Antidepressive Mechanisms of Probiotics and Their Therapeutic Potential

1Approximately 1014 microbes, also known as gut microbiota, reside in the human gastrointestinal tract. The majority of these microbes belong to the Firmicutes, Bacteroidetes, Actinobacteria and Proteobacteria phyla. The gut microbiota flourishes in a symbiotic alliance with the host and, as such, has eminent regulatory effects on the host physiology. The gut microbiota actively engages with the proper development and functioning of both the immune system and brain. This is mediated by the microbiota–gut–brain (MGB) axis that lays the foundation for the intricate communicative pathways between gut microbiota and the nervous, immune and endocrine systems. However, the diversity and richness of gut microbiota are susceptible to change based on the host’s lifestyle. An adverse change induces a gut dysbiosis which disrupts the symbiosis maintained by the MGB axis. Indeed, a gut dysbiosis has been linked to various health conditions, such as obesity, IBS, schizophrenia, Parkinson’s disease and MDD.

2Major depressive disorder is currently the leading cause of disability worldwide and is expected to outrank heart diseases as the number one disease burden by 2030. According to the Diagnostic and Statistical Manual of Mental Disorders-5, MDD is diagnosed when a person experiences most of the following symptoms for at least 2 weeks: depressed mood, anhedonia, excessive guilt, suicidal ideation, changes in appetite and sleep, psychomotor retardation, poor concentration and fatigue. Among these criteria, either depressed mood or anhedonia (or both) must be present for a diagnosis of MDD. In this review, the term “depression” would be used to refer to symptoms that characterize MDD.

3A causal relationship potentially exists between the gut microbiota and MDD. Germ-free (GF) rodents developed depressive-like behaviors following fecal microbiota transplantation from MDD patients, but not from healthy people. As compared to healthy individuals, MDD patients have a different gut microbiota profile. The decrease in Faecalibacterium, Bifidobacterium, Lactobacillus, and Dialister, and increase in Clostridium, Streptococcus, Klebsiella, Oscillibacter, Allistipes, Eggerthella, Holdemania, Gelria, Turicibacter, Paraprevotella, and Anaerofilum genera have been found among MDD patients. This shift in the gut microbiota composition may contribute to a shift in the regulation of the host physiology. It is, thus, worthwhile to tackle MDD from the MGB axis standpoint, with an emphasis on the gut microbiota.

4Probiotics are microbes (usually lactic acid bacteria such as Lactobacilli and Bifidobacteria) that benefit the host physiology upon ingestion. Probiotics are marketed in the form of capsules, powder or fermented products. The global market size of probiotics amount to billions and is increasing annually due to consumers’ interest in optimizing their health with functional foods. Probiotics have been utilized to modulate the MGB axis in an attempt to treat diseases, including MDD. Meta-analyses and systematic reviews have already supported the efficacy of probiotics in reducing clinical depression and depressive-like symptoms in MDD patients and healthy individuals, respectively.

5To what extent are probiotics viable tools to treat MDD/depression? This review addresses this question by first outlining the workings of MGB axis and process by which this axis becomes maladaptive, leading to the development of depression. Antidepressive mechanisms of probiotics are further elucidated by drawing parallels between the physiological outcomes that accompanied the behavioral changes to the MGB axis from animal and human research. Lastly, in light of the heterogeneous nature of both the gut microbiota composition and depression subtypes in the clinical setting, challenges and potentials in translating probiotics for clinical use are discussed.

6The first point of contact between the gut microbiota and host nervous system is likely via the enteric nervous system (ENS). The ENS has been described as “the second brain” due to its neuronal complexity on par with the brain and its ability to function as an independent, discrete unit to regulate gut-related activities and the immune system. Without gut microbiota, the excitability of enteric neurons would likely be attenuated, based on data observed in GF mice. Through the ENS, gut microbiota and the brain communicate bidirectionally through neural and humoral (systemic circulation) pathways. Parasympathetic vagus afferents carry neural information from internal organs, including the gut, to the brain. The vagus nerve also consists of motor neurons that innervate nearly all enteric neurons. This enables the brain to influence the activity of ENS to some extent, particularly the state of intestinal permeability and gut inflammation. Sympathetic spinal nerves also connect enteric neurons to the brain, albeit to a lesser extent than vagal nerves. Additionally, the humoral route allows microbial metabolites to enter the systemic circulation and exert its effects elsewhere, including the brain. Likewise, the brain also sends chemical messengers, such as cytokines and glucocorticoids, via the humoral route to regulate the gut physiology.

7The gastrointestinal tract contains approximately 70% of the immune system. Immune cells express TLRs that respond to foreign antigens, such as LPS, as they penetrate the intestinal mucosal barrier. This promptly triggers production of inflammatory cytokines, mainly ILs, tumor necrosis factor (TNF)-α and IFN-γ. These cytokines enter the brain through various pathways. The humoral pathway enables cytokines to enter circumventricular organs or permeable regions of the BBB or bind to carrier proteins that cross the BBB. The neural pathway allows gut cytokines to stimulate specific brain areas such as the brainstem, hypothalamus and limbic structures via vagus and spinal afferents. The cellular pathway allows cytokines to be transported into the brain by the action of monocytes or macrophages. These cytokines could also bind to receptors on astrocytes and microglia, and subsequently trigger cytokine production within the brain.

8When proinflammatory signals reach the brain, the hypothalamic-pituitary-adrenal (HPA) axis, a sympathetic-neuroendocrine system, is activated to restore homeostasis. In response to stress, the hypothalamic paraventricular nucleus (PVN) synthesizes and releases corticotropin-releasing factor (CRF) to stimulate the anterior pituitary gland to release adrenocorticotropic hormone (ACTH) into the systemic circulation. ACTH stimulates the adrenal cortex to release glucocorticoids (cortisol in humans and corticosterone in rodents) which inhibit the release of CRF, establishing a negative feedback loop. Glucocorticoids are core effectors of the HPA axis that travel by the humoral route to exert its adaptive effects elsewhere; for instance, to reduce gut inflammation.

9Furthermore, neurotransmitters in the brain serve indispensable roles in maintaining proper brain functions. Neurotransmitters such as GABA, glutamate (Glu), serotonin (5-HT), DA, NE, histamine and acetylcholine (ACh) are known to be synthesized by the gut microbiota. Notably, Lactobacillus, a prominent probiotic genus, produces multiple neurotransmitters in a species-dependent manner in vitro (Table 1). It should be noted that gut-derived neurotransmitters are functionally different from brain-derived neurotransmitters. The bioavailability of precursors for these neurotransmitters is also regulated by the gut microbiota. For example, carbohydrate-fermenting microbes secrete butyrate (a SCFA) that stimulates 5-HT synthesis from intestinal enterochromaffin cells (ECs). In contrast, Clostridia metabolites, such as 4-cresol and 4-hydroxyphenylacetate (4-HPA), inhibit dopamine-β-hydroxylase (an enzyme that converts DA to NE in the brain). These microbial neuroactive molecules likely modulate local ENS signaling, which ultimately influence the MGB axis.

