ArticlePDF AvailableLiterature Review


Bipolar disorder (BD) is a long-recognized severe and common psychiatric disorder, with a complex and often diverse range of presentations. BD is a heterogenous disorder that has traditionally, if rather simply, been defined by the recurrences of manic and depressive episodes, and presents with numerous immune-inflammatory and circadian/sleep abnormalities. A number of different lines of research have investigated the biological underpinnings of BD and demonstrate an heritability about 80-90%. This genetic contribution is thought to be mediated by a wide array of genetic factors, rather than being strongly influenced by a couple of genes. In this context, a clearer formulation of the biological underpinnings of BD is needed in order to encompass the diverse effects of multiple susceptibility genes. The biological underpinnings of BD includes work that has focussed on the role played by increased immune inflammatory activity, particularly changes in pro-inflammatory cytokines, as measured both centrally and systemically. Changes in immune-inflammatory activity are intimately associated with alterations in levels of oxidative and nitrosative stress (O&NS), which are increased in BD. Many of the neuroregulatory changes driven by O&NS and immune-inflammatory activity are mediated by the tryptophan catabolite (TRYCAT) pathways, with changes in TRYCATs being evident both centrally and peripherally. A consequence of increased pro-inflammatory cytokines, is their induction of indoleamine 2,3-dioxygenase (IDO), which takes tryptophan away from serotonin, N-acetylserotonin and melatonin synthesis, driving it to the synthesis of neuroregulatory TRYCAT. Most work exploring such changes has emphasized the role of TRYCATs in enhancing or decreasing neuronal activity. However, a relatively overlooked consequence of cytokine induced IDO and TRYCAT pathway activation is the impact that this has on aryl hydrocarbon receptor (AhR) activation and in decreasing melatonergic pathway activity. Melatonin is classically associated with night-time synthesis by the pineal gland, in turn regulating circadian rhythms. However, melatonin is produced by many, if not all mitochondria containing cells, with consequences for gut regulation, as well as glia and immune cell reactivity. The melatonergic pathways are genetic susceptibility factors for BD. Interactive changes in O&NS, immune-inflammatory activity, TRYCATs and the melatonergic pathways form an emerging biological perspective on the etiology, course and management of BD. Here, we review such changes in BD, and how this better integrates the diverse array of BD presentations and comorbidities, including addiction and cardiovascular disorders as well as decreased life-expectancy. We then look at the future directions such research may take.
Send Ord ers for Reprints to
Current Pharmaceutical Design, 2016, 22, 987-1012 987
Bipolar Disorder: The Role of the Kynurenine and Melatonergic Pathways
George Anderson1*, Aude Jacob2, Frank Bellivier2,3 and Pierre Alexis Geoffroy2,3
1CRC Scotland & London, Eccleston Square, London, UK; 2Inserm, U1144, Paris, F-75006, France; Université
Paris Descartes, UMR-S 1144, Paris, F-75006, France; Université Paris Diderot, Sorbonne Paris Cité, UMR-S
1144, Paris, F-75013, France; 3AP-HP, GH Saint-Louis - Lariboisière - F. Widal, Pôle de Psychiatrie et de Mé-
decine Addictologique, 75475 Paris cedex 10, France; Fondation FondaMental, Créteil, 94000, France
Abstract: Bipolar disorder (BD) is a long-recognized severe and common psychiatric disorder, with a complex and
often diverse range of presentations. BD is a heterogenous disorder that has traditionally, if rather simply, been de-
fined by the recurrences of manic and depressive episodes, and presents with numerous immune-inflammatory and
circadian/sleep abnormalities. A number of different lines of research have investigated the biological underpin-
nings of BD and demonstrate a heritability of about 80-90%. This genetic contribution is thought to be mediated by
a wide array of genetic factors, rather than being strongly influenced by a couple of genes. In this context, a clearer
formulation of the biological underpinnings of BD is needed in order to encompass the diverse effects of multiple
susceptibility genes. The biological underpinnings of BD includes work that has focussed on the role played by increased immune in-
flammatory activity, particularly changes in pro-inflammatory cytokines, as measured both centrally and systemically. Changes in im-
mune-inflammatory activity are intimately associated with alterations in levels of oxidative and nitrosative stress (O&NS), which are in-
creased in BD. Many of the neuroregulatory changes driven by O&NS and immune-inflammatory activity are mediated by the tryptophan
catabolite (TRYCAT) pathways, with changes in TRYCATs being evident both centrally and peripherally. A consequence of increased
pro-inflammatory cytokines, is their induction of indoleamine 2,3-dioxygenase (IDO), which takes tryptophan away from serotonin, N-
acetylserotonin and melatonin synthesis, driving it to the synthesis of neuroregulatory TRYCATs. Most work exploring such changes has
emphasized the role of TRYCATs in enhancing or decreasing neuronal activity. However, a relatively overlooked consequence of cyto-
kine induced IDO and TRYCAT pathway activation is the impact that this has on aryl hydrocarbon receptor (AhR) activation and in de-
creasing melatonergic pathway activity. Melatonin is classically associated with night-time synthesis by the pineal gland, in turn regulat-
ing circadian rhythms. However, melatonin is produced by many, if not all mitochondria containing cells, with consequences for gut
regulation, as well as glia and immune cell reactivity. The melatonergic pathways are genetic susceptibility factors for BD. Interactive
changes in O&NS, immune-inflammatory activity, TRYCATs and the melatonergic pathways form an emerging biological perspective
on the etiology, course and management of BD. Here, we review such changes in BD, and how this better integrates the diverse array of
BD presentations and comorbidities, including addiction and cardiovascular disorders as well as decreased life-expectancy. We then look
at the future directions such research may take.
Keywords: Bipolar disorder, depression, mania, oxidative and nitrosative stress, inflammation, melatonin, indoleamine 2,3-dioxygenase,
tryptophan catabolites, aryl hydrocarbon receptor, kynurenine.
Bipolar disorder (BD) is a serious and common psychiatric
disorder. Traditionally, BD has been characterized by depressive
and (hypo)manic episodes, with a prevalence of between 2-5% [1].
The genetic contribution to BD is estimated about 80-90%, al-
though thought to be mediated by a wide array of genetic factors
[2], with a meta-analysis of genetic susceptibility studies indicating
that no individual genes are defining or event significant after mul-
tiple testing corrections [3]. Such data on the diverse, but individu-
ally insignificant genetic influence on BD, are likely indicators that
a clearer formulation of the biological underpinnings of this disor-
der is needed in order to encompass the diverse effects of multiple
susceptibility genes. A more integrated and encompassing biologi-
cal model would refine classification, at the very least complement-
ing the current use of patient and family reports as well as behav-
ioural observations. Recent works have proposed changes in oxida-
tive and nitrosative stress (O&NS), immune-inflammatory activity,
mitochondria, tryptophan catabolites (TRYCATs) and the melaton-
ergic pathways in the etiology, course and treatment of BD [4,5].
Pioneering research indicated that the two poles of mood epi-
sodes in BD (i.e mania and depression) show increased immune-
inflammatory pathway activity, as indicated by raised pro-
inflammatory cytokine levels, positive acute phase proteins, com-
plement factors and heightened T-cell-related activation marker
*Address correspondence to this author at the Head of Research, CRC Scot-
land & London, UK; E-mail:
levels [6-9]. Such data indicated that an acute manic episode is
accompanied by increased levels of the soluble interleukin-2 recep-
tor (sIL-2R), a T cell activation marker; sIL-6R; positive acute
phase proteins, including alpha-1 acid glycoprotein, fibrinogen,
haptoglobin and hemopexin; immunoglubulins (Igs), IgG1; and the
complement factors (C3 and C6) [4-6,8,9]. The depressive episode
of BD has a similar immune-inflammatory profile as that of major
depressive disorder (MDD), showing raised levels of sIL-2R, IL-6
and sIL-6R [7]. As such, both poles of BD can show heightened IL-
6 trans-signalling as well as T cell activation [4,10,11].
