The term “gut-brain axis” refers to the bidirectional communication between the bacteria in the gastrointestinal tract and the brain. Various animal studies have demonstrated that mediators produced by bacterial populations in the gut can alter the brain directly affecting neuroplasticity, epigenetics, and gene expression, and indirectly by influencing endocrine signalling through actions on the gut epithelium [1]. Similarly, other research in the field has demonstrated that neuroactive agents and other metabolites produced by bacteria can affect the development of the enteric nervous system and may even influence the CNS by signalling through the vagus nerve [1]. Current research in the field has tried to elucidate the role of the gut-brain axis in the pathogenesis and severity of numerous mood disorders such as MDD, however, the mechanisms that underlie MDD are still unknown and there are no established biomarkers for use in the clinical setting.
Cross-sectional studies have been fundamental in understanding the significant differences in the gut microbiota composition between healthy individuals and MDD patients. According to a systematic review by Nudsen et al., across 17 different studies, there was a clear difference in gut microbiota composition between MDD patients and healthy controls [2]. However, the species found to be more prevalent or less prevalent varied across studies, consistent with findings in other systematic reviews on this topic [3]. This has proven to be a significant issue as the inability to identify causal relationships has prevented the identification of specific bacterial species and pathways associated with MDD. Additionally, transplanational studies such as one by Zheng et al. aid in demonstrating that depression-like behaviours can be induced via transplantation of germ-free mice with microbiota derived from MDD patients [4].
As a result, our team decided to focus on examining the changes metabolites produced in the gut-microbiota and their differential expression in MDD patients relative to controls.
Short chain fatty acids (SCFAs) are the main products of anaerobic fermentation by microbiota in the large intestine. After production, most are absorbed and metabolized by colonocytes, with a smaller amount being sent into systemic circulation within other tissues; as a result there is a measurable presence of SCFAs in faecal samples. Major SCFAs include acetate, propionate, and butyrate whose roles in the gut consist of maintaining the integrity of the intestinal barrier, mucus production, and inflammation protection, and in the brain consist of maintaining blood brain barrier integrity and development, and preservation of the central nervous system’s (CNS) homeostasis [5]. There are multiple butyrate producing bacteria found in the gut which are associated with higher quality of life indicators, but depleted in those with depression [6,7]. The genera include Faecalibacterium, Coprococcus, Ruminococcus, and Roseburia [6,7].
SCFAs can be detected in serum and plasma but since most are found in the colon, faecal samples are preferred [8]. A study on colorectal cancer using stool sample profiles detected a difference in levels of butyrate producing species with pyrosequencing in addition to differences in SCFAs using gas chromatography-mass spectrometry [9]. In plasma and serum samples, butyrate and propionate are found in trace amounts, while acetate is ubiquitous, making it difficult to detect [10]. A variety of methods are implemented to extract derivatized SCFAs such as hollow fibres and gas or liquid chromatography [10]. Moreover, it was discovered that different tubes containing different additives yielded a variety of SCFA concentrations [10]. For instance, blood samples in tubes with polyacrylamide separator gels such as PST and SST showed higher concentrations of butyrate and propionate, whereas tubes with EDTA like EDTA K2 tubes had higher acetate concentrations [7]. In short, these complications make it challenging to accurately measure concentrations of SCFA levels in plasma and serum, resulting in discrepancies across studies.
A study published in 2018, demonstrated that faecal samples of Polish women with depression had significantly lower concentrations of SCFAs compared to those who were non-depressive [11]. SCFAs also seem to affect neurotransmitter levels, with acetate playing a potential role in appetite regulation through the glutamate-glutamine transcellular cycle, which increases the production of lactate and GABA [12]. Propionate and butyrate, on the other hand, seem to control some level of tryptophan 5-hydroxylase-1 and tyrosine hydroxylase expression, which aid in the synthesis of serotonin and the catecholamines, dopamine, noradrenaline, and adrenaline respectively [13]. It was also found that acetate metabolism decreases lipopolysaccharide (LPS) induced microglial inflammatory signalling of IL-6, IL-1β, and TNFa expression, all of which are related to the pathogenesis of certain neuroinflammatory diseases [14]. Furthermore, they found decreased phosphorylation levels of NF-κB, and the MAPK, p38 and JNK, which regulate pro-inflammatory cytokines [14].