10Acute psychological stress increases the release of ACh from cholinergic nerves and glucocorticoids from the HPA axis, both of which loosen tight junctions of the intestinal barrier (Figure 1). Other stressors such as poor diet, sleep deprivation, antibiotics, environmental pollutants and excessive exercise also increase the intestinal permeability. Additionally, exposure to stress stimulates sympathetic spinal nerves to release NE into the gut which expedites quorum sensing systems and iron uptake of bacteria, leading to increased virulence and growth of pathogenic bacteria (e.g., Escherichia coli, Salmonella, Campylobacter, etc.). These factors facilitate penetration of bacteria and their toxins, such as LPS, through the weakened intestinal barrier. Administration of LPS increased proinflammatory cytokines and caused anxiety and depression in healthy males in a dose-dependent manner. This phenomenon is only transient due to the adaptive response of the immune system and HPA axis. However, chronic stress prevents this homeostatic restoration and causes prolonged inflammation and HPA axis overactivity, both of which aggravate the disrupted intestinal barrier. During this process, chronic inflammation renders the immune system insensitive to inhibitory signals from glucocorticoids. Excess proinflammatory cytokines, in turn, disrupt the negative feedback inhibition of circulating glucocorticoids of the HPA axis. Indeed, MDD patients often show increased intestinal barrier permeability and elevated serum antibodies against LPS.

11Excessive glucocorticoids hyperactivate monoamine oxidases (MAOs; enzymes that degrade 5-HT, NE, and DA). An overactive HPA axis can also induce gut dysbiosis and impairment of brain neurotransmitter systems. Higher baseline levels of cortisol, an indicator of an overactive HPA axis, were detected in more than 70% of MDD patients. Proinflammatory cytokines and glucocorticoids upregulate indoleamine 2,3-dioxygenase (IDO) and tryptophan-2,3-dioxygenase (TDO) enzymes, respectively. Both enzymes metabolize TRP into KYN and quinolinic acid, which reduce the bioavailability of TRP to cross the BBB, thereby lowering 5-HT synthesis. This is evidenced by low plasma TRP levels that were also correlated to a heightened proinflammatory state found in MDD patients. Furthermore, proinflammatory cytokines can decrease levels of DA, 5-HT and NE in the brain by upregulating their reuptake via presynaptic transporters and downregulating enzymatic cofactors required for their synthesis. Indeed, administration of cytokines consistently induced neurotransmitter imbalances in the brain and behavioral changes that are reminiscent of depression in animals and humans. Similarly, higher levels of proinflammatory cytokines were observed in depressed individuals as reported using meta-analyses of the data available in the literature.

12A stress-induced inflamed gut adversely alters the relative abundances of preexisting bacteria in the gut (Figure 1). Acute psychological stress stimulated the release of inflammatory mediators that were correlated with the lowered abundance of Coprococcus, Pseudobutyrivibrio, Dorea, and Lactobacillus in mice. This, in turn, allowed the proliferation of Clostridium species in the gut. The gut microbiota of chronic-stressed mice also deviated from the baseline, whereby an increase in proinflammatory bacteria, such as Helicobacter and Streptococcus, and a decrease in butyrate-producing bacteria, such as Roseburia and Lachnospiraceae species, were observed. Altered gut microbiota composition consequently exacerbates gut inflammation and further increases intestinal permeability and production of proinflammatory cytokines. The precise mechanism underlying vulnerability of certain bacteria to inflammation remains poorly understood. It is hypothesized that inflammation disrupts β-oxidation of intestinal epithelial cells (IECs, both enterocytes and colonocytes) to increase oxygen content in the gut lumen. This promotes formate oxidation that favors the growth of facultative anaerobes, such as E. coli, that are pathogenic and inflammatory at the cost of obligate anaerobes, such as Bacteroides and Firmicutes.

13A dysregulated gut microbiota translates to a shift in the production of neuroactive metabolites and alters host neurotransmitter circuitry. This corresponds with disrupted levels of neurotransmitters in the brain of GF mice. Altered neurotransmitter profile (e.g., GABA, Glu, 5-HT, DA, and NE) has been associated with the pathophysiology of depression. Therefore, pharmaceutical antidepressants function to restore synaptic levels of neurotransmitters. In addition, impaired neurotransmitter systems within the ENS may alter gut motor function. This has direct implications as gut motility is a determining factor in the size and diversity of gut microbiota. Therefore, chronic stress sets up a vicious cycle of increased intestinal permeability, chronic inflammation, hyperactive HPA axis, altered gut microbiota profile and neurotransmitter imbalances – forming a maladaptive MGB axis (Figure 1). Furthermore, MDD patients perceive stress as more threatening and challenging to cope with compared to healthy individuals. These negative emotions can increase their sensitivity to stressors, such as an elevated cortisol response. To restore this malfunctioned axis, probiotics have been demonstrated by meta-analyses and systematic reviews as a potential treatment for MDD/depression. Potential antidepressive mechanisms of probiotics are elucidated in the following section.

14Probiotics secrete a wide range of signaling molecules that operate via distinct pathways to exert their effects, be it antidepressive, immunomodulatory or modulation of neurotransmission. This review classifies probiotic-associated signaling molecules into four types: neurotransmitters, bacterial secreted proteins, butyrate and other bioactive molecules (Figure 2). Some probiotics can secrete signaling molecules of different types. In this regard, the mechanisms of individual probiotics will be presented in the order of pertinence and similarity to each other.

15Lactobacillus rhamnosus JB-1, the typical experimental strain of L. rhamnosus, was formerly referred to as Lactobacillus reuteri. Orally administered L. rhamnosus reduced depressive-like behaviors in normal, healthy mice and chronic-stressed mice. Postpartum women and obese individuals that were supplemented with L. rhamnosus reported lower depressive thoughts compared to the control group. In vagotomized rats, behavioral and physiological benefits of L. rhamnosus were abolished. This substantiates the vagus nerve as an essential conduit in the signaling pathway of L. rhamnosus. Introduction of L. rhamnosus into the gut lumen heightened the firing rate of vagus nerve and enteric neurons in mice. These findings suggest that L. rhamnosus signals to the brain via the neural route, which may influence the central GABAergic system and HPA axis to manifest an antidepressive effect (Figure 2A). However, it is unclear whether neurotransmitters, cytokines or other molecules are involved in the neural signaling of L. rhamnosus.

16Glutamine is a precursor to Glu while Glu is a precursor to GABA. Reduced levels of GABA and Glx (Glu + glutamine) have been consistently reported in cortical regions of MDD patients. A dysfunctional glutaminergic system, that is partly responsible by a decreased GABAergic tone, is also implicated in MDD. N-acetyl aspartate (NAA) is regarded as a marker for neuronal vitality. In MDD patients, decreased NAA levels in the PFC and hippocampus have been detected. These neurochemical (i.e., Glx, NAA, and GABA) levels in the PFC and hippocampus of mice increased when administered with L. rhamnosus, implicating its antidepressive potential.

