, Demler, & Walters, 2005) (Kessler, et al., 2005). According to the WHO, depression is the "leading cause of disability [worldwide] as measured by YLDs [years lost to disability] as well as the 4th leading contributor to the global burden of disease". In high-income countries, the public health impact is even greater: MDD is the leading cause of disability in the U.S. for individuals aged 15-44 (World Health Organization, 2010). Depressive disorders are also a significant contributor to mortality. Blair-West, et al., have calculated a lifetime suicide risk of individuals suffering from MDD at 3.5%(Blair West, Mellsop, & EyesonAnnan, 1997).

Though much studied, the etiology of depression is still unknown. It is thought to be complex and multifactorial, with contributions from hereditary factors, personality type, physical disease, and adverse life experiences (Belmaker & Agam, 2008). These and other factors ultimately cause changes in brain chemistry that result in alterations in mood, behavior, and physiological state.

The nervous system in humans is divided functionally into the central nervous system (CNS), which serves to mediate behavior, and the peripheral nervous system (PNS), which regulates internal functions and transmits information to the CNS (Kolb, 2011). The enteric nervous system (ENS) is a division of the PNS that consists of approximately one hundred million neurons located within the gastrointestinal tract. Although the GI tract receives motor input from the brain, the majority of neurons within the enteric nervous system are afferent (sensory) neurons, transmitting signals to the CNS via the vagal and splanchnic nerves (Goyal & Hirano, 1996). The field of neurogastroenterology examines the interaction of the CNS and ENS and the effects of bidirectional communication along the gut-brain axis (GBA). The effects of stress and negative emotions on digestion have been established, and the co-morbidity of psychiatric symptoms with certain gastrointestinal disorders documented. Because communication along the GBA is bidirectional, researchers have proposed that disease in the gut and concomitant signals from the ENS can have significant effects on emotion and mood.

The gut lumen is colonized by several hundreds of bacterial species. The number of microbial cells within the GI tract has been estimated as outnumbering the eukaryotic cells in the body by a factor of 10 (Guarner & Malagelada, 2003). These microbes serve various essential functions to the human body, including nutrient absorption, immune regulation, and protection from pathogenic bacteria. As such, the collection of intestinal microbiota has been termed a "forgotten organ", responsible for maintaining mucosal homeostasis and controlling inflammatory responses (O'Hara & Shanahan, 2006). Researchers have hypothesized that prokaryotes within the GI tract are able to influence human behavior, either directly, through "interkingdom signaling" (bacterial production of neuroactive chemicals), or indirectly, through stimulation of the immune system and activation of inflammatory, mood-altering cytokines (Bienenstock & Collins, 2010; Collins & Bercik, 2009). It has been proposed that the normal microbiota and their host are in a homeostatic state, but when microbial composition is disturbed (a state termed "dysbiosis"), changes may be observed in host mood or behavior (Collins & Bercik, 2009). In some instances, experimental perturbation of gut flora in animal models has been shown to alter behavior, but due to the methods employed it is questionable to what degree these results can be used to infer the relationship between human behavior and intestinal microbiota. I propose the following research activities to investigate the association between dysbiotic state and depression.


There is a statistically significant association between abnormal gut microbial composition and depression.

Specific Aims

A three-part approach will be used to assess the relationship between intestinal microbial population and mood disorder. This proposal includes one epidemiologic and two experimental approaches using a mouse model.

(1) Hypothesis: The species composition and relative abundance of intestinal microbiota of people affected by depression differs significantly from that of people without disordered mood.

Samples of intestinal microbiota, obtained through fecal specimens, will be characterized by 16s rRNA analysis to determine bacterial species composition and relative abundance. Volunteers will be screened for depression and anxiety disorder using a self-administered personality inventory. Results will be analyzed to determine whether an association can be established between gut bacterial population makeup and psychiatric symptoms.

(2) Hypothesis: Disruption of mouse intestinal microbiota by oral antibiotic administration will result in behavioral changes.

Two experimental groups of mice will be fed non-absorbable antibiotics to alter composition of the intestinal microbial population; one of the two groups will then receive oral probiotic bacteria to conventionalize gut microbiota. A third control group will not be treated with antibiotic and shall retain normal flora. Mouse mood and behavior will be evaluated by measurement of nutrient intake and multiple behavioral assessments, including tail suspension, forced swim, light/dark choice test, and elevated plus maze. Results will be analyzed to determine whether there is an association between gut bacterial population and behavior in mice.

(3) Hypothesis: Nature and abundance of biomarker chemicals present in the peripheral blood and cerebrospinal fluid (CSF) of dysbiotic animals will differ significantly from those of healthy animals.

Experimental group of mice will be fed non-absorbable antibiotics to disrupt intestinal microbe population makeup; control group will retain normal flora. Blood and CSF samples will be taken from each mouse and analyzed to identify chemicals present in the systemic circulation of dysbiotic and healthy individuals. Substances of greatest interest are physiological markers associated with anxiety and depression such as inflammatory cytokines, certain hormones and growth factors, soluble factors produced by bacteria, and neurotransmitter metabolites, precursors, and associated enzymes. Presence of different chemicals in each group may help elaborate the biological mechanism of induction of behavioral change.

