Inflammatory markers associated with neuroinflammation in colitis
Several studies in animal models of colitis have identified inflammatory markers in the hippocampal and cortical brain regions [
126,
135,
136,
137,
138,
139,
140,
141,
142,
143]. There is a significant increase in IL-1β and IL-6 mRNA expression in the cortex and IL-1β and TNF-α in the hippocampus of mice with DSS-induced colitis [
126]. This is accompanied by significantly higher serum levels of IL-6 and TNF-α in mice with colitis [
126,
135]. Moreover, using TNBS-induced model of colitis in rats, Wang et al. (2010) reported intestinal morphological damage, increased myeloperoxidase activity, and increased mRNA and distribution of IL-6 in the inflamed colon and specific regions of the brain including the cerebral cortex and hypothalamus [
144]. Neuroinflammatory changes are considered to be an indicator of alterations in animal behaviour in in vivo models of IBD [
136]. In mice with dinitrobenzene sulfonic acid (DNBS)-induced colitis, a significant increase in the expression of TLR-2 and -4, TNF-α, IL-6 and damage-associated molecular patterns like high mobility group box protein 1 (HMGB1), intracellular signalling proteins such as myeloid differentiation primary response 88, and brain-derived neurotrophic factor (BDNF) was found in the hippocampal regions [
136]. Enhanced innate immune responses in the brains of animal models of colitis have been associated with depressive behavioural traits seen as decreased mobility time in the forced swim and tail suspension tests, decreased grooming in the splash test and sucrose intake in the sucrose preference test [
136]. Moreover, the inflammatory activity associated with anxiety and depression in mice with colitis was accompanied by alterations to hippocampal mitochondrial parameters [
136]. These include decreased antioxidant glutathione (GSH) and adenosine triphosphate levels together with overproduction of reactive oxygen species (ROS) (Fig.
2), suggesting mitochondrial dysfunction and possible oxidative stress in the hippocampus of mice with colitis [
136].
These findings are pertinent in IBD-associated depression given that brain metabolism impairments characterised by mitochondrial dysfunction and the generation of ROS have been implicated in the pathogenesis of depression and anxiety [
145,
146]. Moreover, GSH, a major brain antioxidant that ameliorates oxidative species, is reduced in the prefrontal cortex of MDD patients [
147].
In another study, mice with TNBS-induced colitis showed heightened behavioural despair and increased hippocampal TNF-α, inducible nitric oxide synthase (iNOS), and nitrite expression [
137]. In this study, only male mice were used in order to remove the confounding variable of high estrogen in females given its correlation with elevated serum cytokine production in chemically-induced colitis [
148]. Future studies should determine the influence of estrogen on neuroinflammatory changes associated with colitis. Elevated iNOS activity in the hippocampus region and associated behavioural despair in this study [
137], may suggest elevated nitric oxide (NO) production. NO is thought to play a central role in the neurobiology of depression [
149]. In line with this, iNOS-inhibitors reduce behavioural despair of mice with colitis [
137]. NO-associated depression may be due to impaired neurotransmitter synthesis and/or neurodegeneration [
150,
151]. However, currently no studies in animal models of colitis have elucidated this mechanism. Indeed, in rats inhibition of NOS elevates levels of extracellular serotonin and dopamine in the ventral hippocampus, a major brain region correlated with depression [
150]. During inflammatory pro-oxidant states, excessive NO in the brain can combine with superoxide anions to create peroxynitrate which induces neural degeneration and cell death via protein nitration [
151] (Fig.
2). Given that there is evidence of ROS production and oxidative damage in animals with colitis [
136,
137], further studies should investigate CNS changes indicative of oxidative damage in animals with colitis.