17Intake of L. rhamnosus altered the central mRNA expression of GABAA and GABAB receptors while reducing depressive- and anxiety-like behaviors in mice. These effects were also dependent on an intact vagus nerve. With prebiotics, L. rhamnosus intake decreased hippocampal GABAAα2 mRNA expression in stressed mice. L. rhamnosus produced GABA and Glu efficiently from microbial glutamate decarboxylase and glutaminase, respectively, in vitro. These biosynthetic machineries utilized by microbes to synthesize Glu and GABA are mutual in neurons, which support the interkingdom communication of microbial GABA. It was demonstrated in vitro that gut microbial GABA can cross the intestinal barrier via H+/GABA symporter. The microbial GABA may subsequently interact with GABA receptors and transporters that are widely expressed on enteric neurons and vagus afferents.

18Administration of L. rhamnosus reduced stress-induced plasma corticosterone levels in mice that averted depression. This could be due to the innervation of PVN neurons by GABAergic synapses that can be desensitized by acute stress. Inhibited GABA signals allow continuous release of CRF by PVN neurons, which ultimately leads to cortisol overproduction and HPA axis overactivity. Impairment of GABA receptors also inhibits hippocampal neurogenesis, which has been shown to activate the HPA axis and induce depression in mice. Such effects may be possibly prevented by the production of GABA by L. rhamnosus.

19Individuals with low mood reported feeling happier after consuming milk containing L. casei, but not the placebo. Intake of mixed-species probiotics that included L. casei also reduced clinical depression and depressive-like symptoms in MDD patients and healthy individuals, respectively. Similar to L. rhamnosus, evidence suggests that L. casei may also regulate the HPA axis via the neural route (Figure 2A).

20Intake of L. casei stimulated vagus afferents and decreased both the activity and quantity of CRF-expressing cells in PVN of rats. Intragastric injection of L. casei downregulated the activity of sympathetic efferents to adrenal glands and liver, and this effect ceased upon vagotomy. In clinical trials, L. casei supplementation lowered salivary cortisol levels, feelings of stress and frequency of abdominal- and flu-related symptoms in stressed individuals. These studies imply that L. casei prevents HPA axis overactivity via the vagus nerve, which may consequently lower stress-related feelings and illnesses. L. casei produced GABA in vitro, indicating a possibility that it may share an antidepressive mechanism of L. rhamnosus. Stressed individuals that consumed L. casei showed improvements in mental health and gut microbiota composition, characterized by increased Lactobacillus and Bifidobacterium populations. As most of the antidepressive probiotics belong to Lactobacillus and Bifidobacterium genera, the potential antidepressive capacity of L. casei is highly supported.

21Similar to L. rhamnosus and L. casei, L. brevis produces GABA via glutamate decarboxylase in substantial amounts. This indicates that L. brevis may share a mutual mechanism of action with L. rhamnosus and L. casei (Figure 2A). Although L. brevis has been shown to influence neither the central GABAergic system nor the HPA axis, L. brevis appears to promote sleep.

22Milk fermented with L. brevis had increased GABA content. This L. brevis-fermented milk demonstrated an antidepressive potency on par with fluoxetine, a SSRI, in depressed rats. Intriguingly, intake of L. brevis-produced GABA improved sleep duration in mice. Another study also showed that dietary L. brevis enhanced sleep quality and voluntary physical activity in mice. GABA is the main inhibitory neurotransmitter that is widely associated with sleep, and GABA receptors are frequent targets for pharmaceutical drugs, such as benzodiazepine, to treat insomnia. GABA-enriched foods and GABA extract have also been shown to improve sleep quality in insomniacs and healthy individuals. Therefore, L. brevis has therapeutic value for insomnia, which reflects one of the diagnostic criteria for MDD.

23Treatment of L. reuteri ameliorated depressive-like behaviors in chronic-stressed and immobilization-stressed mice. The former study further elucidated the mechanism of L. reuteri which involves regulation of IDO, a rate-limiting enzyme of immune cells that catabolizes TRP to KYN. It is also well documented that L. reuteri exhibits anti-inflammatory activities. It is, thus, conceivable that L. reuteri may also prevent activation of IDO by proinflammatory cytokines.

24The etiology of depression is partly attributed to a dysregulated KYN/TRP pathway. An elevated ratio of plasma KYN/TRP often correlates positively with the depression severity in human. It was demonstrated that L. reuteri intake improved behaviors of depressed mice by reversing the stress-induced (1) decrease in fecal H2O2 levels and Lactobacillus populations, and (2) increase in intestinal IDO1 expression and plasma KYN levels. KYN administration attenuated this antidepressive effect, which indicates that L. reuteri ameliorates depression by reducing plasma KYN levels. This study also showed that L. reuteri generated high amounts of H2O2 in vitro, and the author proposed that H2O2 is the key metabolite in mediating antidepressive effect of L. reuteri. This is because H2O2 catalyzes peroxidase-mediated reactions that inhibit IDO activity. H2O2 is transported by aquaporin-3 transporters that are expressed on IECs and immune cells. These findings suggest that microbial H2O2 can potentially cross the intestinal barrier to suppress IDO activity in immune cells, which would lower circulating KYN levels (Figure 2D).

25Lactobacillus reuteri possesses histidine decarboxylase that converts dietary L-histidine to histamine, which inhibits the production of TNF-α in vitro. The microbial histamine suppressed proinflammatory cytokine activities in IECs via the histamine-2 receptor signaling pathway in mice. This effect disappeared when the histidine decarboxylase gene of L. reuteri was inactivated by mutagenesis. Intriguingly, microbial histamine also activated histamine-1 receptors to initiate downstream proinflammatory pathways in mice. However, the substrate for this pathway, diacylglycerol, is metabolized to phosphatidic acid by diacylglycerol kinase produced by L. reuteri. Thus, L. reuteri secretes both histamine and diacylglycerol kinase that act on histamine receptors to produce an anti-inflammatory effect. Orally administered L. reuteri simultaneously alleviated colitis and behaviors indicative of anxiety and depression in stressed mice. These effects were also accompanied by a decrease in colon inflammation and blood levels of LPS, interleukin-6 (IL-6) and corticosterone. In the same study, this reduction in peripheral inflammation prevented the infiltration of activated microglia into the hippocampus and increased hippocampal BDNF expression (; Figure 2A). BDNF has been extensively studied for its vital role in neuronal function and its causal link to depression. Antidepressants such as SSRI and ketamine also increase hippocampal BDNF expression as part of their mechanism of action. Furthermore, this anti-inflammatory effect of L. reuteri may prevent IDO activation by proinflammatory cytokines.