Background and Significance


Depression is a significant public health problem. Approximately eight percent of all Americans suffer from a mood disorder, including MDD and dysthymia (Kessler, et al., 2005). At some point in their lives, up to 20% of women and 10% of men experience a depressive episode, which may result in severe disability and, in some cases, mortality. Between 50 and 70% of all suicides are precipitated by mood disorders (Akiskal, 2005). These psychiatric disorders involve changes in emotional state, decreased functioning, and various physical symptoms.

Major depressive disorder is characterized by alterations in both physical and psychological functioning, including sadness or irritability, anhedonia (loss of ability to experience pleasure), disturbances in sleep and appetite, and suicidal ideation (Belmaker & Agam, 2008). It has been stated that mood disorders such as depression should be viewed as syndromes rather than diseases, as symptoms may not be consistent or have an identifiable cause but do recur periodically (Akiskal , 2005). Because there can be a high degree of symptomatic heterogeneity between individuals, diagnosis requires that a patient meet five or more criteria from a list of nine, one of which must be either depressed mood or loss of pleasure. Dysthymia is a chronic, low-grade depression that is less severe than MDD and is often comorbid with a medical illness or another psychiatric disorder (U.S. Department of Health and Human Services, 1999). Approximately half of all individuals diagnosed with MDD also suffer from an anxiety disorder (US DHHS, 1999). Comorbidity of these conditions is so common that it has been proposed that they share a common genetic basis, with each disease manifestation resulting from differing environmental conditions (Kendler, Neale, Kessler, Heath, & Eaves, 1992).

Depression is thought to have a complex, multifactorial etiology. While a hereditary component is widely accepted, a single "depression gene" has not been identified. Other factors that are thought to influence susceptibility include brain chemistry, personality type, chronic illness and inflammation, obesity, and stressful life experiences. The stress/diathesis model of mental illness proposes that each individual has an inborn threshold for tolerating stress, and that this level of vulnerability will determine whether or not stressful events induce disorders (Zubin & Spring, 1977).

Various hypotheses have been set forward for the mechanism of pathogenesis in depression. (1) The monoamine hypothesis holds that diminished concentrations of monoamine neurotransmitters such as 5-HT, norepinephrine and dopamine in the CNS cause depression (Schildkraut, 1965). The effectiveness of monoamine oxidase inhibitors (MAOIs) and selective serotonin reuptake inhibitors (SSRIs) in treating depression provide evidence for this hypothesis. (2) More recently, the molecular and cellular theory of depression has pointed to stress-induced decreases in brain-derived neurotrophic factor (BDNF) as the ultimate cause of depression, resulting in decreased neural plasticity and disturbances of neurotransmission (Duman, Heninger, & Nestler, 1997). (3) Other researchers are exploring the role of the hypothalamic/pituitary/adrenal (HPA) axis in depression. The HPA is a neuroendocrine system that coordinates an organism's adaptive responses to stress. Activation of the stress system results in corticotropin releasing hormone (CRH) secretion by the hypothalamus. Increased production and secretion of CRH in certain brain region is hypothesized to be involved in the causality of depression (Holsboer, 2000). (4) The cytokine hypothesis of depression posits that proinflammatory cytokines have a neuromodulatory effect. The frequent comorbidity of depression with illnesses of chronic inflammation such as rheumatoid arthritis, and the observed depressogenic effects of IFN-α therapy for hepatitis C and some types of cancers are potential evidence of a causal role for cytokines in depression (Schiepers, Wichers, & Maes, 2005). Several mechanisms by which cytokines affect the nervous system have been suggested, and will be discussed later in this paper.

The enteric nervous system

The enteric nervous system (ENS) is a subdivision of the autonomic nervous system, composed of two major nerve plexuses and connecting fibers within the gastrointestinal tract (Gershon, 1981). Because the ENS contains sensory, motor, and interneurons, it has been called the body's "second brain", autonomously controlling digestion, gastrointestinal endocrine and exocrine secretion, and regulation of immune and inflammatory processes in the gut (Goyal & Hirano, 1996). The ENS uses many of the same neurotransmitters as the brain (5-HT, acetylcholine, and dopamine) and contains afferent as well as efferent nerve fibers. Despite the ability of the ENS to operate independently, gastrointestinal system function can also be modulated by CNS lesions, stress, and other emotional states (Aziz & Thompson, 1998). Because communication along the gut-brain axis is bidirectional, it has been proposed that events within the GI tract may also have an effect on CNS functioning. Psychiatric co-morbidity is present in many gastrointestinal disorders: patients with inflammatory bowel disease are significantly more likely to have a lifetime diagnosis of major depression than healthy controls (Walker et al., 2008). The ENS may modulate CNS activity through nerve signaling or release of neurotropic chemicals. For example, ghrelin, a peptide produced in the GI tract, can have significant effects on behavior, including stimulation of appetite, decreased duration of sleep, and enhancement of spatial memory and learning (Diano et al., 2006; van der Lely, Tschop, Heiman, & Ghigo, 2004).