Other lines of study on the biological underpinnings of BD have
investigated alterations in TRYCATs [12], O&NS [13], mito-
chodrial dysfunction, driven by reactive oxygen species induced
inflammasome (e.g. a multiprotein oligomer) [14], hypothalamic
pituitary adrenal (HPA) axis [15] and circadian dysregulations [16]
as well as neurotrophins [17]. Prenatal infection may also raise
offspring BD risk [18,19], which further indicates the importance of
immune-inflammatory processes in the etiology of BD. It is also
recognized that BD cases experience pervasive sleep and circadian
abnormalities during acute, but also remission, phases (i.e. when no
or minimal other bipolar symptoms are present), which are associ-
ated with dysfunctions of the molecular circadian clock and mela-
tonergic pathways [20,21]. Current work also suggests that signifi-
cant changes occur over the course of BD, a process of neuropro-
gression [22], which is thought to contribute to the development of
autoimmune responses, in turn altering the nature of the biological
underpinnings of BD [23]. It is of note that no single biomarker is
/16 $58.00+.00 © 2016 Bentham Science Publishers
988 Current Pharmaceutical Design, 2016, Vol. 22, No. 8 Anderson et al.
determinant of pole or stage over the course of BD or, indeed, of
any accompanying neuroprogression; rather, if biomarkers are to be
found, it is likely that they will be formed by a composite of meas-
ures. It is clear that the association of such a wide array of systemic
processes with BD is highly likely to necessitate a reconceptualiza-
tion of BD as a multi-system disorder, rather than a centrally de-
termined affective or mood disorder. Consequently, the wide array
of BD comorbidities, including cardiovascular disorders (CVD)
[24], alcohol abuse/addiction [25], sleep disorders [26] and lowered
life-expectancy [27] can be seen as arising from the association of
BD with such systemic processes.
Here, we review how such a diverse array of central and sys-
temic processes that are altered in BD, may ultimately be linked to
changes in TRYCATs and the melatonergic pathways. Following a
brief overview of the kynurenine and melatonergic pathways, we
review immune-inflammatory activation in BD, before looking at
alterations in O&NS, mitochondrial dysfunction, TRYCATs, and
the melatonergic pathways. Each of these sections is accompagnied
by the treatment implications that may ensue, before finally we look
at some important future directions that such data indicates.
The kynurenine pathway is activated when tryptophan is driven
by indoleamine 2,3-dioxygenase (IDO) and tryptophan 2,3-
dioxygenase (TDO) to the synthesis of TRYCATs, such as the neu-
roregulatory kynurenic acid (KYNA) and quinolinic acid (see Fig.
1) [12]. IDO is activated by pro-inflammatory cytokines, especially
interferon-gamma (IFN-), whereas TDO is predominantly induced
by cortisol, allowing changes in stress and stress hormone levels to
regulate some key neuroregulatory processes, with pertinence to the
changes occuring in BD [4,12]. By taking tryptophan away from
serotonin synthesis, IDO and TDO decrease serotonin levels and
effects, including the need of serotonin as a precursor for the syn-
thesis of N-acetylserotonin (NAS) and therefore melatonin. As
such, the activation of the TRYCATs pathways, as well as produc-
ing neuroregulatory factors such as KYNA, is intimately associated
with alterations in levels of serotonin, NAS and melatonin.
Melatonin is a neurohormone that regulates circadian and sleep-
related responses [28], which is coupled to anti-inflammatory and
immune regulatory effects [29]. Melatonin is released by many, if
not all, mitochondria containing cells with proven antioxidant and
anti-inflammatory effects, although its night-time release by the
pineal gland has been most extensively investigated. Pineal mela-
tonin synthesis is also synchronized and regulated by the suprachi-
asmatic nuclei (SCN), regarded as the “master circadian pace-
maker” [30]. Melatonin secretion is high during the periods of dark-
ness and inhibited by the effects of light in the retina [31,32]. Con-
sequently, seasons of the year also regulate melatonin secretion due
to changes in the number of daylight hours. Melatonin is syn-
thesized from serotonin with the SCN modulating serotonin catabo-
lism into melatonin [33]. Melatonin synthesis is determined by the
activity of the serotonin N-acetyltransferase (AANAT) enzyme,
which leads to NAS, in a reaction requiring acetyl coenzyme A.
The formation of melatonin from NAS is driven by acetylserotonin
O-methyltransferase (ASMT) also named hydroxyindole O-
methyltransferase (HIOMT) (see Fig. 1) [34]. As is elaborated on
below, melatonin synthesis can also be immune regulated, and, as
well as the pineal gland, can be released by many organs and tis-
sues, including the gut, with relevance to a host of immune-
inflammatory processes and immune surveillance.
Increased immune inflammatory activity is present in BD, even
during remission phases [35]. This may be the result of reciprocated
interactions with O&NS. In an investigation looking at 27 euthymic
female BD patients that measured T helper (Th)-1, Th-2 and Th-17
cytokines from stimulated lymphocytes, it was shown that regula-
tory T cells were decreased, whilst Th1 (IL-2, IL-6) and Th17 (IL-
17) cytokines were increased, compared to age- and sex-matched
Fig. (1). Shows the major components of the kynurenine and serotoninergic pathways.
Omethyltransferase (ASMT)
Bipolar Disorder: The Role of the Kynurenine a nd Melatonergic Pathways Current Pharmaceutical Design, 2016, Vol. 22, No. 8 989
healthy controls. These changes were linked to increased mitogen
activated protein kinase (MAPK) activity, which is generally in-
dicative of enhanced lymphocyte activation [35]. Interestingly these
investigators also showed an increase in senescence associated cells
(CD8+ CD28-), suggesting that BD may be linked to early senes-
cence. Other data also suggests an increased senescence in patients
with long-standing BD, as indicated by serum levels of the
chemokine eotaxin/CCL11 [36], which the authors suggest indi-
cates a role for immune inflammatory processes in accelerated age-
ing, which may be part of neuroprogression over the course of BD
[36], as well as contributing to the array of factors that decrease
life-expectancy in BD [37]. Another indicator of altered immune
responsivity in BD is shown by data demonstrating an attenuated
immune suppressive effect of dexamethasone in BD T-cells [38],
suggesting differential responsiveness to the stress hormone, corti-
sol. In addition, increased monocyte [38] and macrophage [39]
activation are also present in BD, which will further enhance the
pro-inflammatory mileau [40].
Other studies have shown that wider immune cell dysregulation
occurs in BD, including decreased neuronal calcium sensor-1
(NCS-1) expression in CD4+T lymphocytes, CD19+ B lympho-
cytes and CD14+ monocytes [41]. Whilst in the manic phase pre-
liminary work indicates lower levels of natural killer (NK) cells
[42], being indicative of decreased protection afforded by these
cells in viral infection regulation, as well as the management of
cancers. Overall, BD is associated with many immune changes/
alterations that are relevant to BD per se, as well as to alterations in
ageing and in the appropriate immune response to other medical
A meta-analysis of BD cytokine alterations indicated raised
levels of soluble IL-2R, tumor necrosis factor alpha (TNF-), solu-
ble TNF receptor 1 (TNFR1) and soluble IL-6R in BD patients
versus controls [43], with no consistent significant alterations in IL-
1, IL-2, IL-5, IL-6, IL-8, IL-10, IL-12, IL-1, IL-1 receptor antago-
nist (IL-1RA), interferon-gamma (IFN-), transforming growth
factor-beta1 (TGF-1) and sTNFR2. Another meta-analysis of BD
cytokine alterations found IL-4, sIL-2R, sIL-6R and TNF- to be
significantly raised with IL-1RA showing raised levels only during
manic states [44]. This latter meta-analysis showed a trend increase
in IL-1 and IL-6, whilst there were no significant alterations in IL-
2, IL-8 or IFN-. Medication status was also found to have no im-
pact on IL-2, IL-4, sIL-6R, and INF- in this meta-analysis. Over-
all, the results of these two meta-analyses indicate that changes in
immune activation, as indicated by cytokine changes, are common
in BD, although with considerable differences between individual
studies [43, 44].