26Lactobacillus plantarum supplementation decreased depressive-like symptoms in chronic-stressed mice and stressed adults with mild depression, though the latter study did not reach statistical significance. Following L. plantarum intake, reduction in plasma corticosterone levels and inflammation were seen in mice with reduced depressive-like behaviors. Another study reported that mice fed with L. plantarum displayed an increase in cecum SCFAs levels (acetic and butyric), and a decrease in intestinal permeability and level of MAOs in the brain. These physiological changes can be unified into a mutual mechanism that L. plantarum likely mitigates systemic inflammation (Figure 2B).

27Chronic-stressed mice fed with L. plantarum exhibited reduced depressive-like behaviors, coupled with an increase in butyrate and butyrate-producing bacteria, such as Lactobacillus, Bacteroidetes, and Roseburia. L. plantarum synthesizes butyrate via fatty acid synthase II–thioesterase, a glutamine-mediated butyrogenic pathway. Butyrate can enter IECs through cholesterol-rich microdomains and/or monocarboxylate transporter 1 protein, and promote synthesis and assembly of tight junction proteins of IECs. Butyrate also has anti-inflammatory properties; for instance, butyrate inhibited proinflammatory activities of IECs in vitro and interacted with IECs to regulate host T cell responses. Butyrate may also diffuse into the systemic circulation to exert anti-inflammatory effects on various organs and tissues, including the brain. Indeed, butyrate has been shown to normalize behavior of depressed rodents through epigenetic regulations of hippocampal BDNF expression. These outcomes are consistent with the finding that L. plantarum intake increased hippocampal BDNF expression and cecum butyrate levels in chronic stress-induced depressed mice.

28Lactobacillus plantarum supplementation decreased MAOs levels in brain tissues of mice with reduced depression. This is in line with another finding that L. plantarum intake in mice increased levels of DA and its metabolites (HVA and 3,4-dihydroxyphenylacetic acid, DOPAC) in the PFC, along with reduced depressive-like behaviors. However, another study showed that L. plantarum increased DA levels in the striatum of mice while alleviating anxiety-like behaviors. These studies suggest that L. plantarum likely affects the central DA system in a context-dependent manner. It was also proposed that L. plantarum increases DA levels in the PFC to prevent HPA axis overactivation. DA neurons in the PFC and ventral tegmental area (VTA) form the mesocortical pathway which regulates reward-seeking behaviors and the HPA axis. Glucocorticoids from the HPA axis can also influence the DA system either directly or indirectly, via epigenetic control and MAOs inhibition, respectively. Taken together, L. plantarum may regulate both the DA system and HPA axis by attenuating glucocorticoid-induced MAOs activity.

29Recently, it was discovered that oral gavage of F. prausnitzii exerted antidepressive and anxiolytic effects in chronic-stressed mice. F. prausnitzii, as the sole species of Faecalibacterium genera, represents around 5% of the total human gut microbiota. Low populations of F. prausnitzii correlated with the disease severity of those with MDD and bipolar depression. In a recent large cohort study, fecal levels of F. prausnitzii correlated negatively with depressed mood and positively with quality of life. Therefore, F. prausnitzii seems to have pertinent contributions to mental health.

30Faecalibacterium prausnitzii produces butyrate in large quantities from fermenting glucose and fiber. F. prausnitzii also secretes microbial anti-inflammatory molecules that suppress the proinflammatory nuclear factor (NF)-κB pathway in IECs. These immunomodulatory effects are consistent with neurochemical changes observed in F. prausnitzii-treated depressed mice, whereby cecum SCFAs and plasma IL-10 levels increased, while corticosterone and IL-6 levels decreased. Moreover, intragastric administration of F. prausnitzii decreased colonic cytokine levels and intestinal permeability in mice with colitis. Thus, butyrate produced by F. prausnitzii potentially strengthens the intestinal barrier (similar to L. plantarum; Figure 2B). However, whether local immunomodulatory effects of F. prausnitzii extend to the brain remains unknown. Nevertheless, the ability of F. prausnitzii to attenuate gut inflammation is sufficient to reduce depressive- and anxiety-like behaviors in mice.

31Lactobacillus helveticus intake enabled the recovery of chronic- and subchronic-stressed rodents from their state of depression. Probiotic sticks containing L. helveticus, in addition to Bifidobacterium longum, reduced clinical depression and depressive-like symptoms in MDD patients and healthy individuals, respectively. Most of the animal and human studies also showed that L. helveticus intake enhanced memory and, sometimes, attention and learning. Cognitive impairments, such as poor memory and concentration, represent one major cluster of MDD symptoms. Evidence suggests that L. helveticus may modulate the central NE system and HPA axis to improve cognition, and the central 5-HT system and BDNF expression to reduce depression (Figure 2A).

32Supplementation of L. helveticus improved memory and cognitive performance in chronic-stressed rats, comparable to the SSRI citalopram-treated rats. This memory improvement correlated with increased plasma IL-10 and hippocampal NE levels, and reduced plasma corticosterone and ACTH levels. A previous study also showed that ingestion of L. helveticus enhanced memory and mitigated gut inflammation in neuroinflammation-induced rats. However, another study reported that memory improvement in L. helveticus-treated mice did not correlate with the state of gut inflammation. Despite this discrepancy, it is well established that the hippocampal NE system and HPA axis both interact to regulate hippocampal glucose metabolism for memory consolidation. This mechanism may be affected by microbial NE as L. helveticus produced NE in vitro in amounts that exceed the human bloodstream. It was also shown in vivo that gut bacteria are responsible for converting conjugated NE into its biologically active form. This neuroactive NE likely influences the MGB axis, but the exact mechanism remains unknown.

33showed that elevated hippocampal 5-HT levels correlated with reduced depression severity in L. helveticus-fed rats. The same study also demonstrated that treatment with SSRI citalopram alleviated depression and increased hippocampal BDNF expression and 5-HT levels. Hence, the antidepressive mechanism appears similar between L. helveticus and citalopram. Cultures of L. helveticus produced 5-HT at concentrations close to that in the human bloodstream. As shown in vivo, the gut microbiota has an indispensable function in deconjugating glucuronide-conjugated 5-HT to generate their free, biologically active counterparts in considerable amounts. It is hypothesized that gut luminal 5-HT may sensitize 5-HT 3A receptors of enteric neurons by stimulating the glial cell-derived neurotrophic factor of IECs. 5-HT3 receptors are also expressed on IECs and vagal afferents. Therefore, it can be speculated that L. helveticus influences the central 5-HT circuitry via the neural route. This is supported by a recent study showing that L. helveticus intake increased expression of 5-HT 1A receptors in the nucleus accumbens while restoring behaviors of depressed mice.

34Chronic-stressed mice that ingested L. helveticus displayed an increase in hippocampal BDNF levels and neurogenesis in the nucleus accumbens. Nucleus accumbens is a brain region implicated in reward behavior. The central BDNF and 5-HT systems are synergistic, whereby 5-HT upregulates hippocampal BDNF–TrkB signaling to increase expression and synthesis of BDNF. The elevated BDNF, in turn, facilitates neurogenesis of 5-HT neurons. Therefore, L. helveticus likely increases hippocampal BDNF levels via modulation of 5-HT circuitry, in a similar manner to SSRIs.