Gut flora

In the healthy host, intestinal microbiota serve essential functions to the GI system: protecting the gut from pathogens, increasing the bioavailability of nutrients in our food, and influencing the absorption of fat. The influence of intestinal microbiota also extends beyond the gut, and includes regulation of the mucosal immune system, pain perception in the skin, and fat deposition in the liver (Collins & Bercik, 2009).

While the species composition of gut microbiota varies between individuals, it tends to remain consistent within an individual unless disrupted by infection, changes in diet, or exposure to antimicrobials. Disruption of the microbial "organ" by a course of antibiotic treatment can have long term effects. In one study, ciprofloxacin treatment affected the abundance of about one third of microbial species in the large intestine, decreasing the "taxonomic richness, diversity, and evenness of the community" (Dethlefsen, Huse, Sogin, & Relman, 2008). The dysbiotic gut is characterized by an increase in anaerobes such as Bacteroides, Clostridium, and Enterobacteriacae and a decrease in beneficial or "probiotic" bacteria like Lactobacillus and Bifidobacterium. Breakdown of this symbiotic relationship results in altered gastrointestinal function in the host, and can lead to disease (Collins & Bercik, 2009).

The effects of normal and disturbed gut microflora on the host nervous system are just beginning to be explored. One commonly cited example of a neurological disorder which is caused by commensal bacteria in the gut is hepatic encephalopathy. Characterized by symptoms of confusion, irritability, altered levels of consciousness and, in its most severe manifestations, coma and death, hepatic encephalopathy is caused by the accumulation of toxic metabolic byproducts of gut bacteria (Cash et al., 2010). It is caused by impaired functioning of the liver, which normally removes these metabolites, and can be effectively resolved by removing the intestinal bacteria with laxatives and/or non-absorbable antibiotics.

Researchers have observed numerous sensory and behavioral changes resulting from experimental disruption of intestinal flora in animals. Gut microbial composition appears to have an effect on visceral pain perception and sensory function in rats (Verdu et al., 2006). Mice with diet-induced alterations in gut bacteria showed significant differences in memory and learning behavior (Li, Dowd, Scurlock, Acosta-Martinez, & Lyte, 2009). Germ free mice experienced greater stress when tested than colonized controls, implicating commensal bacteria in the development of the hypothalamic-pituitary axis (Sudo et al., 2004).

In humans, changes in gut flora have also been associated with disordered mood. An altered microbiome, evidenced by measurable deficiencies in the ability to ferment specific carbohydrates, has been observed in patients with depression; removal of those carbohydrates from the diet improved patient mood (Ledochowski et al., 2000).

Additionally, there is a strong epidemiological correlation between the gastrointestinal disorder irritable bowel syndrome (IBS) and depression. Antibiotic use has been linked to IBS, and the resultant disruption of gut microbial communities has been proposed as a cause for its symptoms. Probiotic bacteria are known to be helpful in treatment of the condition (Kassinen et al., 2007; Tamboli, Neut, Desreumaux, & Colombel, 2004).

Two major mechanisms by which the intestinal microbiota may affect neurological functioning, mood, and behavior will be considered in this proposal: (1) activation of the inflammatory immune system and (2) neurotropic effects of soluble factors produced by microbial metabolism.


Because the intestinal mucosa is so frequently exposed to potentially antigenic substances, the mucosal arm of the immune system tends toward tolerance and away from immune response. In this way, food allergies and inflammatory responses to innocuous commensals are avoided (Simecka, 1998). It is the presence of intestinal microflora within the gut that helps to control inflammation. Commensal bacteria such as Lactobacillus provoke an immunoregulatory response, promoting the development of regulatory T cells and suppressing production of inflammatory cytokines (Chiba et al., 2010; Livingston, Loach, Wilson, Tannock, & Baird, 2010). One group of researchers has even identified a specific bacterial molecule, polysaccharide A of Bacteroides fragilis, which prevents induction of inflammatory disease by the pathogen Helicobacter hepaticus (Mazmanian, Round, & Kasper, 2008). Disruption of the microbial balance can result in increased inflammation (Collins & Bercik, 2009; Collins, Denou, Verdu, & Bercik, 2009; Verdu, et al., 2006).

Inflammatory cytokines (IL-1β, TNF-α, IL-6, IFN-α) have been shown to have a depressogenic effect, and a role has been proposed for chronic inflammation in major depression (Anisman, 2009). The depressogenic effect of inflammatory cytokines has been explained from an evolutionary medicine perspective as "sickness behavior", a recognized "motivational system that reorganizes the organism's priorities to facilitate recovery from the infection" (Dantzer, O'Connor, Freund, Johnson, & Kelley, 2008). Because sickness behavior has been observed in all mammal and bird species, it is thought to be an evolved strategy, conserved because it favors organism survival. It has been suggested that the immune system functions as a sensory organ, notifying the CNS when infection has occurred. Activation of the immune system by a pathogen stimulates the endocrine system and affects central neurotransmitters (Blalock, 1984). When gut microbiota are disrupted, potentially pathogenic species can flourish, causing chronic low-grade inflammation. A study of effects of chronic gut inflammation on the brain showed activation of the amygdala/anxiogenic center and other areas of the brain associated with behavior (Welch et al., 2005).