In a recently published study, DSS-induced colitis found upregulation of inflammatory-related proteins S100A8, S100A9 (also known as myeloid related protein (MRP) 8 and MRP14), and lipocalin-2 (Lcn2, also known as neutrophil gelatinase-associated lipocalin) in the brain. Though S100A9 and Lcn2 are upregulated in colitis, this is the first study to observe evidence of these proteins in the brain of mice with colitis [
138]. Neutrophils, macrophages, and monocytes are the main source of these proteins, although other cells can also release them during infection [
152]. The S100A8 and S100A9 proteins form a heterodimeric complex S100A8/A9 (also known as calprotectin), which has antimicrobial activity by sequestering trace metals essential for bacterial growth [
152,
153]. Following inflammatory stimuli, these proteins are significantly upregulated and released into the extracellular environment, where they can activate immune and endothelial cells [
152,
154]. Up regulation of S100A8 and S100A9 has been observed in many inflammatory diseases [
155,
156,
157,
158], and faecal calprotectin is used as a marker for IBD severity as its level significantly correlates with intestinal inflammation [
153,
159]. Paquinimod, an orally active immunomodulatory quinoline-3-carboxamide derivative, which blocks the interaction of S100A9 with TLR4 and receptors for advanced glycation end products (RAGE), prevented upregulation of Lcn2 and S100A8/A9 in the brain and S100A8/A9 in the colon and ameliorated symptoms of colitis [
138,
153,
160,
161]. The upregulation of S100A8 and S100A9 in the brain could be related to the infiltration of peripheral inflammatory cells into the brain (monocytes and neutrophils) observed in this study [
138]. However, S100A8 and S100A9 upregulation may be stressed-related as mice with chronic stress have upregulated genes encoding these proteins in the hippocampus [
162].
Whilst S100A8 and S100A9 may play a central role in propagating neuroinflammation in colitis models, it has been suggested that NLR family pyrin domain containing 3 (NLRP3) protein activation may be involved [
139]. NLRP3 largely functions as an intracellular sensor that detects microbial motifs, endogenous danger signals and environmental irritants [
163]. Activation of NLRP3 results in the assembly and activation of the NLRP3 inflammasome, which leads to cleaving of inactive pro-IL-1β and pro-IL-18 into their active forms via caspase 1 [
163,
164].
Increased NLRP3 inflammasome activity, microglial and astrocyte activation in the hippocampus and cortex, accumulation of gut-derived T cells along meningeal lymphatic vessels observed in the brains of wild type mice with DSS-induced colitis, were not found in NLRP3 knockout mice [
139]. These findings could be the results of a NLRP3 facilitating dysfunction of the “glymphatic” system. The glymphatic system facilitates the entry of subarachnoid cerebrospinal fluid (CSF) into the brain interstitium where it mixes with brain interstitial fluid (ISF) [
165,
166]. CSF-ISF then flows through the interstitium, being drained via para venous pathways to the meningeal lymphatic vessels, reaching the cervical lymphatics [
165,
166]. Astrocytes allow the movement of fluid between paravascular spaces and the interstitium via water channels such as aquaporin-4 (AQP4), which requires the polarization of AQP4 [
167]. This movement of fluid through the brain allows the removal of extracellular proteins, such as amyloid-β peptide (Aβ) and tau, from deeper areas of the brain, where interstitial solutes cannot normally reach the BBB [
165,
168]. Impaired glymphatic drainage may lead to Aβ and tau protein accumulation, which has been associated with triggering and propagating neuroinflammation playing a central role in neurodegenerative conditions such as Alzheimer’s disease [
169,
170,
171]. Glymphatic dysfunction, leading to impaired clearance of Aβ and aggravated cognitive decline seen in mice with DSS-induced colitis were attenuated in NLRP knockout mice [
155]. This may be due to the binding of IL-1β to cognate receptors on astrocytes leading to the loss of AQP4 polarity [
152]. Moreover, this study [
139] and others [
129,
138,
139] suggest a role of immune cells migrating from the gut to the brain in colitis-induced neuroinflammation. Meninges localised T cells have been shown to infiltrate the CSF, induce microglial activation, and enhance local pro-inflammatory cytokine production [
139,
172]. Additionally, other peripheral immune cells have been shown to be elevated in brain samples from animals with colitis including monocytes and neutrophils [
129,
138]. Whilst the evidence of neuroinflammation in colitis models is apparent, the underlying mechanisms still require exploration.