There is considerable heterogeneity in BD per se, including in
the levels of cytokines and wider immune responses, as these two
meta-analyses indicate [43, 44]. Many factors may contribute to
this, including: cytokines and wider changes that may be specific to
the depressive versus manic phases; first or early presentations
versus those with long-standing BD and recurrent episodes; the
severity of mood swing, including as to whether psychotic features
were evident; the presence of wider co-morbidities, such as addic-
tion or obesity, which are known modulators of immune responses
[45]; age of BD onset; differential effects of different medications,
used alone or in combination; and neuroprogression. Neuroprogres-
sion refers to the changes that occur over time during a number of
disorders, including BD. As such, neuroprogression is more evident
when recurrent episodes occur and correlates with increased apop-
tosis and neurotoxicity as well as decreased neuroplasticity and
neurogenesis. Many of the changes occurring over the course of
neuroprogression are mediated by O&NS driven oxidative damage,
leading to lipid peroxidation, protein oxidation, hypernitrosylation,
DNA damage and often to increased autoimmune responses
[13,22]. Treatment resistance is often thought to be at least partly
driven by such neuroprogressive changes [10,46]. Overall, hetero-
geneity across a number of variables can lead to sub-groups that
may be lost in such meta-analysis designs. Nevertheless, cytokine
changes are a robust finding in BD, suggesting the relevance of
immune inflammatory processes in the course and neuroprogression
of BD.
Other immune activation indicants are also evident in BD, in-
cluding: increased chemokines and the chemoattractant cytokines,
which regulate the attraction of circulating leukocytes to sites of
inflammation [47]; indicants of endothelial cell activity, including
osteoprotegerin and von Willebrand factor [47]; increased adipose
tissue derived adipokines, such as adiponectin, resistin and leptin
[48], with relevance to mood stabilizer- and antipsychotic-induced
obesity as well as to lithium's lowering of adiponectin levels [49];
acute-phase proteins, such as C-reactive protein (CRP) [50], which
associates with deficits in cognition in BD patients [51] and which
can also increase blood-brain barrier (BBB) permeability [52],
thereby contributing to CNS inflammation. Recent work suggests
that alterations in BBB permeability may be important in BD, pro-
viding a ready link between systemic immune-inflammatory activ-
ity and central changes [53].
The BBB is a crucial barrier between the brain parenchyma and
the systemic circulation, with the aim of maintaining brain homeo-
stasis and primarily composed of brain endothelial cells covered by
a basal membrane, astrocytes end-feet and pericytes, in close prox-
imity to neuronal endings. At this interface, astrocytes may act as a
hub for collating the effects of systemic alterations in immune and
cytokine activity in BD [54]. In line with this, alterations in frontal
cortex astrocytes have been shown to occur in BD autopsy studies
[55], suggesting that alterations in glia may be a relatively over-
looked aspect of the central changes occurring in BD, including as
to how such astrocyte changes may link alterations in neuronal
regulation by astrocytes with systemic processes.
BD Clinical Correlates and Immune Inflammatory Processes
General, global functioning (as well as psychotic symptoms)
correlate with IL-1RA and sTNF-R1 levels in both BD and schizo-
phrenic patients [56]. As suggested for the role of pro-inflammatory
cytokines in neuroprogression, this could indicate that immune
inflammatory processes may also modulate symptom severity as
well as cognition and wider aspects of day-to-day functioning. In a
comparison of cytokine changes in different phases of BD, eleva-
tions in IL-6 occurred during manic phases, whilst raised levels of
IL-10 were evident when manic patients switch to euthymic states
[57]. This may be relevant to changes occurring centrally, as IL-10
can interact with the serotonin transporter in specific brain regions
in euthymic BD, which can then modulate inter-area connectivity
and brain patterning [58]. Other cytokine correlates with BD clini-
cal presentations are also evident, with raised IFN- levels during
the course of mania associating with symptom severity as measured
by the Young Mania Rating Scale (YMRS). IFN- is also increased
during the depressive phase, as was IL-1. Given that IL-1, IL-6,
IL-18 and TNF-, but especially IFN- induce IDO [59], such data
suggests that TRYCATs are likely to be significantly altered in both
poles of BD, perhaps differentially. Preclinical data suggests that it
is the pro-inflammatory cytokine induction of IDO that mediates
the shift from chronic stress to depression [60]. However, the cyto-
kine and cytokine receptor changes occurring in BD are not neces-
sarily co-ordinated in a simple manner with clinical presentations,
as raised sIL-6R and sTNF-R1 levels are found in BD irrespective
of subsyndromal symptom levels during the euthymic phase [61].
By comparing cytokines across patients in depressive, manic and
euthymic states in comparison to healthy controls, Becking and
colleagues showed that only IL-6 levels were raised in the de-
pressed state, whilst IL-6, IL-2 and IL-4 levels were higher during
manic states [62]. In this study, symptoms of mood dysregulaton
positively correlated with IL-2 and IL-6 levels. In euthymic patients
versus controls, only IL-4 levels were increased. Other investiga-
990 Current Pharmaceutical Design, 2016, Vol. 22, No. 8 Anderson et al.
tions have compared IL-1, IL-2, IL-4, IL-6 and TNF- levels in
BD depression versus mania, finding that IL-6 levels were in-
creased and IL-1 and IL-4 levels decreased in the depressed state,
with this pattern being reversed in mania [63]. Interestingly, altera-
tions in cytokine levels have been shown to link the depressed and
manic phases of BD, with raised levels of pro-inflammatory cytoki-
nes, especially IL-6, during the depressive pole, increasing the like-
lihood of a subsequent manic phase [64]. Such data would suggest a
role for immune-inflammatory activity in the shift between BD
poles. Such work has also highlighted the association of pro-
inflammatory cytokines with variations across BD states [64], al-
though with raised levels of sIL-2R, sIL-6R, CRP, sTNF-R1, p-
selectin receptor and monocyte chemotactic protein-1 (MCP1) be-
ing evident in all BD phases of BD versus controls. In a study by
Bai and colleagues sTNF-R1 was shown to decrease along the BD
dimension, being highest in BD-I intermediate in BD-II and lowest
in the depressive phase [65]. Such data again highlights the poten-
tial importance of alterations in immune inflammatory processes in
the different poles, and variable presentations, of BD.