Due to their size, cytokines are not normally able to directly access the brain, although they may gain limited access through certain transport mechanisms (Anisman, 2009). Infiltration of the brain may also occur due to the ability of proinflammatory cytokines and chemokines to affect permeability of the blood-brain barrier. Anisman has also proposed that peripherally circulating cytokines may trigger de novo synthesis of cytokines and their receptors from within the brain (Anisman, 2009).

Several mechanisms have been suggested by which pro-inflammatory cytokines might contribute to the development of depression. Immune system activation affects the expression and uptake of multiple neurochemicals and growth factors that influence behavior:

(1) When the HPA is activated, corticotropin releasing hormone (CRH) is secreted by the hypothalamus, stimulating the pituitary to release adrenocorticotropic hormone (ACTH, once known as corticotropin). ACTH acts on the adrenal glands, stimulating release of corticosteroids such as cortisol, which serve to inhibit CRH secretion through a feedback loop. Proinflammatory cytokines may induce hyperactivity of the HPA axis by disrupting the negative feedback inhibition of cortisol, causing increased levels of corticotropin-releasing hormone (CRH) to be produced by the brain (Schiepers, et al., 2005). Sustained hyperactivity of the HPA is observed in patients with major depressive disorder (Holsboer, 1999). In mice, increased CRH levels result in elevated anxiety levels; when the CRH receptor is knocked out, anxiety is decreased (van Gaalen, Stenzel-Poore, Holsboer, & Steckler, 2002).

(2) Systemically administered INF- α is associated with reduced levels of the neurotransmitter 5-HT in the prefrontal cortex. Decreased production or lowered availability of precursors can result in a 5-HT deficit and depressed mood. The effects of INF- α on neurotransmitter levels are attenuated in mice that have been pretreated with a non-steroidal anti-inflammatory drug (Anisman, 2009). It is also thought that proinflammatory cytokines decrease levels of available tryptophan (TRP), the precursor of 5-HT, by stimulating activity of indoleamine-2,3-dioxygenase (IDO), an enzyme that metabolizes TRP (Schiepers, et al., 2005). Studies in cancer and hepatitis C patients undergoing cytokine therapy show decreased blood TRP levels, as well as increased levels of kynurenine (a product of tryptophan metabolism) (Capuron et al., 2002).

(3) Brain-derived neurotropic factor (BDNF) is a growth factor responsible for regulating neural development, supporting the survival of existing neurons, and facilitating plasticity of neural networks (Numakawa, Kumamaru, Adachi, & Kunugi, 2008). Both psychological stressors and inflammatory cytokines decrease levels of BDNF, promoting depression and anxiety. Glucocorticoid, a stress hormone produced in response to HPA activation, acts as an antagonist to BDNF, blocking its effects (Numakawa, et al., 2008). Infusion of BDNF can produce antidepressant-like behavioral effects, and antidepressant therapy increases BDNF expression. The behavioral effects of antidepressant drugs are attenuated in BDNF knockout mice (Anisman, 2009).

Inflammation may lead to altered behavior through one or, more likely, some combination of these pathways.

The anti-inflammatory and anti-depressant properties of certain species of gut flora termed "probiotic" can be interpreted as evidence for the role of gut flora in depression. Probiotic bacteria, defined as "live microorganisms, which when consumed in adequate amounts confer a health benefit on the host", are a fundamental component of the normal gut microflora (Guarner & Schaafsma, 1998). Representative probiotic genera include Lactobacillus and Bifidobacterium. Probiotics in the gastrointestinal system such as L. reuteri 100-23 serve to downregulate the inflammatory response, decreasing levels of circulating proinflammatory cytokines (Livingston, et al., 2010).

Probiotic bacteria have been demonstrated to have beneficial effects on depression in animal models. Administration of Bifidobacterium infantis to rats in the maternal separation model of depression resulted in normalization of the immune response and reversal of behavioral deficits (Desbonnet et al., 2010). B. infantis therapy was also shown to increase blood tryptophan (5-HT precursor) and regulate the hypothalamic-pituitary axis response to stress in test animals (Desbonnet, Garrett, Clarke, Bienenstock, & Dinan, 2008; Sudo, et al., 2004).

In humans, the probiotic Lactobacillus casei strain Shirota was used to successfully treat anxiety symptoms in patients with chronic fatigue syndrome (Rao, 2009). Anxiety and depression are commonly comorbid with CFS, as is altered intestinal microflora, with decreased levels of Bifidobacteria and increased levels of aerobic bacteria as compared to healthy controls (Logan, Rao, & Irani, 2003). Post-treatment stool samples showed a significant rise in both Lactobacillus and Bifidobacteria (Rao, 2009). Due to their ability to decrease production of depressogenic inflammatory cytokines, probiotic bacteria have been proposed as adjuvant to antidepressant therapy (Logan & Katzman, 2005).