Microglial cells during neuroinflammation in colitis
After entering the CNS, inflammatory mediators may modulate local neuroglial cells in specific brain regions triggering neuroinflammation in animals with colitis [
126,
136]. Among the neuroglial cell populations, microglial cells can migrate and become activated during cytokine-induced neuroinflammation [
173]. Microglia are derived from the embryonic mesoderm and are closely related to peripheral macrophages [
174]. Functionally, they eliminate cell debris, remove damaged cells and destroy pathogenic agents [
175]. Moreover, they support and regulate neurogenesis, maintain oligodendrocyte progenitor cells, neuronal morphology, neural circuitry pathways, and neuronal outgrowth and positioning in the developing brain. During neuroinflammation, the inflammatory milieu activates microglial cells initiating their immunological response [
176,
177] (Fig.
2).
Several studies provide evidence that microglia are activated in the brains of mice with colitis [
126,
135,
138,
139,
140,
141,
142,
143]. In DSS-treated mice, significantly higher cortical and CA1 hippocampal immunofluorescence for a microglial marker, ionized calcium-binding adaptor protein-1 (Iba-1), has been observed [
126,
141]. Since Iba-1 is a marker of both resting and activated microglia, an increase in Iba-1 immunoreactivity in DSS-treated mice was attributed to a change in microglia morphology and localisation as opposed to increased cell number and consistent with increased microglia reactivity [
138]. Increased Iba-1 immunoreactivity may be transient as mice with DSS-induced colitis revealed increased hippocampal Iba-1 expression in acute colitis (day 7 post initial DSS treatment), and showed no difference after chronic colitis (day 29 post initial DSS treatment) [
135]. Moreover, DSS administered to weaning (postnatal day 21) mice revealed enhanced gene expression for markers associated with microglia such as
Iba-1, Nos2, and
IL-1β along with increased microglia cell numbers, decreased numbers of dendritic processes, and decreased length of processes [
140]. However, DSS was administered at the weaning stage, which is a critical point for the maturation of gut microbiota and may be due to gut dysbiosis [
140,
178]. Furthermore, rats with TNBS-induced colitis, displayed microglial activation, increased excitability of hippocampal neurons, altered hippocampal glutamatergic transmission, and lowered seizure threshold [
142,
143]. An intracerebroventricular injection of anti-TNF-α antibody and minocycline (an inhibitor of microglial/macrophage activation), reversed these findings, which may suggest CNS microglial/macrophage and/or TNF-α involvement in neuroinflammation associated with colitis [
142,
143]. In another recent study, the number of microglia was significantly increased in the cortex and hippocampus in DSS-fed WT mice but reduced in the NLRP3 knockout mice [
139].
Thus, the proinflammatory cytokines and NLRP3 appear to play a critical role in perturbation in microglia activity in these models. Whether these microglia are activated via mediators originating from the gut is yet to be confirmed and requires more investigation.
Microglia activation, migration and neurodegeneration
Increased expression of microglia, suggested in models of colitis, may be due to CNS-derived or circulating cytokines or antigens, which may activate neuroglial cells. As discussed, patients with IBD present with elevated serum levels of many pro-inflammatory cytokines and antigens inclusive of TNF-α, IL-1β, IL-6, and LPS [
114,
115,
116,
117]. If BBB dysfunction is indeed found in patients with IBD, these circulating factors may enter the brain parenchyma and alter microglial function conducive to the progression of neuroinflammation. Importantly we see evidence of upregulation of TNF-α, IL-1β, and IL-6 in various brain regions of colitis models [
126,
129,
136,
137], which may be sourced from or interact with neurons, microglia, and other cells.