35Dietary intervention of heat-killed L. paracasei prevented mood deterioration in times of stress in healthy individuals. In corticosterone-induced depressed mice, oral gavage of either live or heat-killed L. paracasei exhibited antidepressive efficacy equivalent to or better than fluoxetine. The same study also showed that live and heat-killed L. paracasei operated via different mechanisms. Live L. paracasei increased 5-HT levels whereas heat-killed L. paracasei increased DA levels in the brain. The signaling mechanism of L. paracasei appears independent of the HPA axis or vagus afferents. The remaining evidence suggests that L. paracasei potentially functions via an immune-mediated humoral pathway.

36Lactobacillus paracasei secretes lactocepin, a PrtP-encoded serine protease, that selectively degrades proinflammatory chemokines in inflamed ileal tissue of mice. Lactocepin is most likely a heat-labile cell surface protein unique to L. paracasei. Mice fed with live L. paracasei exhibited lower inflammatory markers in serum, such as increased IL-10 and glutathione peroxidase and decreased TNF-α and MCP-1. Another study showed that oral gavage of live L. paracasei with its bacterial products prevented adverse effect of stress on intestinal permeability in rats. This can be linked to a suppressed IDO activity, resulting in higher TRP bioavailability for 5-HT synthesis in the brain. Following this, it was shown that live L. paracasei delivered via gavage increased 5-HT and 5-HIAA (the main metabolite of 5-HT) levels in the hippocampus and striatum of mice. As 5-HT facilitates BDNF synthesis, the upregulated central 5-HT expression presumably explains the accompanying increase in hippocampal BDNF expression of mice alleviated of depression from L. paracasei intake. Therefore, L. paracasei may upregulate the central 5-HT-BDNF system (similar to L. helveticus; Figure 2C).

37Treatment of live L. paracasei also increased fecal Bifidobacterium populations while normalizing behaviors of depressed mice. The gut microbiota profile, inflammatory markers and levels of acetate and butyrate were improved in IBS patients supplemented with live L. paracasei. Reduction in systemic inflammation, coupled with an improvement in hippocampal function, was also observed in obese rats fed with live L. paracasei. Thus, live L. paracasei may facilitate the colonization of butyrate-producing bacteria to reduce systemic inflammation (similar to L. plantarum) and increase 5-HT secretion from ECs (similar to Bifidobacterium infantis; Figure 2B).

38In naïve rats, intake of B. infantis was shown to alter depression-related biomarkers. The same group later showed that chronic-stressed mice no longer displayed depressive-like behaviors after B. infantis intake. In flood victims with IBS, B. infantis consumption did not affect their IBS symptoms but improved their mental health instead. B. infantis did not influence corticosterone levels in mice, implying that the effect of B. infantis is likely to be independent of the HPA axis. Evidence suggests that B. infantis has immunomodulatory effects that regulate the central NE system. A recent study also provided support for the antidepressive mechanism of B. infantis that involves the hippocampal 5-HT system.

39Bifidobacterium infantis treatment manifested two physiological changes in vivo. First, B. infantis decreased plasma IL-6 levels in mice and patients with inflammatory conditions. In depressed mice, the IL-6 release also correlated positively with the severity of depression. Second, B. infantis increased NE levels in the murine brainstem containing the majority of NE neurons. Therefore, B. infantis likely regulates plasma IL-6 and central NE system to exert an antidepressive effect.

40Bifidobacterium infantis secretes bioactive factors (probably polysaccharides) that enhance transepithelial resistance of IECs. Other studies involving rodents also showed that B. infantis treatment enhanced the intestinal barrier by strengthening the formation of tight junction proteins and anti-inflammatory activities of immune cells. Indeed, bacterial DNA translocation from the gut lumen into the circulation was reduced in B. infantis-fed rodents. Bacterial DNA is a potent inducer of TLRs which facilitate the release of proinflammatory cytokines, including IL-6. Administration of IL-6 induced depression in mice, and this outcome was prevented by pharmaceutical blockage of NE neurons in the brainstem. Hence, B. infantis potentially modulates the NE system via an immune-mediated humoral route to reduce depression (Figure 2D). This mechanism appears to be independent of the vagus nerve as oral gavage of B. infantis also decreased proinflammatory cytokine (including IL-6) levels in vagotomized mice with an inflamed colon.

41Treatment of B. infantis upregulated mRNA expression of Tph1 in RIN14B cells, a cell line that mimics ECs. Tph1 converts TRP to 5-hydroxytryptophan (5-HTP) and aromatic amino acid decarboxylase subsequently converts 5-HTP to 5-HT. B. infantis-fed mice displayed reduced depressive-like behaviors, along with an increase in TRP biosynthesis and hippocampal 5-HT and 5-HTP levels. In the same study, B. infantis increased cecum butyrate levels and the abundance of butyrate-producing Bifidobacterium. The elevated butyrate levels also correlated with increased hippocampal 5-HTP and PFC BDNF levels. This could be due to the ability of butyrate and other SCFAs to increase Tph1 activity of ECs, thereby promoting 5-HTP and 5-HT secretions. This is consequential as ECs contribute about 95% of the bodily 5-HT, and that mice with a gut microbiota had 2.8-fold higher plasma 5-HT levels than GF mice. The evidence for the ability of 5-HT to cross the BBB is conflicting. In contrast, 5-HTP readily crosses the BBB and can be converted into 5-HT. Therapeutic 5-HTP has also been shown to treat clinical depression with a potency equivalent to or better than SSRIs.

42Furthermore, B. infantis intake increased plasma TRP levels in healthy rats, but another study with chronic-stressed rats reported otherwise. The author then suggested that B. infantis regulates TRP metabolism differently, depending on the rat strain. Therapeutic TRP can improve symptoms of mood, sleep and cognitive disorders as TRP readily passes through BBB to regulate numerous brain functions, such as 5-HT synthesis. The elevated plasma TRP levels from B. infantis intake is most likely a result of reduced proinflammatory cytokines, which reduces IDO activity and prevents over-catabolism of TRP. Thus, B. infantis may upregulate the hippocampal 5-HT system via modulation of peripheral 5-HTP, 5-HT and/or TRP levels. As 5-HT promotes BDNF synthesis, this presumably explains the concomitant increase in BDNF levels in PFC of rats ameliorated of depression with B. infantis treatment. Taken together, L. helveticus, L. paracasei and B. infantis upregulate the central 5-HT-BDNF system as their mutual antidepressive mechanism, although via different pathways (Figure 2B).