Another factor that may influence the variations in levels of
cytokines and immune-inflammation is the level of stress at the
time of measurement. Using a social stressor test, Wieck and col-
leagues showed that stress differentially regulates cytokine changes
in BD patients versus healthy controls [66]. In this study, BD pa-
tients showed an attenuated rise in IL-2 and sTNF-R1 versus con-
trols [66]. This study also indicated that another cytokine, IL-33,
shows raised levels in all stages of BD, which is accompanied by a
decrease in the expression of its receptor, sST2, irrespective of
stress states [66]. As other data also indicates increased levels of
IL-33 in BD [67], it will be important to further investigate the
relevance of this cytokine to the different BD poles and clinical
Alterations in the stress response in BD may be driven by
changes in the levels of, and responsiveness to, the stress hormone,
cortisol [68]. Fries and colleagues showed that BD patients might
show a hypo-responsiveness to cortisol that is driven by epigenetic
factors [68]. As cortisol may also modulate the immune response -
especially under stress- it will be important to determine how such
variations in cortisol and hypothalamic-pituitary adrenal (HPA) axis
regulation interact with the stress response to modulate the im-
mune-inflammation changes occurring in the different phases of
BD presentations. When lipopolysaccharide (LPS) is used to stimu-
late monocytes from untreated BD patients, there is a relative rise in
the IL-6/IL-1 ratio, which lithium treatment prevented [69], sug-
gesting that some o f the efficacy of lithium may be via the pattern-
ing of immune-inflammation. As lithium does not modulate mono-
cytes IL-6/IL-1 ratio in vitro, it is likely to be acting indirectly on
monocyte function in vivo. Likewise lithium treatment normalizes
the increased IL-6 in mitogen stimulated leukocytes [70]. Anti-
inflammatory effects are usually evident following lithium treat-
ment, although not always [71]. As such, lithium is likely to be
having immune regulatory effects that are dependent on wider im-
mune regulatory processes acting on a given individual at a given
point in time. This would suggest that it does not have any universal
or standardized impact on immune inflammatory processes. In-
creased IL-6 in BD patients is often associated with a decrease in
brain derived neurotrophic factor (BDNF) [72], suggesting that
wider neurotrophin changes are occuring in the presence of altera-
tions in immune-inflammation. It requires investigation as to
whether lithium modulates BDNF synthesis and autocrine effects in
immune cells more widely, given BDNF effects in macrophages
and T cells [73,74].
Alterations in immune-inflammation over the course of BD
have also been shown in comparison to chronic schizophenic pa-
tients and controls [75]. In this study, Brambilla and colleagues
showed, in patients with chronic BD, an increase in the generally
pro-inflammatory M1 type of monocytes (indicated by increased
IL-6 and CCL3) coupled to a decrease in the generally anti-
inflammatory pro-phagocytic M2 phenotype (indicated by CCL1,
CCL2 and IL-10), with these monocyte phenotype changes being
associated with a decrease in regulatory T cell activity (indicated by
lowered CCL2 and TFG-B1 levels). These results led to the authors
proposing that chronic BD is typified by heightened M1 versus M2
monocyte activation and a rise in the Th1/2 ratio, with this being
coupled to a decrease in regulatory T cell activity [75]. Other data
also acts to support this [39]. As we will discuss below, such data
on an alteration in the M1/M2 monocyte phenotype ratio may be
dependent on the levels of local melatonin synthesis by these cells,
given that melatonin acts in an autocrine manner to induce a M2
monocyte/macrophage phenotype, which is driven by the activation
of the transcription factor NF-B [76]. Indeed, recent work suggests
the existence of an immune-pineal axis [77], whereby the inhibition
of local melatonin synthesis by immune cells at sites of inflamma-
tion, leads to an increase in TNF-, which can switch off pineal
melatonin synthesis [78], leading to alterations in the circadian
rhythm, that are well-known to be associated with BD [79]. As
such, the immune-pineal axis may then act as a switch that deter-
mines the site of melatonin synthesis. The synthesis and release of
melatonin in M2 macrophages will reduce the production of pro-
inflammatory mediators, such as TNF, with a failure to induce a M2
phenotype leading to the maintenance of TNF levels, thereby nega-
tively regulating pineal melatonin synthesis. Upon the resolution of
local or systemic inflammation, local melatonin synthesis in im-
mune cells decreases TNF- and other pro-inflammatory cytokines,
leading to the reinstatement of pineal melatonin synthesis and the
normalization of the circadian rhythm. In this context it is of note
that increased NF-B activity is evident in BD immune cells, which
correlates with levels of plasma TNF- [80]. As such, genetic al-
terations in the melatonergic pathways in BD [3,4], may be interact-
ing with wider systemic processes, including as to how such sys-
temic processes interact with central processes, including circadian
[3,4]. This implies that a decrease in, or blockage/inhibition of, NF-
B induced melatonin will be a contributory factor to heightening
pro-inflammatory activity in BD, with relevance to processes of
chronicity and neuroprogression. As well as lithium, valproate is a
commonly used mood stabilizer, with efficacy in BD partly medi-
ated by immune inflammation regulation [81]. Valproate attenuates
the rises in IL-1, IL-2, IL-4, IL-6, IL-17, and TNF- in stimulated
human whole blood [82]. Although requiring investigation, it is not
unlikely that these effects are at least partly mediated by valproate's
regulation of melatonin and its receptors [83].
CNS Immune Inflammatory Activity in BD
In comparison to systemic immune activation, relatively few
studies have investigated cerebrospinal fluid (CSF) cytokine altera-
tions in BD patients. However, CSF neuroinflammatory markers,
including those of microglia activation, have been found in BD in
association with cognitive deficits [84,85], suggesting that such
central measures have relevance to clinical presentations and neu-
roprogression in BD. In euthymic BD patients, Jakobsson and col-
leagues showed that both neuroinflammatory, including MCP1, as
well as systemic inflammatory processes, are involved in BD
pathophysiology, with these authors also importantly showing that
markers of central immunological processes may be, at least partly,
independent of systemic immune-inflammatory activity [86]. In a
separate study these authors also measured the concentrations of
CSF neurofilament light chain (NF-L), myelin basic protein (MBP),
S100B, and heart-type fatty acid binding protein (H-FABP) in rela-
tion to diagnosis, clinical characteristics and current medications
[87]. They found in BD patients an elevation in NF-L, a marker of
subcortical axonal damage, positively correlated with specific an-
tipsychotics, whereas H-FABP showed a positive correlation with
lithium dosage. Such work indicates that axonal damage may be a
neuropathological component of BD, with alterations in central
inflammatory processes significantly interacting with specific
Bipolar Disorder: The Role of the Kynurenine a nd Melatonergic Pathways Current Pharmaceutical Design, 2016, Vol. 22, No. 8 991
medication used [87]. Effects of medications may be mediated via
alterations in blood-CSF barrier dysfunction [88]. In an animal
model of mania, lithium modulates both peripheral and cerebral
cytokine synthesis, which requires replication in BD patients [89].
Euthymic BD patients show raised levels of IL-1 and lowered
IL-6 versus controls, with the heightened IL-1 associating with the
numbers of hypomanic/manic episodes in the previous year [90].
Although undoubtedly complicated by the array of different phar-
maceutical treatments used, this work does indicate significant CNS
cytokine changes in BD patients. IL-1 is usually released with
another cytokine, IL-18, via the co-ordinated cleavage of their pro-
forms following inflammasome activation. IL-18 is also known as
IFN-inducing factor and has significantly increased central mRNA
in unipolar depression. Along with other pro-inflammatory cytoki-
nes, IL-18, especially via the induction of IFN, increases IDO,
therefore driving tryptophan away from serotonin and melatonin
synthesis and leading to the synthesis of neuroregulatory TRY-
CATs. IL-18 also potentiates IL-12 induced Th1 responses, sug-
gesting a role for ROS regulated inflammasome activation in the
modulation of wider aspects of immune-inflammation in BD [91].
However, there is relatively little work looking at IL-18 and the
inflammasome in BD.
A recent study by Kim and colleagues suggests that the nod-like
receptor pyrin domain-containing 3 (NLRP3) inflammasome may
have a powerful role in co-ordinating the increased levels of im-
mune-inflammation in BD with mitochondrial dysregulation [14].
Mitochondrial dysfunction is also commonly found in BD [92].
These authors suggest that altered levels of inflammatory cytokines
and lower mitochondrial complex I subunits, found both centrally
and peripherally, will contribute to increased mitochondrial ROS
production. Upon sensing the production and release of mitochon-
drial ROS, NLRP3 assembles the NLRP3 inflammasome, releasing
caspase 1, which initiates the inflammatory cascade via the release
of IL-1 and IL-18, thereby suggesting that the NLRP3 inflamma-
some forms a link between complex I dysfunction and immune-
inflammatory activity in BD. As such, wider aspects of well-proven
alterations in BD, namely mitochondrial dysfunction and ROS pro-
duction, may be intimately involved with immune-inflammatory
activity via NLRP3 inflammasome activation. As indicated above
the presence of released IL-1 and IL-18, as well as IL-18 induced
IFN-, will act to increase IDO and neuroregulatory TRYCATs,
whilst also decreasing tryptophan availability for the melatonergic
pathways [14,59,92].