Indeed, in vitro studies demonstrate the capability of cytokines to activate microglial cells and induce their release of pro-inflammatory and neurotoxic mediators [
176,
177,
179]. For instance, stimulation of microglia with recombinant TNF-α induces upregulation of many pro-inflammatory mediators such as TNF-α, Nos2, and Il-1β via an NF-kB p65 pathway [
176]. As mentioned above, excessive NO may trigger neuronal cytotoxicity through protein nitration [
180]. Whether NO neurotoxicity from microglia occurs in the CNS of colitis models is yet to be explored, but, as discussed earlier, TNBS-induced colitis was associated with a significant increase in hippocampal TNF-α and iNOS protein levels which could reflect reactive microglia activity [
137]. Furthermore, stimulation of mouse microglial cell line BV-2 with IL-1β induces expression of pro-inflammatory markers such as COX-2, chemoattractant protein-1 (CCL2), and IL-6 via a PI3K/Akt pathway [
177]. This pathway and associated inflammatory markers could be relevant in colitis as NLRP3 inflammasome activity (critical for caspase 1-dependent release of IL-1β and IL-18) in the CNS has been implicated in the exacerbation of neuroinflammation by DSS-induced colitis in aging mice [
139,
181]. Moreover, increased levels of mRNAs for TNF-α, IL-1β and COX-2 protein expression were found in isolated rat cerebral cortex microglial cell cultures treated with recombinant IL-6 compared to untreated control [
179]. COX-1 and -2 catalyse the formation of prostaglandins, thromboxane, and levuloglandins [
182]. In vivo, systemic TNF-α and LPS administration activated microglia and increased expression of brain pro-inflammatory factors in WT mice, but not in TNF R1/R2 deficient mice [
183]. Indeed, this may be consistent with studies that observed normalisation of synaptic transmission following either anti-TNF-α or minocycline treatment in animals with colitis [
142,
143,
184].
Enhanced prostaglandin activity might contribute to the mechanisms involved in the increased BBB permeability observed in models of colitis [
126,
127]. Different prostaglandin receptors appear to have varying functions in terms of BBB permeability. In ischemic stroke models, pharmacological or genetic inhibition of PGE2 receptors suggests that EP1 and EP3 receptors contribute to BBB breakdown observed in these models [
185,
186,
187]. EP4 was reported to attenuate BBB dysfunction induced by stroke [
188,
189]. However, in animals administered with LPS, prostaglandins show varying effects including blocking, enhancing, or having no effect on the actions of LPS on BBB permeability [
190,
191,
192]. Furthermore, systemic LPS challenge has been shown to induce upregulation of prostaglandin enzyme COX-1 in microglia and perivascular macrophages with PGE2 increase seen primarily in the hippocampus and thalamus [
193]. Mice with DSS-induced colitis exhibited more anxiety and less social behaviour than control mice and occurred in parallel with increased circulating IL-6, NPY, and IL-18 levels as well as an increase in hypothalamic
Cox-2 mRNA [
194]. In a recent study using DSS-induced colitis, elevated expression of the
Ptgs2 gene, which encodes COX-2, was noted [
138].
Systemic LPS challenge in mice elicits the increased amounts of CCL2 mRNA and protein in the hypothalamus and hippocampus, in conjunction with upregulation of chemokine receptor 2 (CCR2) expression by microglia [
195]. CCR2 studies in the CNS of mice with colitis are very limited and require further investigation. However, no changes in seizure threshold in colitis mice with impaired CCR2 functioning were found, which suggested that monocytes do not play a major role in colitis-induced neuronal hyperexcitability [
129]. CCR2 appears in two isoforms (CCR2A and CCR2B) with CCR2B being the dominant isoform making up 90% of all CCR2 expression and is observed on microglia, astrocytes, and neurons, while CCR2A is observed in certain mononuclear and smooth muscle cells [
196,
197,
198]. CCL2-CCR2 axis can induce the secretion of pro-inflammatory cytokines, such as IL-1β and IL-18 by microglia [
197]. Moreover, CCR2 appears critical for microglial accumulation as indicated in CCR2 knockout models [
199]. Studies should entice to investigate whether CCR2 is upregulated in brain tissue from animals with colitis, which may help elucidate possible mechanisms underpinning microglial activity seen in animals with colitis. Given that in vitro and in vivo studies evidence the capacity of circulating inflammatory mediators and endotoxins in inducing microglial changes, it may be plausible that alteration in microglia noted in animal models of colitis could be due to systemic infiltration of antigens and immune mediators. Moreover, COX2, PGE2, and CCR2 would be plausible future targets to investigate in the CNS of animals with colitis, given their role in the progression of events relevant to neuroinflammation.