However, compared to acetate and propionate, butyrate is more widely studied in relation to depression. In particular, butyrate is involved in epigenetic mechanisms by way of histone deacetylase inhibition which promotes processes such as apoptosis and suppress proliferation [15]. By this mechanism, sodium butyrate was also found to decrease activation of LPS induced hippocampal microglia, explaining its antidepressant-like effects [16]. Additionally, it decreases pro-inflammatory cytokine secretions, which relates to neuroinflammation, a trait seen in mouse models for depression [16]. The Flinders Sensitive Line (FSL) is an established model for depression in rats, and when used to test rats treated with NaB during a behavioural despair test, they found a decrease in immobility time, demonstrating antidepressant-like outcomes. Similarly, another study revealed that in mice induced with chronic and unpredictable stress, treatment with NaB could ameliorate anhedonia-like symptoms, promote locomotive behaviours, and lessen immobility time behavioural despair tests in [17]. Treatment with NaB also heightened the concentration of serotonin and the expression levels of BDNF and occludin, both of which are related to BBB repairment [17].
Many interactions have been found between butyrate and tryptophan. For one, research demonstrated that butyrate reverses the effects of oxidative stress on tryptophan uptake in human fibroblast cells [18]. Additionally, butyrate downregulates indoleamine 2,3-dioxygenase 1 (IDO-1), an enzyme involved in tryptophan metabolism that oxidizes L-tryptophan into N-formylkynurenine [19]. The mechanism of action involves histone deacetylase inhibition and decreasing STAT1 expression which inhibits IFNγ-dependent and phosphoSTAT1, thus preventing IDO-1 transcription [19]. Finally, sodium butyrate affects the expression of tryptophan hydroxylase 1, a key enzyme in 5-hydroxytryptamine (5-HT) synthesis in mice [20]. At treatments of 0.5 to 1 mM, TPH1 mRNA expression was significantly increased while treatments between 8 to 16 mM significantly decreased TPH1 transcription [20].
Butyrate can be made from carbohydrates which undergo glycolysis to form acetyl-CoA, leading to butyryl-CoA which undergoes further processing to form butyrate [21]. There are also several amino acids like glutamate, lysine, and cysteine which can be turned into butyrate. Consequently, substrates for butyrate production consist of starch, wheat, oat, rye, and arabinoxylan-rich whole grains [21]. But consumption of substrates like xylan, pectin, and cellulose result in low butyrate levels [21]. Since butyrate is produced by bacteria, antibiotics play a large role in affecting SCFA levels. More specifically, use of antianaerobic antibiotics lowered butyrate and propionate levels [22]. Antibiotic-induced microbiome depletion resulted in a butyrate depletion to undetectable concentrations [23]. Compared to controls, mice which exercised, had an elevated ratio of butyrate to acetate and increased levels of acetate CoA transferase gene, a gene in butyrate producers, responsible for aiding in the conversion of acetate to butyrate [24]. A similar result was found in a study on seniors, in which butyrate levels increased 5.44 mmol/g after 24 weeks of cardiovascular and resistance exercise [25].
Gamma-aminobutyric acid, also known as GABA, is an amino acid-based inhibitory neurotransmitter directed toward inhibiting or reducing signals of the central nervous system (CNS) [26]. Inhibitory neurotransmitters decrease the action potential of a neuron below its threshold potential, preventing that neuron from further exciting neighbouring neurons. Once bound to GABA-receptor proteins, GABA is able to inhibit nerve transmission and reduce excitatory signals to target neurons [27,28]. This produces a relaxing effect on the body, and therefore, the absence or deficiency of GABA is associated with anxiety, stress, depression, and anhedonia [27,28]. Research shows no indication of a scarcity of GABA detecting methods as it can be measured and monitored by tracking CSF GABA levels and examining faecal samples for GABA-producing bacteria [29,30].
Studies conducted using faecal samples of MDD patients and healthy individuals concluded an association between low GABA concentrations and MDD. In MDD faecal samples, there was a negative correlation between high bacteroides concentration and MDD brain response, compared to that of the healthy control [31]. The GABA-producing bacteria, parabacteroides, escherichia, and primarily bacteroides, were found in abundance in faecal samples of the healthy control compared to patients with MDD [31]. Another study further detected that bifidobacterium abundance was associated with high faecal GABA content in healthy human subjects [32]. As a result, previously conducted research illustrated a negative correlation between GABA concentration and depression.