Despite extensive documentation of the depressogenic effect of inflammatory cytokines, the role of inflammation in mood disorders is controversial. In one study, when confounding factors such as age, BMI, gender, and smoking habits were controlled for, researchers found no difference between circulating cytokine levels in depressed and healthy patients (Haack et al., 1999). On the other hand, depressive behavior has been observably induced subsequent to secretion of IL-6 and TNF-α in response to endotoxin vaccines and after IFN-α therapy for hepatitis C and some types of cancers (Anisman, 2009). This suggests that inflammatory cytokines are just one of many factors in the complex etiology of mood disorders, and that their effects can be modulated by the influence of other (e.g. cognitive) factors.

Soluble factors

The possibility of "interkingdom signaling" between intestinal microbiota & the human CNS has been recently proposed (Bienenstock & Collins, 2010). Bacteria are able to synthesize neurotransmitters such as 5-HT, GABA, and melatonin (Bienenstock & Collins, 2010). The presence of bacterially generated neurotransmitters in the gut may activate signaling pathways in afferent nerves, modulating mood and behavior. Other soluble factors such as short-chain fatty acids (SCFAs) may affect function of the nervous system. Butyrate, a SCFA which is the product of anaerobic bacterial fermentation, has demonstrated anti-depressant effects when administered systemically to mice (Schroeder, Lin, Crusio, & Akbarian, 2007). Elevated production of propionic acid, a SCFA formed by Clostridium and Bacteroides species in the gut, has been associated with increased anxiety and aggression in animals (Hanstock, Clayton, Li, & Mallet, 2004). The ability of soluble factors to escape the GI tract and enter systemic circulation is increased in the dysbiotic gut. Animals with experimentally-perturbed gut bacteria, characterized by decreased levels of Bifidobacteria, showed increased intestinal permeability (Cani et al., 2009).

Public health

The possibility that microbial communities in the gut can affect mental health prompts consideration of potential public health interventions ranging from dietary to medicinal.

Despite numerous educational campaigns, physician mis-prescription of antibiotics (that is, in the absence of a bacterial infection) is still a major problem in the United States (Mangione-Smith, McGlynn, Elliott, Krogstad, & Brook, 1999). Whether due to misdiagnosis, promotional efforts by pharmaceutical corporations, fear of the repercussions of not using an antimicrobial, or a desire to appease a perceived patient desire for prescription, inappropriate and unnecessary antibiotic use is widespread (Marr, Moffet, & Kunin, 1988). In addition to the significant consequences this practice has in terms of creating antimicrobial resistance, it is possible that an unnecessary course of antibiotics could trigger a depressive episode through destruction of protective gut bacteria.

The threat of cultivating antibiotic-resistant microbes through uncontrolled use of antimicrobials as growth promoters in food animals has also been well described. Regulations exist to minimize antibiotic exposure by humans through dietary intake; however the established "no-effect levels" considered acceptable to human health may not properly take into consideration the effects of these chemicals on the microbial "organ". Carman measured the effects of sub-therapeutic doses of several antibiotics on a fecal and colonic microbial community intended to simulate that of the human large intestine, finding that antibiotic exposure at these levels did affect metabolism of the microorganisms(Carman et al., 2005). In this model system, ciprofloxacin in doses of 0.43 µg/mL (equivalent to slightly less than 0.1 μg/kg/d) reduced counts of Bacteroides and E. coli and affected the ability of the microbial community to resist colonization by Salmonella (Carman, Simon, Fernandez, Miller, & Bartholomew, 2004).

Veterinary antibiotic residues can enter the human diet directly through animal tissues and products or indirectly through the environment (Board on Agriculture, 1999). Because antibiotics are often poorly adsorbed by the guts of livestock, they are often excreted unchanged. Waste from these animals, used as an agricultural fertilizer, can result in contamination of surface and groundwater with antibiotic residues (Sarmah, Meyer, & Boxall, 2006). Excretion of unabsorbed anti-microbials and disposal of unused medicines by humans may also contribute to the environmental burden of antibiotics. Residents of poor communities in developing countries without appropriate wastewater treatment facilities are especially susceptible to this type of exposure (Kalter et al., 2010).

One of the major moderators of the normal balance and diversity of gut microbiota is diet (De Filippo et al., 2010). In a mouse model in which communities of intestinal microbes were both stable and intergenerationally heritable, switching from a low-fat, plant-rich diet to a high-fat, high-sugar diet significantly altered gut flora within a single day (Turnbaugh et al., 2009). A study comparing the gut microbiota of rural African children (who ate a predominantly plant-based diet that was low in fat and animal protein) with that of similarly-aged Western European children (whose diet was low in fiber and high in animal protein, sugar, starch, and fat) found greater gut microbial diversity and fewer pathogenic strains in the African population (De Filippo, et al., 2010).

Dietary supplementation with beneficial bacteria such as Lactobacillus and Bifidobacterium has been shown to change the makeup of the gut microbial community, but only transiently. More long-term effects are observed with consumption of "prebiotics": substances such as indigestible fructooligosaccharides (FOS) that selectively stimulate the activity or growth of desired intestinal bacterial species (Gibson & Roberfroid, 1995).