Mechanisms of colitis-associated suppression of hippocampal neurogenesis
Impaired hippocampal neurogenesis has previously been associated with microglial cell activity leading to depression and maybe a neurobiological mechanism underlying IBD-associated depression [
200]. The association between reduced neurogenesis and depression in humans can only be inferred through reduced hippocampal volume noted in depressed individuals [
201]. However, post-mortem cellular changes in depressed humans revealed alterations in the neuropil, altered fluid content, and changes in granule cell and pyramidal cell density. [
202]. This may be responsible for hippocampal volume changes in humans. Further research to confirm neurogenesis as a neurobiological correlate of depression is needed.
Adult neurogenesis mainly occurs in the subgranular (SGZ) and subventricular zones in the dentate gyrus of the hippocampus resulting in the formation of new granule cells from neural progenitor cells [
203]. Microglia play a vital role in facilitating the complex process of neurogenesis. In vitro studies have demonstrated that microglial conditioned media enhance precursor cell differentiation, neuroblast production, and neuronal survival [
204,
205]. In addition, microglia are implicated in eliminating apoptotic neuroblasts and adult neurons through phagocytosis, which is vital given that most of the newborn cells undergo death by apoptosis within the first 1–4 days of their life [
206].
Given that increased microglial expression is noted in in vivo animal models of IBD [
126,
135,
138,
139,
140,
141,
142,
143], microglia-facilitated impairment of neurogenesis may be responsible for triggering or potentiation of colitis-associated depressive symptoms. Imaging studies show an increase in gray matter volume in the hippocampus of CD patients which may be related to immune activation that induces alterations in glial cells activity [
207]. There are limited studies confirming hippocampal dysfunction because of activated microglial cells in animal models of colitis. However, enhanced microglial cell activity in the hippocampus is correlated with a reduction in a neuronal marker, doublecortin (DCX), associated with reduced neurogenesis and behavioural abnormalities in mice with DSS-induced colitis [
140]. Moreover, it has been speculated that reduced neurogenesis seen in animals with colitis may be induced by cyclin-dependent kinase inhibitor p21
Cip1 (p21) activity in the hippocampus. Functionally, p21 restrains cell cycle progression and arrests the cell in the G1 phase [
208]. p21 can be induced in early neuronal progenitors and immature neurons in the SGZ and can function to limit these cells’ proliferation and ultimately suppresses neurogenesis [
135,
204,
209,
210]. In addition, acute systemic inflammation and pro-inflammatory cytokines, originating from microglia or other cells, can increase p21 expression and restrain hippocampal precursor cells of neuronal lineage in the SGZ [
210]. In a study using mice with DSS colitis, acute colitis was correlated with increased p21 expression in the hippocampus [
210]. However, in the chronic phase of inflammation, a fourfold increase in p21 mRNA levels was noted [
210]. Markers of neuronal stem/early progenitor cells, inclusive of nestin and brain lipid-binding protein, and DCX were downregulated [
210]. The nuclear protein Ki-67 and marker of cell proliferation co-labelled with DCX showed a decrease in number during chronic colitis in the SGZ [
210].
Microglial-associated cytokines IL-1β, TNF-α, and IL-6 have been shown in vitro to induce p21 expression in differentiating neuronal progenitors and may be partly responsible for the above findings [
210] (Fig.
2). Importantly, pro-inflammatory cytokines noted in the hippocampus of animals with colitis, whether secreted by microglia or other cells, or peripherally sourced, have been suggested to suppress neurogenesis through different mechanisms. For instance, Cacci et al. (2005) revealed that the co-culture of an embryonic hippocampus‐derived HiB5 cell line with LPS-activated microglia results in TNFα-mediated apoptosis suppressing neuronal development and differentiation [
211]. Moreover, altered hippocampal neurogenesis is seen in vivo in mice with depleted TNF receptor (TNFR)1 and TNFR2 [
212]. TNFR knockout mice showed an increased rate of neural progenitor proliferation and neurogenesis in the hippocampus [
212]. This study suggests that microglial activation may suppress hippocampal neurogenesis via the release of TNF-α binding to TNFR1 on hippocampal progenitors, which is known to be related to a fas-associated protein with death domain-caspase-8/3 which induces apoptosis likely contributing to impaired generation of new neurons [
213] (Fig.