Neurotransmitters, other than GABA, associated with MDD include depleted levels of serotonin, dopamine, and norepinephrine. The focus of this project required a metabolite implicated in MDD whose presence can be easily, accurately, and feasibly detected in faecal samples and has a minimum amount of interactions with the other chosen metabolites to prevent invalid results. In comparison to GABA and serotonin, there was a lack of specific primary research on the implication of the latter two neurotransmitters in the gut microbiota and faecal samples of MDD patients.
Regarding serotonin, research suggested that despite 90% of it being produced by enterochromaffin cells, most of the serotonin produced in the gut is unable to cross the blood-brain barrier and performs many digestive functions, which deviates from our focus on MDD [33]. It was also briefly mentioned that there is an inverse relationship between the amount of serotonin and indole production via the tryptophan pathways in the brain [34]. As such, the amount of circulating tryptophan levels required for the synthesis of serotonin depletes as tryptophan is used for indole formation in the brain [34]. In knowing this, detecting serotonin along with indole, a definite candidate as a metabolite for this biosensing project, may compromise the validity of the results. There is a lack of available research regarding the direct interaction between GABA and indole, however, GABARAP (GABAa receptor-associated protein) has binding sites for indole and indole derivatives though the result of this binding to GABARAP sites remains unknown. In addition, tryptophan residue is found at these GABARAP sites, which were proven to have a high affinity to indole derivatives [35].
Considering that GABA is accompanied by indole and butyrate in our biosensing system, it is vital to investigate any naturally occurring interactions between the three metabolites. If there are any direct interactions, this system would need to be modified to preserve the credibility of its results. An important correlation to note lies between the GABA degradation pathway, the succinate degradation pathway, and butyrate synthesis, which share a direct interaction found in bacteria such as porphyromonas gingivalis and Clostridium difficile [36]. Both degradation pathways share succinate semialdehyde and 4-hydroxybutyrate, to which, adding butyryl-CoA to the next step of the system leads to butyrate production [36]. The same also works for the synthesis of acetate by adding acetyl-CoA [36]. However, it was noted that pH conditions in the gut microbiome vastly impact the journey toward butyrate formation, determining the pathway it chooses to follow. The synthesis of butyrate through these degradation pathways is not a common one and is conducted in specific pH conditions, thus, this did not factor as an issue in the design of the biosensing mechanism [36]. Under hypoxic conditions, butyrate in solution with weak acids of pH 5 can synthesize GABA, however it is unknown whether this process occurs in the body naturally [37]. Interestingly, GABA and succinate degradation share the enzymes succinate semialdehyde and 4-hydroxybutyrate. Upon addition of butyryl-CoA, butyrate was produced, the mechanism of which is proposed to be the presence of a gene related to propionate-CoA transferase [36].
Tryptophan is an essential amino acid that acts as a precursor to several important pathways including the kyrenueric, indole-pyruvate, and serotonin pathways [34]. The gut microbiota plays a key role in regulating the bioavailability of the resulting substrates as it can catabolize tryptophan into the synthesis of metabolites such as indole, serotonin, kynurenine, quinolinate, indole acetic acid (IAA), and indole propionic acid (IPA) [34]. Tryptophan metabolites seem to both, directly and indirectly, alter the function of the gut-brain axis as its consumption by gut bacteria may reduce the level of tryptophan accessible to the brain through the blood-brain barrier or cause imbalance in metabolite production [34]. For example, a study by O’Mahony et al., reported that certain bacterium, such as Pseudomonas, has been shown to synthesize serotonin from available tryptophan, thereby affecting serotonergic neurotransmission and resulting CNS and ENS dysfunction [38]. Low serotonin is a disease specific marker of MDD and is correlated with symptoms such as depression, fatigue, and impaired cognitive and mood functions [39]. Additionally, a 2004 study also showed that depleting tryptophan in patients caused return of MDD symptoms which were not seen in controls [39].