43Treatment of C. butyricum improved depressive-like behaviors in chronic-stressed mice. These treated mice also showed upregulated central 5-HT, BDNF and GLP-1 receptors in the brain. Remarkably, the combination of C. butyricum with antidepressants reduced depression in about 70% of treatment-resistant MDD patients, of which 30% achieved remission. These studies support the antidepressive efficacy of non-pathogenic C. butyricum. It should be noted that certain strains of C. butyricum are pathogenic which may cause botulism and necrotizing enterocolitis.

44Clostridium butyricum, as a resident of healthy gut microbiota, produces butyrate from carbohydrate fermentation. Treatment of C. butyricum increased central 5-HT levels and BDNF expression in mice with reduced depression. Another study also reported that C. butyricum intake upregulated neurogenesis-related pathways, such as BDNF, via butyrate production in mice. Additionally, intragastric inoculation of C. butyricum increased intestinal secretion of GLP-1 and the central expression of GLP-1 receptors in mice alleviated from depression. This effect may also be mediated by butyrate as SCFAs can bind to receptors expressed on intestinal L cells to stimulate GLP-1 secretion into the bloodstream. GLP-1 is known for appetite and glucose control, but the activation of central GLP-1 receptors has been shown to regulate the central 5-HT system and reduce anxiety- and depressive-like behaviors in rats. Therefore, antidepressive mechanism of C. butyricum potentially involves a butyrate-mediated upregulation of central BDNF-5-HT system (similar to L. paracasei and B. infantis) and GLP-1 receptor expression (Figure 2B).

45Lactobacillus kefiranofaciens is isolated from kefir, a type of fermented milk. Oral gavage of L. kefiranofaciens improved behaviors of chronic-stressed, depressed mice. These treated mice also showed several physiological alterations. Levels of circulating TRP, splenic IL-10 and beneficial gut bacteria (e.g., Lachnospiraceae, Bifidobacteriaceae, and Akkermansia) increased, and KYN/TRP ratio, splenic IL-6 and IFN-γ levels and Proteobacteria abundance decreased. What factors mediate such broad effects of L. kefiranofaciens on the TRP/KYN pathway, immune system, HPA axis and gut microbiota remain unclear, but exopolysaccharide is potentially a candidate (Figure 2D).

46The only known metabolite of L. kefiranofaciens is an exopolysaccharide called kefiran. The intake of kefiran modulated the gut mucosal immune system of mice, which could potentially account for changes in splenic cytokines seen in depressed mice. Kefiran was also shown to protect human enterocyte cell lines from adhesion and damage inflicted by toxins of pathogenic bacteria. A further study discovered that L. kefiranofaciens produces a novel exopolysaccharide (not kefiran) that is bactericidal toward enteropathogens Listeria monocytogenes and Salmonella enteritidis. It may be possible that the antibacterial effects of this exopolysaccharide extend to other species in the gut microbiota. This supports the finding that L. kefiranofaciens supplementation ameliorated depressive-like behaviors in chronic-stressed mice by regulating gut microbiota content, which included the decreased abundance of Proteobacteria, a phylum that includes pathogens such as Salmonella. Other mice studies also supported the role of L. kefiranofaciens in modulating gut microbiota composition. Collectively, these changes in gut microbiota profile prevent gut dysbiosis that could lead to chronic inflammation, HPA axis overactivity and depression.

47Bifidobacterium breve treatment improved symptoms of depression in innately anxious mice, chronic-stressed mice and schizophrenic patients with depression. B. breve supplementation also improved mood and cognition in elderly people with mild cognitive impairment. However, none of the accompanying physiological changes among these studies overlapped, making it difficult to identify an exact mechanism of B. breve. In spite of this, one study demonstrated that antidepressive mechanism of B. breve involves the generation of benzoic acid (; Figure 2D).

48Among the 18 bacterial strains isolated from gut microbiota, B. breve was the most efficient converter of albiflorin to benzoic acid via microbial carboxylesterase, at the rate of 75% as compared to L. casei, Lactobacillus acidophilus and B. longum at about 5%. The same study further showed that orally administered benzoic acid alleviated depression in mice. Benzoic acid readily crosses the intestinal barrier and BBB to inhibit D-amino acid oxidase that catabolizes D-serine, a co-agonist of N-methyl-D-aspartate receptor (NMDAR, a type of Glu receptor). Both D-serine and NMDARs are therapeutic targets in neuropsychiatric disorders, such as depression, schizophrenia and cognitive impairment. Indeed, a dysfunctional Glu system is linked to the pathophysiology of depression. In line with this, B. breve intake increased Glu synapses in chronic-stressed mice while treating its depressive-like behaviors.

49Bifidobacterium longum treatment decreased depressive-like symptoms in innately anxious mice and IBS patients with mild to moderate depression and/or anxiety. B. longum supplementation also presented anxiolytic efficacy in numerous human and animal studies. However, B. longum did not affect the gut inflammatory state in animals and humans, indicating a lack of immunomodulatory function. Other physiological changes, such as BDNF expression and plasma KYN/TRP ratio, seen in B. longum-treated mice and humans were inconsistent. Collectively, these data suggest that brain neural activity and HPA axis are possible targets of B. longum signaling mechanisms (Figure 2C).

50Both in vitro and in vivo studies showed that B. longum weakened the excitability of murine myenteric neurons. Mice with inflamed intestines that were fed with B. longum demonstrated reduced anxiety-like behaviors, and this effect ceased upon vagotomy. Intriguingly, B. longum intake also alleviated anxiety in colon-inflamed mice that were vagotomized before treatment. The author postulated that vagus afferents are an essential conduit when B. longum signals enterocytes, but not colonocytes. The genome of B. longum encodes serpin, a serine protease inhibitor. Serpin can inhibit the activation of enteric neurons by suppressing the secretion of elastase-like proteases from IECs. These studies support the premise that B. longum interacts with the host via the neural pathway (similar to L. rhamnosus). Following this, the neural activity and HPA axis of the brain may be altered. Individuals consuming B. longum had increased neural activity in the PFC and decreased neural activity in the amygdala and fronto-limbic regions. Anomalies in the anatomy and activity of the amygdala and PFC are also commonly observed among depressed patients. Furthermore, B. longum intake exerted simultaneous glucocorticoids-lowering and anxiolytic effects in humans and mice, suggesting that B. longum potentially modulates the HPA axis.

51Supplementation of L. gasseri improved mood and depressive-like symptoms in stressed individuals. However, no studies have evaluated the effect of L. gasseri on clinically depressed individuals. Interestingly, L. gasseri is the only dietary probiotic which showed consistent sleep-enhancing effects in humans. Irregular sleeping patterns are frequently associated with MDD, supporting the use of L. gasseri as a potential treatment for MDD-related sleep disturbances.