2). Increased hippocampal expression of TNF-α has been noted in DNBS, TNBS, and DSS-induced colitis [
129,
135,
136,
137,
142,
143,
204]. Importantly, in DSS-colitis, cleaved caspase-8 was found upregulated in the brain and cleaved caspase-3 was found upregulated in the hippocampus, which may suggest the action of the above pathway [
135,
138].
Cytokine or antigen challenge can induce microglia to release IL-1β, which has been implicated in the modulation of neurogenesis. Studies have shown IL-1R1 expression in vitro in rat embryonic forebrain NPCs [
214] and adult rat hippocampal cells [
215]. The binding of IL-1β is associated with decreased proliferation in hippocampal progenitor cells [
216]. Furthermore, mice with chronic stress-induced depression display increased IL‐1β expression in the dorsal hippocampus that decreases dentate gyrus hippocampal neurogenesis [
217]. Moreover, IL-1β dysregulation dampens BDNF secretion associated with neurodegeneration [
218]. Reduced BDNF mRNA expression in the dentate gyrus and CA3 region of the hippocampus was seen in mice exposed to contextual fear conditioning followed by social isolation [
218]. In vivo treatment of contextual fear-conditioned mice with an IL-1R antagonist, suppressed IL-1β signalling improving BDNF expression and preventing impairments in hippocampally-dependent contextual fear conditioning tests following social isolation [
218] (Fig.
2). Limited studies in models of chemically-induced colitis provide evidence that BDNF expression is reduced in the hippocampus (DNBS) and forebrain (DSS) with IL-1β expression elevated in DSS models [
135,
136,
138].
Alteration in hippocampal neurogenesis in IBD animal models may be due to abnormal excitatory synaptic properties in the hippocampus. Hippocampal tissue from Sprague–Dawley rats with TNBS-induced colitis revealed enhanced Schaffer collateral-induced excitatory field potentials in CA1 stratum radiatum [
142]. Schaffer collaterals are axon collaterals from CA3 pyramidal cells projecting to CA1 area [
219]. This was associated with larger-amplitude miniature excitatory postsynaptic currents (mEPSCs), but unchanged mEPSC frequencies and paired-pulse ratios, suggesting altered postsynaptic effects. Both α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid receptor (AMPA)- and N-methyl-
d-aspartate (NMDA)-mediated synaptic currents were enhanced in the rats [
142]. Moreover, AMPA-mediated currents revealed increased contribution of GluR2-lacking receptors and mRNA and protein levels of the glutamate ionotropic receptor AMPA type subunit 2 (GluR2) subunit were reduced in the hippocampus [
142]. Interestingly, the chronic administration of minocycline, a microglial/macrophage activation inhibitor, lowered the level of TNF-α in the hippocampus and completely abolished the effect of peripheral inflammation on observed transient electrical signals and synaptic plasticity [
142]. The authors had previously shown in vivo that enhanced brain excitability during colitis requires both elevated cytokines TNF-α and microglial activation [
143]. Indeed, TNF-α has been evidenced to facilitate the insertion of GluR2-lacking AMPA receptors in the membrane [
204,
220,
221].
The increase in hippocampal NMDA and AMPA receptors may make neurons more prone to glutamate-induced excitotoxicity, though no evidence of increased release or availability of glutamate was found [
142]. The presence or absence of the GluR2 subunits determines the Ca
2+ permeability of the AMPA receptor [
222]. The low expression of GluR2 enables the formation of AMPA receptors with high Ca
2+ permeability, which contributes to neuronal degeneration [
222,
223,
224]. In relation to NMDA receptors, the 2A and 2B subtypes are widely distributed in the hippocampus. Moreover, quinolinate phosphoribosyltransferase, which converts NMDA agonist quinolinic acid (QA) into nicotinamide adenine dinucleotide, is low in the hippocampus reducing the capacity to clear QA in the hippocampus [
225]. QA may then function as an excitotoxin and damage the hippocampal neurons [
207].