Some literature available on tryptophan metabolites suggests that disturbances to kynurenine and quinolinate levels have been shown to affect brain function and consequently lead to depression-like symptoms. However, this mechanism is not yet clear. One study proposed that in depressed individuals, tryptophan was massively diverted to the production of kyrenueric acid, thus limiting serotonin production in the brain [40]. Another study showed that simultaneously low overall serum kynurenic acid and quinolate acid concentrations have been observed in dMDD (major depressive disorder in current depressive episode) and rMDD (major depressive disorder in current remission) patients when compared to healthy controls [41]. Additionally, it is suggested that relative levels of kynurenic acid and quinolate acid in neurotoxic ratios appear to be a critical pathophysiological mechanism involved in MDD, in contrast to exploration of their absolute values [41]. Although the kynurenic acid pathway has potential for further investigation, it was not chosen as a metabolite for our biosensor due to insignificant stool levels. Indole is the only compound whose stool concentrations showed a significant difference between healthy individuals and MDD patients [42].
Indole is synthesized by tryptophanase, an enzyme that catalyzes the hydrolytic B-elimination of tryptophan to indole [43]. Indole and its derivatives, including IAA and IPA, are implicated in neuropathogenesis by altering CNS metabolism, and thus receiving increased attention in its relationship to disorders such as MDD [44]. Several studies also show a general increase in indole and its derivatives in faecal samples of MDD patients [45]. Derivatives such as IAA were considered as a potential marker as it is found at the end of the indole pathway, however, it was eliminated as IAA is only found in urine and not in faecal samples; therefore, not fulfilling the criteria established when choosing the biological markers for study [44]. Furthermore, E.Coli senses indole in the range of 0.1-10 mM, and the indole concentration of human stools is around 0.25-1 mM [46,47]. Generally, it was found that there was more research conducted on the relationship between indole and MDD compared to indole derivatives and indole was chosen as the last metabolite of measure in our project.
A 2021 study investigating the correlation between gut microbiota and depression suggested that indole increase in the gut microbiome was linked to increased vulnerability to chronic stress and interference with neurotransmitter synthesis, including serotonin, in mice led to anxiety-like and depressive behaviour [44]. Another study by Jaglin et al. demonstrated that this anxiety and depressive-like behaviour in mice, due to continuous overproduction of indole, may have been the result of oxindole augmentation in the brain and increased activation of the vagus nerve (rapid eye movements) [48]. Furthermore, indole supplementation was associated with improved neurological function in rats [49]. Other literature, such as the metadata performed by Skonieczna-Żydecka et al. (2018), showed that increased indole concentrations were present in the stool of Polish women with depressive symptoms, while Naudon et al. (2019) observed 28-877 nmol/g of faecal indole in a cohort of adolescents with MDD [43, 50, 51].
Additionally, due to the high interindividual amplitude of variation of the non-redundant tnA gene, there is a large degree of heterogeneity among healthy human subjects [52]. A study by Brydges et al., showed that only few people in the population carry levels of indole concentration that are recognized as abnormal, either carrying a high-indole or low-indole-producing gut microbiota [52]. Therefore, if the biosensor system model designed can detect indole levels reliably, indole could be effectively utilized as a biomarker for MDD.
As higher concentrations of indole have been observed in MDD patients, we hoped to utilise and develop a receptor system that was specific for indole that excluded its derivatives [45]. Our team was interested in several indole receptor mechanisms aryl hydrocarbon receptors (AhRs). AhR mediates the stimulatory effect of indole on the regulation of neurogenesis and is a more conventional method for indole detection that is more widely studied in literature [53]. However, the AhR receptor system was not chosen for further analysis due to its lack of specificity: thousands of structurally different ligands can bind to this system other than indole [53].
A study by Matulis et al., showed a promising system as they demonstrated the development and characterisation of the indole-inducible gene expression system PpTrpl/PPP_RS00425 [54]. In this system, the tryptophan synthase alpha subunit (TrpA) is an enzyme that catalyses the reversible conversion between indole and I3GP in the body; the gene expression system explored in the study can be activated in the presence of indole and therefore a similar system can be used in our goal to develop an indole-specific biosensor [54]. Furthermore, researchers assembled the PpTrpl/PPP_RS00425 system into the vector pBRC1 to create the plasmid construct pPM0081, which was consequently transformed into E.Coli. Fluorescence activation occurred 60 minutes after the indole was added to the plates [54]. This study summarises that the PpTrpl/PPP_RS00425 system had a strong affinity for indole, and more importantly, other indole derivatives such as L-tryptophan, 3-IAA, 3-IPA, and 3-IBA did not induce the system with 99.0% confidence in E.Coli, pointing to the specificity of this system [54]. Specifically, the PpTrpl/PPP_RS00425 system was shown to be induced up to 639.6-fold by indole. There was no significant growth inhibition of E.Coli by indole [54]. Thus, our team decided to base our receptor system for indole on this paper, as we utilized TrpI in our final design.