Given the raised levels of O&NS in BD and the role of ROS in
inflammasome regulation, such a potentially integrated set of fac-
tors will be important to investigate in BD patients. Also the lower
levels of central IL-6 in euthymic BD requires investigation [90], as
a lowering of central IL-6 means loss of its predominantly anti-
inflammatory and generally homeostatic effects via IL-6R activa-
tion. MDD may be associated with decreased central IL-6, coupled
to increased serum IL-6 and sIL-6R trans-signaling [10,11]. It is the
trans-signaling activation of sIL-6R that drives the pro-
inflammatory effects of IL-6 and its association with many inflam-
matory conditions, including MDD [10,11]. The balance of the
homeostatic IL-6 effects with the detrimental trans-signaling via the
sIL-6R requires investigation in the context of O&NS and NLRP3
inflammasome changes in BD. Given that IL-6 can increase both
IDO and autoimmune associated Th17 cells, it is likely that central
and peripheral IL-6 is important to integrate into models of the
biological underpinnings of BD.
As indicated above, melatonin has autocrine and paracrine ef-
fects that decrease immune cell pro-inflammatory activity, includ-
ing in CNS glia. Work by Niles and colleagues [83,93] show that
valproate, a commonly used mood stabilizer in the management of
BD, increases melatonin receptors in CNS cells. As melatonin nor-
mally increases its own receptors, coupled to evidence showing that
local melatonin is produced in the CNS [29,94], it is of need of
investigation as to whether valproate mediates some of its benefi-
cial effects via the upregulation of melatonin production in CNS
cells, thereby enhancing its autocrine and paracrine effects that
lower immune inflammatory activity. As such, variations in local
melatonin synthesis and regulation, as well as that of its immediate
precursor, NAS, may be an important hub in glia for the many fac-
tors that act to modulate the course and treatment of BD. This will
be further elaborated on below.
Immune Inflammation and Cognition in BD
Decrements in cognition are defining features of neurodegen-
erative disorders. Recent work has investigated as to whether the
biological factors associated with Alzheimer's disease, including
hyper-phosphorylated tau and amyloid  (A), have any role to play
in the cognitive deficits evident over the course of BD. Indeed,
Rolstad and colleagues have investigated this, including looking at
changes in CSF concentrations of total and phosphorylated tau,
A1-42, ratios of A42/40 and A42/38, soluble amyloid precursor
protein and , and NF-L chain protein in euthymic BD patients
versus controls, including as to whether these CSF measures had
any relation to neuropsychological performance in five cognitive
domains. The results indicate that the Alzheimer's disease associ-
ated factors explained a significant proportion of the variance in
cognition in euthymic BD patients, across all cognitive domains,
with effects being independently of age, medication, disease status,
and BD subtype I or II. However, hyperphosphorylated-tau and
A1-42, did not individually contribute significantly [85]. Such
results may be parsimonious with the role that O&NS, immune-
inflammation, mitochondrial dysfunction and TRYCATs have been
proposed to play in Alzheimer's disease, including in driving co-
ordinated changes in A and hyperphosphorylated tau [59].
Other data links specific immune inflammatory markers, in-
cluding CRP [51] to cognitive deficits in BD. Some of the CRP
effects may be mediated by increasing BBB permeability [52],
thereby more directly increasing th e influence of systemic inflam-
matory factors on central processing. Many of these same processes
may be evident in recurrent MDD, which also shows evidence of
cognitive deficits and neuroprogression [95]. It is also of note that
increased gut permeability may contribute to raised CRP levels,
being another means by which variations in gut permeability may
drive mood and neuroprogressive changes centrally [96]. Raised
levels of serum IL-1RA, in all BD phases, correlates with cognitive
deficits in BD, including after increased IL-6 and decreased BDNF
have been controlled for [97]. Another cytokine, TNF-, and its
receptor levels may also modulate specific cognitive deficits in BD
[98]. It should be noted that such peripheral cytokine measures
assume that there is a direct or indirect effect of such cytokines on
central processes, including in interaction with factors that increase
BBB permeability. However, it is important to emphasize that not
all BD patients show cognitive deficits, indicating heterogenous
presentations. Consequently, sub-groups have been proposed, in-
cluding groups that are partially determined by sleep disturbance
and sub-syndromal depressive symptom levels [99]. Interestingly,
sleep disturbance and depressive symptomatology is also associated
with the early etiology of Alzheimer's disease [100], again implicat-
ing overlapping processes of cognitive decline in BD with those of
this classical neurodegenerative disorder. As with MDD, the fre-
quency and duration of BD episodes modulate levels of cognitive
deficits [101], linking to concepts of neuroprogression [102].
Other dietary-associated inflammatory factors may be linked to
decrements in cognition and neuroprogressive processes. Increased
homocysteine associates with Alzheimer's disease and is commonly
found to have raised levels in bipolar depression [103]. Homocys-
teine is commonly increased due to folate deficiency and is fre-
quently associated with a decrease in B-vitamins. Interestingly,
melatonin has been shown to decrease levels of blood homocysteine
in preclinical studies [104]. Folate depletion leads to increased ho-
992 Current Pharmaceutical Design, 2016, Vol. 22, No. 8 Anderson et al.
mocysteine, as the remethylation of homocysteine to methionine
requires folate. Wh en folate is depleted, the resultant increase in
homocysteine acts as a marker for decreased S-adenosylmethionine.
S-adenosylmethionine is a necessary methyl source for melatonin
formation from NAS [105]. As such, increased homocysteine, as
evident in BD, especially during depressive episodes, is likely to
occur in association with decreased melatonin and a relative in-
crease in the NAS/melatonin ratio. As discussed below, such altera-
tions in the NAS/melatonin ratio could be important to the under-
standing of shifts between poles in BD. Increased homocysteine is
also likely to contribute to CVD risk in BD.
Other Drivers of Immune-In flammation in BD
Obesity is recognised to significantly influence immune in-
flammatory processes, perhaps especially peripheral IL-6 levels.
This is of some importance in BD, where increased levels of obe-
sity are highly prevalent in individuals with BD, often as a result of
treatments and reduced physical activity. Recent work shows that
obesity negatively regulates outcome in BD [106]. As well as IL-6,
adipocytes are a significant source of other pro-inflammatory cyto-
kines and adipokines. As such, alterations in food regulatory proc-
esses that contribute to obesity are likely to interact with other im-
mune-inflammation regulators over the course and treatment of BD
[107]. Also high body mass index (BMI) is associated with a vari-
ety of sleep disturbances in BD compared to controls, showing an
exaggeration of phenomena observed in non-clinical populations
that might be due to these inflammatory processes and/or melaton-
ergic pathway alterations [108]. Levels of pro-inflammatory cyto-
kines and increased monocyte activation change from adolescence
to adulthood in BD offspring, irrespective of BD symptoms [109].
This study showed an increase in the cytokine pentraxin-3 and
chemokine ligand-2 (CCL2) as well as monocyte activation [109].
This could suggest that the genetic modulation of immune changes
in BD may contribute, over the course of development, to the wide
array of immune inflammatory activity levels in BD. As indicated
previously, stress levels at the time of testing/investigation are also
likely to be an unmeasured confound that modulates the diverse
immune inflammation responses in BD [65]. As to how levels of
stress interact with obesity in the etiology, course and treatment of
BD is unknown.