The above factors suggest nutritional recommendations as well as adoption of a judicious attitude toward antibiotic use for self-limiting conditions, especially for individuals with a predisposition to depression.

Research design and methods

(1) Hypothesis: The intestinal microbial communities of individuals affected by depression differ significantly from those of mentally healthy people.

Approach: A cross-sectional study will be conducted, comparing individual mental health status and exposure (intestinal dysbiosis). Samples of intestinal microbiota, obtained from fecal specimens, will be characterized by 16s rRNA analysis to determine bacterial species composition and relative abundance. Volunteers will be screened for mood disorders using two short self-administered personality inventories, the Beck Depression Inventory (BDI-II) and the State/Trait Anxiety Inventory for Adults (STAI). Results will be analyzed to determine whether an association can be established between gut bacterial population and psychiatric symptoms.

Considerations: There are some challenges associated with determining the exposure status of study subjects. No definitive diagnostic criteria have been established for intestinal dysbiosis (Blue Cross, 2010). Furthermore, while bacterial species present tend to stay constant within an individual unless disrupted by diarrheal illness, antibiotic treatment, or changes in diet, species composition can vary widely between individuals (Guarner & Malagelada, 2003). In this respect, comparison of microbiota between subjects may not be as meaningful as comparison between samples taken from one individual at different time points. However, researchers have identified a limited number of species that are present in the majority of healthy individuals which might represent "the phylogenetic core of the human intestinal microbiota"(Tap et al., 2009). These included members of the genera Faecalibacterium, Ruminococcus, Eubacterium, Dorea, Bacteroides, Alistipes, and Bifidobacterium.

It has been estimated that up to 80% of bacterial species residing within the human gut cannot be cultured by conventional techniques (O'Hara & Shanahan, 2006). Because so few species are recoverable by culture, modern molecular techniques are used to identify commensal populations (Tamboli, et al., 2004). Instead of using prohibitively expensive full-length Sanger sequencing of the small subunit ribosomal RNA (16S rRNA) gene of each of the hundreds to thousands of bacterial species in the GI tract, researchers now use pyrosequencing of a specific, highly variable region of the gene instead (Dethlefsen, et al., 2008). Pyrosequencing employs a sequencing-by-synthesis approach, in which PCR primers for conserved regions of the gene are used to amplify the intervening variable sequence, which is then compared to a reference library and assigned an identity. Because the sequences obtained in this way do not directly correspond to traditional taxonomic classifications, organisms with different sequences are referred to as operational taxonomic units (OTUs) rather than species. A study in which the variable region of the 16S rRNA gene of gut flora from seventeen individuals was sequenced detected 3,180 OTUs (Tap, et al., 2009). Of these, only 2.1%, or 66 OTUs, were present in more than half of the individuals studied. Using the presence of these 66 OTUs as an operating definition of "orthobiosis" or normal microbial makeup, samples will be evaluated and assigned exposure status: ≥33 OTUs = not exposed; < 33 OTUs = exposed).

Certainly, one potential confounder is reverse causality: alterations in GI microbial populations can also be a consequence of fear or stress (Knowles, Nelson, & Palombo, 2008). If an association can be identified between disordered mood and dysbiosis of the GI tract in this cross-sectional study, a prospective cohort study could be undertaken, with periodic measurements made of exposure and disease in order to infer the temporal sequence and establish the direction of causality. The experimental animal model aspects of this proposal will also serve to increase understanding of causality.

Because depression probably results from multiple etiological factors, the effect size of disrupted gut microbiota in humans is predicted to be small. Additionally, while the BDI-II and STAI have been validated by other measures of depression, they are obviously not as sensitive as evaluation by a mental health professional. For these reasons, to increase confidence in results, a large (n=1500) sample size will be used.

(2) Hypothesis: Disruption of mouse intestinal microbiota by administration of rifaximin, a broad spectrum non-absorbable oral antibiotic, will result in quantifiable behavioral changes (relative to a control group). These changes will be reversed upon conventionalization with oral Bifidobacterium.

Approach: Two experimental groups of mice will be fed non-absorbable antibiotics to disrupt intestinal microbe population makeup; one of the two groups will then receive oral probiotic bacteria to conventionalize gut microbiota. A third control group will not be treated with antibiotic and retain normal flora. Mouse behavior will be evaluated by measurement of nutrient intake and multiple behavioral assessments, including tail suspension, forced swim, light/dark choice, and elevated plus maze. Results will be analyzed to determine whether there is an association between gut bacterial population and behavior in mice.

Considerations: Li, et al., noted behavioral changes in mice that had been fed different diets to induce changes in the intestinal microbiota (Li, et al., 2009). However, because dietary factors (such as quantity of the nutrient taurine) differed between the control and experimental groups, it is impossible to confirm that the effects observed were due solely to changes in gut flora.