52Stressed individuals that were given probiotic-based milk containing either heat-killed or live L. paracasei showed alterations in the gut microbiota profile. Heat-killed L. gasseri decreased Bacteroides vulgatus and increased Dorea longicatena populations, whereas live L. gasseri decreased growth of inflammatory Enterobacteriaceae and Veillonella. Both studies also showed that L. gasseri enhanced sleep quality of participants. Another study reported that heat-killed L. gasseri (in milk) increased the population of Clostridium cluster IV group and SCFAs levels in individuals with altered bowel movements. Using a similar methodology, decreased Clostridium cluster IV and increased Bifidobacterium populations were found in another group of participants. Taken together, these results suggest that heat-killed L. gasseri does not have a specific microbial target, but rather modifies the preexisting gut microbiota that is unique to each individual. Nevertheless, these changes in the gut microbiota composition favor an anti-inflammatory state. L. gasseri likely alters the gut microbiota profile through its unique, heat-resistant gassericins A and T with potent antibacterial properties against enteric pathogens.

53Heat-killed L. gasseri decreased expression of leukocytic stress-responsive microRNAs and salivary cortisol levels in stressed individuals. L. gasseri intake also prevented downregulation of EIF2-related genes in IBS patients. These studies suggest that L. gasseri confers protection against detrimental effects of stress. Moreover, heat-killed L. gasseri intake promoted parasympathetic nerve activity while improving sleep quality of stressed individuals. In healthy individuals, administration of either live or heat-killed L. gasseri increased their parasympathetic activity. Therefore, L. gasseri may modify the gut microbiota profile in such a way that lowers gut inflammation and stress response, which may consequently promote parasympathetic activity to facilitate sleep (Figure 2C).

54The existence of different gut microbiota compositions, depression subtypes and probiotic formulations complicate treatment outcomes and necessitate an individualized approach when using probiotics to treat depression. Despite these challenges, probiotics confer some benefits over antidepressant drugs, and there are more promising candidate probiotics that can potentially treat depression.

55Several factors are known to influence the gut microbiota composition, such as diet, medications, genetics, age, geographical location and smoking. Recently, approximately 1000 gut-derived putative bacterial species that do not belong to any existing genus were discovered in humans. Such tremendous diversity complicates the understanding of how introduced probiotics affect the overall gut microbiota. One study showed that tolerability of individuals’ gut microbiota toward the colonization of probiotics ranges from permissive to resistant. This appears to depend on the baseline abundance of probiotic species in the host gut microbiota. For instance, those who were permissive toward the colonization of Lactobacillus had prior low levels of Lactobacillus populations before treatment. Similarly, B. longum colonized the gut for a longer period in 30% of users who initially had low levels of B. longum. Another study showed that the antidepressive effect of multi-species probiotics (MSP) only manifests when the administered MSP successfully colonized the gut of rats. This is consistent with the observation that lower levels of two main probiotic genera, Lactobacillus and Bifidobacterium, are commonly found in individuals with MDD.

56Despite most studies supported the effectiveness of probiotic supplements in reducing depression, not all randomized controlled trials reported the same outcome (Table 2). For instance, L. rhamnosus did not affect scores of anxiety, depressions, sleep, cognition, inflammatory and stress responses among healthy adults. L. rhamnosus also did not affect perceptions of wellbeing, anxiety and stress among healthy older adults. In healthy individuals, L. helveticus exhibited no antidepressive effect. These results imply that probiotics are less efficacious among the healthy population, which agree with a meta-analysis that reported an insignificant effect of probiotics on mood, particularly in healthy individuals. Therefore, probiotics could be generally more effective in colonizing gut microbiota of depressed individuals that are different from healthy people. In some cases, probiotic colonization may be optional for their effects to manifest. For instance, heat-killed L. paracasei benefited the human and animal host, in terms of neurochemical and behavioral changes. Some probiotics, such as L. reuteri, L. paracasei, L. plantarum, L. gasseri, L. kefiranofaciens, B. breve, and B. infantis, promoted the colonization of other beneficial microbes that contributed to the reduction of depressive-like symptoms in animals.

57Major depressive disorder is characterized by depressed mood and/or anhedonia, in addition to excessive guilt, suicidal ideation, changes in appetite and sleep, psychomotor retardation, poor concentration and fatigue. From these diagnostic criteria, approximately a thousand combinations of symptoms and 19 depression subtypes can be derived. These subtypes of depression are often grouped as a single term, namely depression, which should not be the case when evaluating therapeutic potential of probiotics.

58Some associations can be drawn by matching behavioral benefits of probiotics to the characteristics of depression subtypes (Table 2). For instance, the sucrose preference test in rodents reflects the anhedonia subtype. Probiotics that have been shown to improve the outcome of this test include L. helveticus, L. plantarum, L. paracasei, L. kefiranofaciens, B. infantis, and B. breve. Among these probiotics, L. plantarum and L. paracasei also modulated the central DA system, whereas B. infantis and B. breve upregulated tyrosine (precursor to DA) biosynthesis. An impaired DA system represents the hallmark pathophysiology of anhedonia. This provides a proof of concept that these probiotics may be effective in treating anhedonia.

59Somatic depression subtype is characterized by psychomotor agitation/retardation (i.e., locomotion), changes in weight/appetite, insomnia/hypersomnia and fatigue without physical exertion. Probiotics that improved locomotor activity of rodents include L. plantarum, L. helveticus and L. brevis. Intake of L. brevis increased sleep duration in healthy mice, and L. gasseri enhanced sleep quality in medical students with mild depression. L. rhamnosus supplementation modulated appetite-associated genes and attenuated appetite in zebrafish. In combination with prebiotics, L. rhamnosus exerted antidepressive effect and appetite control in obese individuals. Hence, symptoms of somatic depression are rather distinct and may be improved differently with different probiotics.

60Cognitive depression subtype is distinguished by poor concentration and memory function as well as indecisiveness. Behavioral assessments for memory function in mice include the Morris water maze, Barnes maze and other behavioral tests. Administration of probiotics including L. helveticus, L. plantarum, and L. paracasei enabled animals to perform these memory test more effectively. Attention, memory and learning behaviors in humans are assessed by cognitive tests, such as the Stroop, verbal-learning and digit-symbol tests. Improvements in these tests have been shown with the intake of (1) L. helveticus and B. longum in healthy adults; (2) L. plantarum in MDD patients and stressed adults with mild depression; and (3) B. breve in elderly with mild cognitive impairment. Thus, some probiotics appear to improve cognition regardless of depression.

61Anxious depression subtype refers to major depression that comorbid with high levels of anxiety. In mice, anxiety can be measured by behavioral tests, such as the elevated plus maze and open field tests. In humans, anxiety is generally assessed with questionnaires. Probiotics that exhibit anxiolytic effect include L. rhamnosus, L. helveticus, L. plantarum, B. longum and B. breve. Moreover, MSPs intake often decreased depression and anxiety simultaneously in randomized controlled trials.