As well as obesity and the genetic and epigenetic alterations in
immune-inflammatory pathways, other factors may act to increase
immune-inflammation in BD patients, including infection, gut per-
meability and gut microbiota. Some of the central inflammatory
processes may be mediated by infection-driven immune activity.
Stich and colleagues found evidence for increased antibody produc-
tion in the CSF of BD patients, suggesting a role for the activation
of the intrathecal humoral immune system in a subgroup of BD
patients, with links to processes known to be integral to CNS auto-
immune disorders [110].
Alterations in how the gut microbiome and associated gut per-
meability contribute to immune-inflammation in a number of medi-
cal conditions is an area of intense investigation, including as to the
role of these factors in MDD [111]. It is generally accepted that
increased gut permeability leads to bacteria or partially digested
tiny fragment of food crossing over the gut barrier and thereby trig-
gering an immune reaction and associated heightened immune-
inflammatory activity [112]. Bacteria, by acting on toll-like recep-
tor (TLR)4, will also directly trigger innate immune activity, which
may then act to modulate wider immune responses, including those
of the adaptive immune response [113]. As such, many factors such
as infection and gut permeability are likely to associate with the
etiology, course and/or management of at lease some BD patients.
In this context, it is of note that gut melatonin levels, predominantly
from enterochromaffin cells, can be a hundred-fold higher than
night-time pineal gland melatonin synthesis, with melatonin acting
to decrease gut permeability [114].
Oxidative and Nitrosative Stress (O&NS)
Raised levels of O&NS are present in BD [13,115], as well as
in mood disorders more widely [116]. In a meta-analysis of serum
O&NS in BD, a number of indicants of increased O&NS are pre-
sent in all BD phases, including increased lipid peroxidation, nitric
oxide (NO) and DNA/RNA damage [117]. Increased myeloperoxi-
dase and decreases in the anti-oxidant catalase are also evident in
BD, suggesting wide changes in the regulation of O&NS in BD,
including in its association with immune-inflammation [118]. Ge-
netic alleles of the NO synthase (NOS)3 gene are associated with
increased levels of violent suicidal behaviours [119], linking NO
with the high levels of violent suicide found in BD [120].
O&NS increases are also evident centrally, with mitochondrial
dysfunction and the lipid peroxidation products, 4-hydroxy-2-
nonenal (4-HNE) and 8-isoprostane (8-Iso) also being increased in
the prefrontal cortex of BD patients at autopsy [121], which again
couples suboptimal mitochondrial functioning with increased levels
of O&NS. Central alterations also occur in BD white matter, with
white matter changes correlating with peripheral measures of lipid
peroxidation, which have also been mooted as BD peripheral bio-
markers [122].
Interestingly, MDD and BD depression are associated with
different aspects of paraoxonase (PON)1 functional status [123].
PON1 (EC is an antioxidant enzyme that is bound to high
density lipoprotein (HDL) and protects against oxidation of HDL
and low density lipoprotein thereby protecting against the develop-
ment of atherosclerosis. PON1 enzyme activity is in part deter-
mined by a PON1 Q192R polymorphism coding three genotypes,
i.e. QQ, QR and RR. While MDD is accompanied by lowered
PON1 activity, no such changes are detected in BD. Nevertheless,
an interaction between smoking and PON1 Q192R polymorphism
(i.e. the QQ genotype) significantly increased risk of BD but not
MDD [123]. Brain imaging studies suggest that preventing alcohol
and tobacco abuse in adolescence may prevent or slow the neuro-
progression in BD [124].
The relevance of increased O&NS in BD is supported by lith-
ium treatment decreasing O&NS levels in BD patients [125], con-
trols [126] and animal models of BD [127]. Similarly the mood
stabilizer, valproate, also decreases central O&NS as well as pro-
inflammatory cytokines in stressed rodent models [128]. As a con-
sequence of decreasing O&NS, mood stabilizers are also neuropro-
tective. Indeed, in a rodent model of ouabain-induced mania, oua-
bain decreased levels of the anti-apoptotic protein, bcl-2, as well as
increasing the O&NS driven apoptotic proteins, BAX and p53, in
rodent brains [129]. Both lithium and valproate improved th ese
ouabain-induced cellular apoptotic pathways, with the effects being
dependent on the protein investigated and the specific brain region
analyzed. These authors suggest that the sodium-potassium pump
(Na(+)/K(+)-ATPase) may be an important link between O&NS-
driven damage and the reduced neuronal and glial density that is
commonly observed in BD, with mood stabilizers exerting protec-
tive effects against ouabain-induced apoptotic pathways. As to how
relevant this is to the management of BD by mood stabilizers re-
quires investigation.
An important consequence arising from raised levels of O&NS
and lipid peroxidation in BD, and mood disorders more widely, is
the damage to DNA that may occur. DNA damage leads to the re-
cruitment of the DNA repair enzyme, poly(ADP-ribose) polym-
erase (PARP). PARP catabolizes nicotinamide adenine dinucleotide
(NAD+), in turn decreasing NAD+ dependent sirtuins. Maintaining
sirtuin levels is important to mitochondrial functioning and in pro-
longing life expectancy, which is decreased in BD. The relevance
of such DNA damage and the response processes induced, is indi-
cated by the positive correlation of oxidised DNA levels with the
number of BD manic episodes [130]. Such data suggests that there
is an association of O&NS with BD, especially mania, which may
Bipolar Disorder: The Role of the Kynurenine a nd Melatonergic Pathways Current Pharmaceutical Design, 2016, Vol. 22, No. 8 993
contribute to the course of BD, as well as to its neuroprogressive
nature. As such, increased O&NS in BD may be associated with a
number of other important changes in BD patients, including in the
regulation of mitochondrial and anti-ageing processes.
In a study looking at the seven sirtuin isoforms (SIRT1-7) in the
peripheral white blood cells of MDD and BD during depressive and
remissive states as well as in healthy controls, Abe and colleagues
found that SIRT1, 2 and 6 mRNA levels were decreased in the de-
pressive phase of both MDD and BD patients, with these mRNA
levels then returning to those of healthy controls during periods of
remission [131]. This data suggest that SIRT1, 2 and 6 mRNA ex-
pression alterations are mood state-dependent, with the study of
sirtuin protein changes and consequences requiring further investi-
gation. This could have important ramifications as sirtuin-1 and
sirtuin-3 are significant modulators of the circadian rhythm as well
as mitochondrial function, suggesting that an attenuation of sirtuin
levels and functioning may contributes to many aspects of BD,
including metabolic, longevity, neuroprogression and circadian.
Such sirtuin changes require investigation in BD CNS cells, includ-
ing over the course of recurrent episodes and the relevance of any
sirtuin changes to alterations in cognition and neuroprogression. As
well as lowering O&NS and optimizing mitochondria functioning,
melatonin can also increase sirtuins, suggesting that this could be
another route through which melatonin may have utility in BD, as
well as to the relevance of melatonin regulation by other factors,
both locally and pineal, in driving changes in BD [132]. It is of note
that O&NS can also lower BDNF levels, thereby reducing the neu-
rotrophic support in BD [133]. Given that the immediate melatonin
precursor, NAS, is a BDNF mimic, via the activation of the BDNF
receptor, TrkB, this could suggest that adequate levels of NAS may
offset the loss of BDNF [134], although variations in where and in
what cells NAS is produced is likely to mean that it is not a direct
substitute for O&NS driven reductions in BDNF in BD.
Increased O&NS in BD is likely to contribute to the decreased
longevity associated with this disorder by a number of means, in-
cluding O&NS driven increased risk of cardiovascular disease
(CVD) in BD patients [135], with Hatch and colleagues also finding
that increased O&NS is strongly associated with a proxy measure
of atherosclerosis. Adolescents with BD have lower indicants of
O&NS, versus BD adults, which gives support to the prevailing
idea of a set of 'staging' processes occurring over the course of BD
[18] and may explain, in part, the increased risk of CVD in BD,
although increased homocysteine in BD is also likely to contribute
to CVD and shortened longevity.