With its antidepressant and anti-inflammatory effects, butyrate has been implicated in MDD risk assessment where lower concentrations of butyrate are observed in MDD patients [55]. Therefore, butyrate is respected as a biological marker for MDD in our project as we aim to track and monitor its changing concentration using a microbial biosensor. Our team took inspiration from Enterohemorrhagic Escherichia coli (EHEC) as it senses butyrate to regulate both flagella growth and T3SS effector adhesin [56]. EHECs are also great receptors for butyrate, as its cell membrane is permeable to a variety of SCFAs, including butyrate, allowing for passive diffusion in and out of the cell [57]. In both cell processes, flagella growth and T3SS effector adhesin, butyrate is selectively recognized by a leucine-responsive regulatory protein (LRP), forming a binary complex that binds to the promoter regions of their respective genes [56]. In flagella growth, LRP promotes the expression of flhD and flhC class 1 promoters to produce flhD4FlhC2 transcription activation complex [58]. This is significant as research has shown that this process is regulated solely by butyrate and is independent of other prominent SCFAs such as acetate and propionate [58]. However, with a lack of previous literature outlining the feasibility of this system, our team decided to find one that is both selective and well-researched.
Researchers Bai and Mansell from Iowa State University designed and tested a microbial biosensor focusing on the Ppcha Promoter in E coli Nissle 1917 that is activated by the aforementioned LRP-Butyrate complex [58]. As a result, the PchA regulator after being transcribed binds to the promoterLEE1 to activate the transcription of the ler gene [58]. This ler gene was replaced by the researchers with a green fluorescent protein (GFP) to communicate the concentration of butyrate. Our project took inspiration from this design and centred our butyrate sensing system around this concept of replacing the ler gene to a reporter gene. One limitation with the E Coli Nissle 1917 microbial biosensor is that the production of butyrate concentration will fluctuate as it is mediated by Ecoli’s native type II biosynthetic pathway (FAS II) [58]. Due to budget and resource restraints, our group could not use CRISPR technology to knock out the genes involved in this pathway (frdA, ldhA, adhE, pta) as proposed in the original study and will have to take in account this imperfection in the processing of our raw data.
GABA, or γ-aminobutyric acid, is one of the main neurotransmitters in the human brain [59]. The implementation of a GABA receptor within this system is based on the findings that patients having MDD also show a deficit - or reduced levels - of GABA in their neural tissue and fluids [60]. This strengthens the theory that lack of GABA neurotransmission plays a role in MDD, fuelling the investigation into measurement of GABA concentrations within this receptor system.
GabR and its natural promoter GabT were chosen for this receptor system due to their previously established prevalence in Bacillus subtilis bacteria. Additionally, GabR specifically binds GABA molecules, avoiding any conflict with promiscuous binding [61]. Non-specific binding proves to be an issue in GABA receptors of the attKLM operon, negating this option within the system among other non-specific receptors [62]. The GABA receptor used in this system was taken from a process endogenous to Bacillus subtilis bacteria. The GabR regulator protein and its gabT promoter have been effectively proven to allow B. subtilis to bind and uptake GABA molecules as a source of nitrogen and carbon for cellular processes [61,63]. This system is also presently known to be the best-studied and most prominent GABA regulators [64], improving the validity of its inclusion in the system. As a part of the GabTD operon, activated through formation of a GabR-PLP(pyridoxal-5′-phosphate)-GABA complex to induce the synthesis of glutamate and initiate gabR transcription [65]. Genes of the GabTD operon have previously been engineered into E. coli bacteria by Park et al for the purposes of GABA reception leading to glutaric acid production [66].
Dysregulation of GABA neurotransmission is a contributing factor not only in MDD, but in other mental illnesses such as schizophrenia, anxiety disorders, and in addition, Huntington’s disease [67,68]. In many of these related disorders, GABA concentrations have been shown in recent research to be notably different between the faeces of individuals with these disorders compared to control populations and unaffected individuals [67,68]. Using the outlined receptor system, GABA levels could be measured using this fluorescent protein signalling consortium in the same way as MDD from faecal samples of these individuals. With further research, indicative baseline levels of GABA can be determined in each of these disorders to gain a better diagnostic outlook for these illnesses and other related disorders.