Other studies have relied on a gnotobiological approach, in which germ-free mice are developed as a model of dysbiosis to characterize the effects of different gut microbiota on behavior. However, because the microbiome at birth is thought to imprint physiological and immunological systems, with lifelong effect, mice whose immune system has not developed normally are not an ideal test subject. A preferable approach is to selectively disrupt the bacterial makeup with a non-absorbable antibiotic. Because even slight absorption of the antibiotic into the circulation may produce confounding systemic toxic effects, this study will use Rifaximin (INN), a derivative of rifamycin exclusively indicated for GI tract infections that is not absorbed by the intestine (Beseghi, 1998).

Behavioral researchers have noted that when measuring mood in test animals, it is important to differentiate between anhedonia, a key symptom of depression, and malaise, which can be caused by stress, illness, and other factors. Animals that won't eat their standard chow have no appetite and may be experiencing sickness behavior; animals that won't work for sugar can be considered to lack motivation and are anhedonic (Anisman, 2009). For this reason, consumption of both chow and sugar by test and control mice will be quantified. In discriminating the behavioral effects of anhedonia from behavior induced by stress, anhedonia is thought to be selectively associated with specific responses in the forced swim and tail suspension tests (Strekalova, Spanagel, Bartsch, Henn, & Gass, 2004).

The forced swim test consists of two trials in which mice are placed in a deep, water-filled cylinder from which they cannot escape. During the second trial, the length of time for which mice are immobile is measured. The tendency toward immobility is interpreted as characteristic of a depressive-like state, and can be reversed through administration of antidepressant drugs (Porsolt, Bertin, & Jalfre, 1977). In the tail suspension test, mice suspended by their tails for six minutes become similarly immobile, reflecting "behavioral despair" or hopelessness (Steru, Chermat, Thierry, & Simon, 1985).

To assess whether dysbiosis induces anxiety-like behavior in the animal model, two rodent tests of anxiety will be used: (1) light/dark choice test, in which it is postulated that anxious animals spend less time in illuminated areas and (2) elevated plus maze, which measures time spent by animals in open vs. closed arms of maze, with less anxious animals spending relatively more time in open arms (Crawley & Goodwin, 1980; Lister, 1987).

The difference between treated and control group outcomes is expected to be larger than in the epidemiological study because of the physical homogeneity of inbred mice as well as the ability to control for other variables. Therefore, the sample sizes for each group can be much smaller (n=50).

(3) Hypothesis: Nature and abundance of biomarker chemicals present in the peripheral blood and cerebrospinal fluid (CSF) of dysbiotic animals will differ significantly from those of healthy animals.

Approach: Experimental group of mice will be fed non-absorbable rifaximin to disrupt intestinal microbe population makeup; control group will retain normal flora. After 24 hours, CSF and blood samples will be taken from each mouse and analyzed to identify chemicals present in the systemic circulation of dysbiotic and healthy individuals. Levels of the following chemicals will be quantified: acute-phase proteins; brain-derived neurotrophic factor (BDNF); corticotropin-releasing hormone (CRH); cortisol; C-reactive protein (CRP); 5-hydroxyindoleacetic acid (5-HIAA); homovanillic acid, (HVA); indoleamine 2,3 dioxygenase (IDO); insulin; inflammatory cytokines (IL-1β, TNF-α, IL-6, IFN-α); tryptophan; kynurenine; matrix metalloproteinase 9 (MMP-9); quinolinic acid; short-chain fatty acids (SCFAs: acetic, propionic and butyric acids); 3-methoxy-4-hydroxyphenylglycol (MHPG).

Considerations: Because the etiology of mood disorders is thought to be multi-faceted and heterogeneous between individuals, identifying universal physiological markers of depression and anxiety is challenging. While there are no definitive biomarkers for depression, various chemicals have been proposed:

A. monoamine neurotransmitter precursors/metabolites/ associated enzymes

It is believed that neuropsychiatric disorders result from disruption of nerve signal transmission in the brain. The molecular components of neurotransmission, including the monoamine neurotransmitters 5-HT, norepinephrine, and in some cases dopamine, are implicated in these disorders, and levels of these chemicals may therefore be an ideal biomarker for depression. Because of the challenges of measuring neurochemical transmission in the brain, concentrations of monoamine metabolites (5-HT ƒ  5-HIAA; norepinephrine ƒ  MHPG; dopamineƒ HVA) and precursors (tryptophan ƒ  5-HT; tyrosine ƒ  dopamine and norepinephrine) in CSF and peripheral plasma will be measured instead (Fisar & Raboch, 2008). Lowered CSF concentrations of 5-HIAA, HVA and MHPG (interpreted as lowered rate of monoamine neurotransmitter use in brain) have been correlated with depression in some (but not all) studies (Widerlov, Bissette, & Nemeroff, 1988). Reduced levels of monoamine precursors have been shown to result in reduced levels of neurotransmitters within synapses, and in patients with a family history of mood disorder, induced depressed mood (Benkelfat, Ellenbogen, Dean, Palmour, & Young, 1994).