Mitochondrial Dysfunction and BD
Mitochondrial dysfunction is implicated in a wide array of
medical conditions, including neurodegenerative disorders such as
Alzheimer’s disease, Parkinson’s disease, multiple sclerosis,
Huntington’s disease and motor neuron disease, as well as other
medical conditions such as the autistic spectrum disorders, schizo-
phrenia and MDD [136-139], which are conditions where impaired
autophagy is also evident. Autophagy is a cellular mechanism in-
volved in the clearance of aggregated and misfolded proteins as
well as of organelles that are dysfunctional. Arguably, the associa-
tions of mitochondrial dysfunction and autophagy with such a wide
range of disorders raises the question as to its specific relevance to
BD. Alterations in TRYCATs and the melatonergic pathway have
similarly been associated with all of the above neurodegenerative
conditions, as well as with BD. It is also of note that melatonin is
associated with the regulation of autophagy across a host of cell
types and organisms [140,141]. Overall, this could suggest that
mitochondria, and the levels of mitochondrial and/or cellular mela-
tonin may be a common focal point for collating changes in the cell
to a wide array of biochemical processes.
It is possible that beneficial effects may be obtained via the
enhancement of autophagy. It is worth recalling that mitochondria
undergo cycles of fission and fusion, with fission leading to segre-
gation into two daughter organelles, and fusion leading to two or-
ganelles assembling into one. Fission enhances the sequestration of
the dysfunctional organelles due to their selective degradation by a
specialized form of autophagy, namely mitophagy [142]. As to
whether increasing autophagy enhances the clearance of dam-
aged/suboptimal mitochondria coupled to an enhancement of mito-
chondria biogenesis from more healthy, maintained organelles re-
quires investigation in BD, including following treatment with
mood stabilizers and antipsychotics.
Genetic polymorphisms in mitochondria DNA are associated
with BD, including via the altered regulation of central and periph-
eral cellular glucose [143]. Mitochondria are crucial for energy
production, which is driven by the activity of the electron transport
chain and mitochondrial oxidative phosphorylation. Mitochondria
can also have a significant role in driving apoptosis as well as in the
modulation of calcium signalling and synaptic plasticity. Mito-
chondria are not static organelles, with about 30% in the process of
moving at any given point in time. As such, modulators of mito-
chondrial transport are also likely to be regulate BD associated
processes. Disrupted-In-Schizophrenia 1 (DISC1) may have rele-
vance in this context. DISC1 is a candidate genetic risk factor for
BD [144], as well as for schizophrenia and severe recurrent MDD.
Ogawa and colleagues showed that DISC1 strongly associates with
trafficking-protein-Kinesin-binding-1 which interacts with the outer
mitochondrial membrane proteins Miro1/2 [145]. This complex of
proteins thereby links mitochondria to the kinesin motor and there-
fore to microtubule-mediated subcellular trafficking. In neuronal
axons DISC1 enhances anterograde mitochondrial transport. By
directly regulating mitochondrial trafficking, DISC1 is essential for
many neuronal processes. DISC1 is also associated with the regula-
tion of hippocampal neurogenesis in preclinical models [146], as is
melatonin [147]. BD autopsies indicate a decrease in electron trans-
port chain activity as well as a decrease in oxidative phosphoryla-
tion and increased O&NS in BD brains, all of which alterations in
DISC1 are likely to impact upon. Neuroimaging studies have con-
sistently shown lowered indicants of energy production as well as
changes in BD brain pH [92], suggestive of alterations in mito-
chondrial functioning.
Suboptimal mitochondrial function is evident in BD, with some
of the beneficial effects of lithium being via improved mitochon-
drial functioning [148], although any causal relationship with either
BD or lithium and altered mitochondrial functioning has still to be
fully established. Valproate can increase mitochondrial biogenesis
in non-CNS cells [149], indicating relevant mitochondrial effects
from the most widely used mood stabilizers. Mitochondrial DNA
variation is evident in euthymic BD patients, along with increased
oxidative stress [150]. In rodent BD models (an amphetamine in-
duced mania model), the behavioural changes following the addi-
tion of a rotenone dose that very mildly inhibits mitochondrial func-
tioning, led the authors to propose that mitochondrial dysfunction
predisposes to manic behaviour [151], with the implication that
drugs targeting suboptimal mitochondrial functioning will have
therapeutic potential in the treatment of BD, especially of manic
episodes. This is not to suggest that suboptimal mitochondrial func-
tion per se is sufficient to induce BD behaviours, but rather, that it
sensitizes the nature of the response (and associated gene induc-
tions) to particular triggers. This is a simple sub-cellular model of
BD. However, should such changes differentially occur in particu-
lar brain regions, it may contribute to the etiology and/or course of
BD via alterations in the inter-area patterning of brain activity.
Other aspects of mitochondrial functioning may be affected in
BD, including the functioning of the 18kDa translocator protein
(TSPO). TSPO regulates cholesterol transport into mitochondria,
which is necessary for steroid synthesis. The rs6971 polymorphism
in the TSPO gene leads to an amino acid substitution (Ala147Thr)
in the cholesterol-binding pocket which is positioned in the trans-
994 Current Pharmaceutical Design, 2016, Vol. 22, No. 8 Anderson et al.
membrane domain. This polymorphism has been shown to regulate
the steroidogenic pathway. Colasanti and colleagues reported that
there is a nominal association between the rs6971 TSPO polymor-
phism and a diagnosis of BD [152]. These authors then went on to
propose that the amino acid substitution will affect the regulation of
the HPA axis. As such, alterations in TSPO may impact on HPA
axis and cortisol changes in BD, therefore impacting on chronic
cortisol regulation of monoamine oxidase (MAO) [153]. Given that
an increase in MAO lowers the availability of serotonin for NAS
and melatonin synthesis, TSPO may then be acting on local mela-
tonergic regulation of mitochondria functioning. MAO is also mito-
chondria located [154].
Overall, alterations in mitochondrial functioning and transport
regulators have significant associations with BD. These mitochon-
drial changes are likely to have two-way interactions with levels of
TRYCAT and melatonergic pathway activations.
In recent decades, a decrease in serotonin has been widely ac-
cepted as a core feature in MDD, but are also recognised as a sig-
nificant contributor to BD. Alleles of the serotonin transporter have
been identified as a susceptibility factor for BD, as for MDD [155].
Serotonin transporter alleles are also associated with white matter
abnormalities in BD [156], with serotonin transporter alleles also
being a risk factor for antidepressant induced mania in BD patients
[157]. The serotonin transporter alleles have also been proposed to
regulate the association of early stress with adult suicidality in BD
patients [158]. As such, alterations in the regulation of serotonin,
perhaps especially the serotonin transporter, are relevant to the
etiology, course and management of BD. This is important, as acti-
vation of the TRYCAT pathways will decrease the availability of
serotonin and therefore NAS and melatonin.
As well as the serotonin transporter, levels of serotonin are
determined by the availability of tryptophan and its conversion to
serotonin by tryptophan hydroxylase. Alleles of the tryptophan
hydroxylase gene (TPH2) are a risk factor for BD [159]. However,
other physiological processes may more powerfully regulate sero-
tonin in BD. Under normal physiological conditions, over 95% of
tryptophan is driven down the kynurenine pathways leading to the
production of a variety of TRYCATs, including KYNA, picolinic
acid and quinolinic acid. These TRYCATs can have a differential
impact on neuronal and glia functioning at different CNS sites.
Therefore, there is usually only a relatively small percentage of
total tryptophan available for the production of serotonin, NAS and
melatonin, which is even lower under chronic stress and heightened
immune-inflammatory activity.