62Conventional SSRIs that target the 5-HT system often fail to treat anhedonic patients and, in some cases, worsen their symptoms. Antidepressant drugs (e.g., SSRI and SNRI) are also ineffective against other depression subtypes, namely the somatic, cognitive and anxious depression. Therefore, certain probiotics may serve as an adjuvant or alternative treatment for MDD and its subtypes. A pilot study showed that MSP, together with a magnesium supplement, decreased depression in SSRI treatment-resistant patients. A clinical trial also reported that the combination of B. longum and L. helveticus decreased depression in MDD patients with prior use of standard antidepressants.

63In studies that investigated behavioral effects of probiotics, about 60% of animal studies and 50% of human studies used single-species probiotics (SSPs). Studies with SSPs promote a better understanding of the function and contribution of individual probiotic, which is difficult to measure in MSPs. However, MSPs may have higher potency in humans. In MDD patients, SSP (L. plantarum) did not reduce depression but improved cognition, whereas MSPs had repeatedly shown antidepressive efficacy. MSPs often gave better therapeutic efficacy compared to that of SSPs in gut-related disorders and pathogen infections, which could be explained by an overall higher dosage. Indeed, MSPs with a higher dosage improved symptoms of depression and anxiety in healthy individuals compared to that of a lower dosage. MSPs are also hypothesized to exhibit synergistic effects that would have an expanded effect on the host physiology. In contrast, SSPs are speculated to promote better colonization as it does not have to compete for nutrient or adhesion sites in the host. This highlights the need for more studies to understand how probiotics in MSPs interact with each other and with existing gut microbiota, and which probiotic(s) is suitable in formulation of MSPs for antidepressive efficacy.

64Probiotics are generally safe for consumption, except for immune-compromised and critically sick individuals wherein probiotics may cause sepsis, pneumonia, endocarditis and allergies. Still, it has been viewed by some that more human trials are required to establish the dosage efficacy and long-term safety profile of probiotics. For antidepressant drugs such as SSRIs, side effects occur in 40-60% of users which include sexual dysfunction, suicidality, emotional numbness and addiction. A meta-analysis data showed that users of antidepressant drugs were associated with a 33% increased risk of mortality. On the other hand, probiotics possess fewer side effects than antidepressant drugs. For instance, rats fed with L. brevis-fermented milk exhibited comparable antidepressive efficacy to fluoxetine-treated rats, but without side effects of fluoxetine (decreased appetite and weight loss).

65Antidepressant usage is also associated with stigma, such as being perceived as emotionally weak and dependent on drugs, which contributes to the disease severity and poor adherence to treatment. In a survey study, 77% of depressed patients prefer to hide their use of antidepressant medication from others. However, the prevalence of perceived stigma against antidepressants differs based on the population studied. To this end, probiotics may help as an alternative treatment for depression, given that probiotics have not been associated with any perceived social stigma.

66Bifidobacterium pseudocatenulatum is known for its regulation of obesity-related changes in metabolism and the immune system. B. pseudocatenulatum intake reversed diet-induced obesity, depression, high corticosterone and low hippocampal 5-HT levels in mice. However, a high-fat diet model is meant to study the pathophysiology of obesity and type 2 diabetes. It is, thus, unclear if B. pseudocatenulatum would decrease depression in mice without obesity. Another study showed that anxiety-like behaviors diminished in chronic-stressed mice fed with B. pseudocatenulatum, but depressive-like behaviors were unevaluated. Therefore, further studies are required to determine whether B. pseudocatenulatum has an independent antidepressive effect.

67Bacillus coagulans supplementation relieved symptoms of both IBS and depression in patients diagnosed with IBS and MDD. This clinical recovery is accompanied by a decrease in serum myeloperoxidase, an inflammatory marker. However, patients might have experienced less depression as a result of reduced IBS symptoms. Interestingly, B. coagulans intake increased levels of circulating IL-10, fecal F. prausnitzii and SCFAs in older adults. As F. prausnitzii and butyrate are associated with antidepressive properties, B. coagulans may also indirectly reduce depression and improve gut health.

68Bifidobacterium bifidum and L. acidophilus were often included in the formulation of MSPs to treat depressive symptoms in patients with MDD and other health conditions, such as polycystic ovarian syndrome, multiple sclerosis and IBS. Surprisingly, B. bifidum and L. acidophilus have not been tested independently for its antidepressive effect. B. bifidum intake improved mood and reduced symptoms of abdominal pain, diarrhea and constipation in patients with gastrointestinal disorders. However, the mood elevation could be due to recovery of gastrointestinal symptoms rather than effect of probiotics solely. Both in vitro and in vivo models showed that L. acidophilus protects the intestinal barrier integrity by preventing pathogen adherence and release of proinflammatory cytokines. Taken together, B. bifidum and L. acidophilus potentially exhibit antidepressive effect and their direct influence on depression warrants further investigation.

69Bacteroides fragilis has been proposed as a potential probiotic, although its pathogenicity needs to be taken into consideration. B. fragilis secretes polysaccharide A and expresses sphingolipids that benefit the host gut health and immune system. Bacteroides genus is likely to be the largest GABA producer amongst human gut microbiota, with B. fragilis produces GABA at low pH. The same study also found that neural patterns of a typical MDD patient correlated with low fecal levels of Bacteroides. Hence, antidepressive potential of B. pseudocatenulatum, B. coagulans, B. bifidum, L. acidophilus, and B. fragilis warrants further investigation. It is also worth noting that Bifidobacterium adolescentis’s antidepressive capability may be a new probiotic candidate. Evidently, an increasing number of probiotics are being presented as a potential treatment for depression. This provides a wide repository of available probiotics, with different species combinations, that can be assessed for clinical efficacy against depression.

70The MGB axis enables the bidirectional communication between the gut microbiota and the brain. When this axis becomes maladaptive, the host physiology is adversely affected which may lead to the development of depression. Probiotics have shown clinical efficacy in the treatment of depression by modulating the MGB axis. Yet, the complexity of gut microbiota and heterogeneity of depression presents a challenge to explain the underlying mechanisms that contribute to this clinical efficacy. Nonetheless, cumulative evidence suggests the therapeutic potential of probiotics for certain depression subtypes, with fewer side effects and less stigma compared to standard antidepressants.

71Limitations of this review include: (1) inferences of probiotic mechanisms were derived from preclinical and in vitro data; (2) interactions of probiotics with other members of gut microbiota were unexplored, therefore the mechanisms of MSPs was unable to be explored; (3) strain-specific effects of bacterial species were neglected; (4) potential applications for probiotics for depression subtypes are hypothesized, however, clinical evidence is limited; (5) effect sizes of probiotics as antidepressants was not evaluated. Notwithstanding these caveats, this review adds further understanding to the potential antidepressive effects and therapeutic potentials of probiotics. stated that it is imperative to grasp the underlying molecular mechanisms of the MGB axis, and which microbial populations are pertinent for this intervention, to advance the marketability of probiotics.

72SY wrote and edited the manuscript. TT, JC, and WL conceptualized and edited the manuscript. All authors contributed to the intellectual input and critical revision of the manuscript.

73The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.