Indoleamine 2,3 dioxygenase (IDO) is an enzyme produced by activated macrophages and other immune cells. IDO degrades tryptophan, forming breakdown products such as the potent neurotoxin quinolinic acid (which can induce apoptosis of glial cells in the CNS) and kynurenine (Leonard, 2010). IDO can be activated by pro-inflammatory cytokines, particularly IFN-γ and IFN-α. In a mouse model, intraperitoneal administration of LPS and subsequent activation of the innate immune system resulted in increased IDO activity in the hypothalamus and frontal cortex (Moreau, Castanon, & Lestage, 2005). In patients with cytokine-induced depression, lowered levels of serum tryptophan and increased serum kynurenine were found, possibly due to cytokine activation of IDO (Capuron, et al., 2002).

B. brain-derived neurotrophic factor

Serum and CSF samples will be evaluated to explore the possibility that dysbiosis-induced inflammation results in decreased levels of brain-derived neurotrophic factor (BDNF). In one study, researchers found that BDNF serum levels were significantly lower in depression cases than controls, suggesting that major depression is characterized by decreased BDNF in serum (Karege et al., 2002).

C. corticotropin-releasing hormone

Another possible group of biomarkers include neuro-endocrine chemicals such as the hypothalamic corticotropin-releasing hormone (CRH). Abnormalities in stress hormone regulation are routinely observed in anxiety and depression, due to dysfunction of the hypothalamic-pituitary axis (HPA) and elevated secretion of CRH (Holsboer, 1999). Increased levels of CRH have been detected in more than half of depressed patients studied, a result of elevated HPA function (Fisar & Raboch, 2008).

D. pro-inflammatory cytokines and acute phase proteins

According to the "cytokine hypothesis", inflammatory immune proteins play a role in the pathophysiology of depression, suggesting the use of cytokines such as IL-1, IL-6, and TNF-α as biomarkers. Plasma concentrations of acute phase proteins are affected by inflammation. Levels of positive acute phase proteins such as C-reactive protein and complement are elevated, while concentrations of negative acute phase proteins like albumin, transferrin, and transthyretin in plasma are decreased (Bankier, Barajas, Martinez-Rumayor, & Januzzi, 2009; Raison, Capuron, & Miller, 2006; Ritchie et al., 1999).

E. short-chain fatty acids: acetic, butyric and propionic acids

Short-chain fatty acids (SCFAs) are byproducts of bacterial carbohydrate fermentation that have been proposed as soluble factors that act directly on the brain to influence behavior. These include acetic, butyric and propionic acids. The latter has been found at higher levels in IBS sufferers than controls (Tana et al., 2010).

F. insulin and matrix metalloproteinase-9

Finally, recent work by Domenici has demonstrated significantly elevated plasma levels of both insulin and MMP-9 in patients with major depressive disorder (Domenici et al., 2010). Depression has been previously linked to insulin resistance, and it has been hypothesized that insulin resistance may result from disruption of hormonal activity associated with depression (Timonen et al., 2005). It is not clear what role MMP-9 plays in depression, but the matrix metalloproteinases are known to process proteins required for synaptic development and neural functioning (Domenici, et al., 2010).

Analysis of results

Aim #1: Relative risks / risk ratios will be calculated to determine the probability of having symptoms of depression in dysbiotic individuals relative to orthobiotic individuals. In order to control for confounding, results will be stratified by variables such as use of alcohol, tobacco, drugs, or psychiatric medication.

Aim #2: A two tailed Student's t-test will be used to compare (1) control and dysbiotic groups and (2) dysbiotic and conventionalized groups to determine whether there is a significant difference in behavior between mice with healthy and disrupted gut flora.

Aim #3: A two-tailed Student's t-test will be used to compare levels of each molecular marker in plasma and CSF of (1) control and dysbiotic and (2) dysbiotic and conventionalized groups and determine whether there is a significant difference in concentration of these markers between groups.


As described previously, because events in the gut may affect the brain, and the brain modulates gut activity, it can be difficult to discriminate cause from effect. A cross-sectional epidemiological study such as the one proposed above can only identify associations between exposure and disease, not causality. However, identifying an association is an important first step in uncovering disease etiology, and would justify further study including undertaking a prospective cohort study in which the temporal sequence of exposure and outcome could be established.

While the reliability of extrapolating results from animal models to humans is often questionable, the experiments proposed in this paper could identify possible biological pathways by which the gut microbiome, inflammation, and neurological functioning interact. These findings could serve to generate hypotheses that could then be tested in human patients.

The challenges of pinning down the causative agents of a disease that is modulated by individual variations in environment, cognitive factors, and heredity are enormous. However, it is worth examining the possibility that biologically plausible mechanisms such as disruption of gut microbes and chronic inflammation may be risk factors. According to the diathesis/stress model of mental illness, depression could be triggered by a relatively minor environmental stressor in an individual with a high degree of inherent vulnerability. Therefore, it is vital to be aware of all types of stressor, emotional, physical or otherwise, that could induce pathology. If it can be shown that disruption of the intestinal microbiota contributes to mental illness, it will be of great benefit to vulnerable individuals and their health care providers. In this case, heightened awareness of the potential consequences of antibiotic use and dietary choices could serve to prevent onset of a depressive episode and even save a life.