It has been revealed that activation of adenosine monophosphate-activated protein kinase (AMPK) can enhance hippocampal neurogenesis through the AMPK/BDNF pathway [
226]. Furthermore, there is evidence indicating that activation of AMPK attenuates inflammation in the CNS [
227]. Neuroinflammation and suppression of hippocampal neurogenesis in models of colitis could be due to impairments in the AMPK/BDNF signalling pathway. A study tested this theory using an activator of AMPK, called liver hydrolysate (LH), that has been shown to increase hippocampal neurogenesis through the AMPK/BDNF pathway and has an antidepressant effect in an animal model of depression [
228]. In a study using DSS-treated mice, LH prevented depressive-like behaviours and enhanced hippocampal neurogenesis through the AMPK/BDNF pathway and hippocampal activation of microglia and astrocytes [
229].
HMGB1 expression could play a critical role in synaptic dysfunction and/or impaired neurogenesis in colitis models. HMGB1 is a 215 residue protein that consists of two consecutive L-shaped basic domains referred to as HMG boxes and a 30 amino-acid long tail with acidic properties [
230]. HMGB1 is commonly found in the nucleus where it binds to the minor group of B type DNA and distorts and bends the double helix DNA of 90 degrees or more. HMGB1 can function to modulate transcriptional activity through its interaction with transcription factors such as p53 and nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) [
230]. Moreover, HMGB1 can function as a damage-associated molecular pattern (endogenous danger molecule released from damaged or dying cells inducing immune response by interacting with pattern recognition receptors) and bind to hippocampal TLR-4 inducing the activation of NF-κB and Activator protein 1, which facilitate the synthesis of pro-inflammatory mediators such as IL-6, TNF-α and iNOS [
231]. As mentioned above, the induction of pro-inflammatory cytokines, including IL-1β, TNF-α, and IL-6, have been shown to inhibit hippocampal neurogenesis [
210,
211,
212,
216,
217]. Using an experimental model of chronic cerebral hypoperfusion in rats, anti-HMGB1 neutralizing Ab reduced hippocampal glial activity and inflammatory cytokines, such as TNF-α, IL-1β,
and IL-6, as well as increased antioxidants superoxide dismutase and catalase, which was associated with improved CA1 neuronal survival and cognitive tasks [
232].
In DNBS-induced colitis, increased expression of the
HMGB1 gene in the hippocampus has been suggested to be detrimental to hippocampal neurogenesis and function [
136]. Additionally, HMGB1 has also been shown to be involved in hippocampal neurobiological functions including memory and long-term potentiation [
233,
234]. HMGB1 can inhibit hippocampal long-term potentiation and memory via TLR-4 and RAGE, which is accompanied by activation of NF-κB and c-Jun N-terminal kinase [
233].
Overall, pro-inflammatory cytokines, HMGB1, glial cells and, synaptic dysfunction could independently or in unison be responsible for alterations in neurobiological pathways which promote suppression of neurogenesis seen in colitis [
140,
210]. However, the exact mechanisms whereby colitis can alter pathways and the relationship between contributing factors is yet to be determined.
Microglia and the serotonin–kynurenine pathway
Impairments in serotonin biosynthesis could be an underlying mechanism behind behavioural changes seen in animals with colitis. However, the serotonin biosynthesis theory of depression is still debated. In treatment for tuberculosis and schizophrenia, iproniazid (inhibits the breakdown of monoamines) and imipramine (blocks serotonin and norepinephrine transport) were found to reduce depressive symptoms [
235]. Moreover, reserpine, which can deplete monoamines, was implicated in triggering depressive symptoms. These observations helped formulate the theory that depression is caused by the depletion of monoamine transmission [
235,
236]. But whilst serotonin biosynthesis has been implicated in depression pathogenesis, many studies have found evidence contradicting this theory. For instance, selective serotonin reuptake inhibitors increase extracellular serotonin within short periods following administration [
237,
238], however, the beneficial antidepressant effects arise following weeks of continuous treatment [
239]. Moreover, reducing serotonin synthesis through dietary reductions in tryptophan fails to induce depression in non-depressed individuals [
240]. This review discusses the microglia-associated reduction of serotonin bioavailability as a possible mechanism underlying IBD-associated depression, however, caution should be taken as this proposed theory is still debated.
Microglial cell activity can modulate the serotonin-kynurenine pathway which plays an important role in depression [
241]. Microglia express the tryptophan-catabolizing enzyme IDO in the presence of pro-inflammatory cytokines [
242]. IDO converts tryptophan (amino precursor of serotonin [5-HT, 5-hydroxytryptamine]), into kynurenine (KYN) which can then be catabolized by the enzyme kynurenine 3‐monooxygenase (KMO) into excitotoxic metabolites 3‐hydroxy‐kynurenine (3‐HK), 3‐hydroxy‐anthralinic acid, QA, and finally the end‐point co‐enzyme nicotinamide adenine dinucleotide [
243]. Conversely, KYN can also be metabolized through a neuroprotective pathway to kynurenic acid (KYNA) by the enzyme kynurenine–aminotransferase (KAT) [
243]. Importantly, KMO is expressed by leukocytes such as monocytes, macrophages, and microglial cells, whereas KAT is present in astrocytes [
242,
244]. Quinolinic acid reduces the expression of astrocyte glutamate reuptake pumps while stimulating release of glutamate from astrocytes which may result in glutamate neurotoxicity and neurodegeneration [
245,
246] (Fig.
2).
Studies linking colitis with the TRY/KYN alteration in the CNS are limited. However, mice infected with a parasite
Trichuris muris (T. muris), which induces colitis, have higher levels of serum kynurenine and an increased kynurenine/tryptophan ratio when compared to non-infected mice [
247]. Moreover,
T. muris-infected mice displayed behavioural abnormalities as assessed by a light/dark preference test and elevated levels of circulating pro-inflammatory cytokines such as TNF-α and IFN-γ which were all alleviated with either a corticosteroid (budesonide) or an anti-TNF-α agent (etanercept) interventions that normalized circulating kynurenine levels [
247]. Similarly, mice with DSS and TNBS colitis revealed a reduction in serum levels of tryptophan and increased intestinal expression of IDO [
248,
249]. In humans, serum obtained from CD and control participants elucidated a marked reduction in tryptophan and an increased K/T ratio in active CD. Whether these findings are due to upregulation in microglial IDO is unknown. Indeed, increased IDO-1 gene expression was observed in the medial prefrontal cortex (PFC) of mice with colitis, however, this was accompanied by reduced microglia expression [
250]. It should be considered that alterations noted in serum tryptophan and K/T ratios could be due to changes in IDO-1 expression in the gut. IDO overexpressed has been noted in lesional biopsies from patients with IBD with CD123+ dendritic cells being the primary cell to express the enzyme [
251]. Moreover, although appearing detrimental in the context of depression, IDO expression appears beneficial in suppressing intestinal inflammation. In TNBS colitis, inhibition of IDO results in more severe colitis and a significantly increased colonic pro-inflammatory cytokine expression [
249]. This may be due to enhanced availability of tryptophan and increased 5-HT synthesis in the intestines. Indeed, 5- HT has been implicated in worsening colitis as mice with tryptophan hydroxylase-1 knockout experienced reduced 5-HT in the GI tract and had reduced severity of DSS-induced colitis [
252]. It appears that there may be paradoxical findings in the brain and gut whereby reduced serotonin worsens depressive symptoms and increased serotonin contributing to more severe colitis. Overall, more research is warranted to elucidate the presence of IDO-expressing neuroglia cells in the brains of animals with colitis and whether serum levels of tryptophan and K/T ratio alteration noted in IBD and animals with colitis contribute to/or are caused by IDO expression in the CNS.