As such, any further activation of the TRYCAT pathways will
deplete serotonin, NAS and melatonin. TRYCAT pathway activa-
tion is primarily driven by two independently-reglulated enzymes,
namely IDO and TDO. Both of these enzymes act initially to cata-
lyze kynurenine synthesis from tryptophan.
IDO is very strongly induced by IFN-, with other pro-
inflammatory cytokines also inducing IDO to some degree, includ-
ing IL-1, IL-6, IL-18 and TNF- [10,160]. As indicated in the
immune-inflammation section above, all of these pro-inflammatory
cytokines and IDO inducers have their levels raised in BD, suggest-
ing that the differential regulation of IDO and the TRYCAT path-
ways occurs in BD, perhaps dependent on the particular pro-
inflammatory cytokine that is increased at a particular site or cell.
TDO is regulated by different factors, including by cortisol [161],
D-amino acid oxidase (DAAO), ROS and cyclic adenosine mono-
phosphate [162]. This is of importance as KYNA levels are in-
creased in the CSF of BD patients [163], which will have a negative
impact on cognition via the inhibition of the alpha7 nicotinic recep-
tor (7nAChR). The activation of the 7nAChR is a known cogni-
tive enhancer, with recently completed phase two trials in schizo-
phrenia patients showing that it improves cognition in this patient
group [164]. Given the raised levels of KYNA in the CSF and its
impact on 7nAChR driven cognition, it is not unlikely that the
activation of IDO and TDO will mediate some of their negative
effects on cognition via KYNA induction. Interestingly, circadian
melatonin is a significant positive regulator of the 7nAChR [165],
suggesting that any genetic, epigenetic or environmental suppres-
sion of melatonin synthesis will modulate the 7nAChR, thereby
having an impact on cognition.
The 7nAChR is a significant treatment target in many neu-
rodegenerative conditions, including Alzheimer's disease [166]. As
to whether local melatonin synthesis by CNS cells has any impact
on levels of the 7nAChR will be important to determine, particu-
larly given that single nucleotide polymorphisms in the 7nAChR
are associated with impaired attention in BD patients [167] and the
rs12916879 allele with BD-I susceptibility in males [168]. It is of
note that the 7nAChR is also a significant regulator of glia and
immune cell reactivity, suggesting that alterations in the TRYCAT
and melatonergic pathways, via 7nAChR regulation by KYNA
and melatonin respectively, will have an impact on the reactivity
threshold of these cells [169-171]. Some of the effects of melatonin
in mitochondria also seem to be mediated via the 7nAChR [172],
suggesting further the close association of these pathways with the
7nAChR. As the production of KYNA will inhibit the 7nAChR,
variations in TRYCAT pathway activation as a consequence of
O&NS and pro-inflammatory cytokines will significantly interact
not only with the levels of melatonin produced but also with its
effects, some of which seem via the 7nAChR. As indicated in the
section on 'mitochondria and BD', some of the beneficial effects of
melatonin may be via its regulation of autophagy. Recent work
shows that melatonin's upregulation of autophagy is mediated via
the upregulation and activity of the 7nAChR [173].
As such, the increased CSF KYNA, decreased melatonin and
genetic associations of the 7nAChR would suggest that the inter-
actions of the TRYCAT, serotoninergic and melatonergic pathways
are of some relevance in BD and relatively overlooked as a phar-
maceutical treatment target. Also given the dysregulation of the
HPA axis in BD and the role of cortisol in increasing TDO, it is of
note that increased central TDO is associated with decreased neuro-
genesis and increased anxiety in rodents [174].
It should be noted that a variety of cells and tissues express
IDO. IDO is important in determining the patterning of the immune
response, including through its activation in dendritic cells, where it
leads to an increased production of regulatory T cells. As such,
given the importance of immune-inflammation in BD, as high-
lighted above, the balance of regulatory T cells to the pro-
inflammatory T-helper (Th)1 and Th17 cells is of some importance.
Centrally, IDO seems mainly expressed in microglia as well as in
infiltrating systemic immune cells. As well as being differentially
regulated, the pattern of TDO expression is also different to that of
IDO, with TDO showing very high hepatic expression, but also
with significant expression in astrocytes and in some neurons.
With over 60% of brain kynurenine being peripherally derived,
and kynurenine being readily taken up over the BBB by the large
amino acid-transporter-1 [175] and thereafter converted in glia to
KYNA and other TRYCATs, the activity of the systemic immune
system and inflammation in other organs and tissues will have an
impact on CNS processing. As such, not all immune-inflammatory
activity is mediated via the effects of immune cell-derived cytoki-
nes in the CNS, but also by the utilization of tryptophan in the
synthesis of kynurenine and downstream TRYCATs.
Following the production of kynurenine, its catabolism leads to
the synthesis of KYNA by the activity of kynurenine 2,3-
aminotransferase (KAT). In astrocytes and in some neurons that
predominantly express TDO, kynurenine and kynurenic acid are the
main TRYCATs that are produced and released. In microglia and
other IDO-expressing cells, further TRYCATs products are pro-
Bipolar Disorder: The Role of the Kynurenine a nd Melatonergic Pathways Current Pharmaceutical Design, 2016, Vol. 22, No. 8 995
duced and released, including following the enzymatic activity of
kynurenine 3-monooxygenase (KMO), the product of which sig-
nificantly inhibits neurogenesis [176]. Other neuroregulatory TRY-
CATs include the excitotoxic quinolinic acid and 3-hydroxy-
Recent work looking at the blood-derived measures of these
TRYCATs found that the putative neuroprotective index, the
kynurenic acid/quinolinic acid ratio was significantly lower in BD
patients versus controls, and independent of medication or a prior
history of psychosis [177]. These authors also found that another
putative neuroprotective ratio index, the KYNA/3-hydroxy-
kyurenine ratio was positively correlated with the volume of the
hippocampus in BD patients after controlling for age, sex, body
mass index and intracranial volume. Kynurenine/3-hydroxy-
kynurenine significantly correlated with total amygdalar volume in
BD, although this association was reduced to that of a trend after
controlling for age, sex and BMI. This study indicates a number of
interesting possibilities, including that alterations in the regulation
of specific TRYCAT patterning in BD impacts centrally on hippo-
campal and amygdalar structure. This would suggest that systemic
and central immune-inflammatory associated cytokines, as well as
TDO regulation by cortisol and chronic stress, may have significant
effects on the structure and activity of key brain regions that are
classically associated with alterations in BD.
Other work on TRYCATs and their ratios in BD patients have
shown significant differences. TRYCAT synthesis in BD skin fi-
broblasts shows raised levels of neurotoxic 3-hydroxykynurenine as
well as increases in the sometimes more neuroprotective KYNA
[178], with the 3-hydroxykynurenine/KYNA being increased, per-
haps indicative of a bias to neurotoxic TRYCATs. The raised 3-
hydroxykynurenine/KYNA ratio was evident at baseline, as well as
following pro-inflammatory cytokines induction of IDO, indicating
that raised pro-inflammatory cytokine levels in BD may contribute
to alterations in TRYCAT patterning. This requires investigation in
CNS cells, especially microglia, which seem the major IDO-
expressing cells in the brain. Autopsies on BD patients also indicate
TRYCAT changes in the prefrontal cortex [179].
Work by Reininghaus and colleagues provides further support
for a significant impact of TRYCATs changes in BD [180]. These
authors showed an increase in blood kynurenine and in the
kynurenine/tryptophan ratio in euthymic BD patients [180]. These
increases were more apparent in overweight BD patients, implicat-
ing increased obesity in BD patients in the modulation of immune-
inflammatory activity driven TRYCATs changes. Given that much
of the increase in obesity in BD may be medication-driven, this
would suggest complex effects of mood stabilizers and antipsychot-
ics via the induction of obesity and obesity regulation of inflamma-
tory responses that may regulate TRYCATs. It is of note that mela-
tonin can significantly attenuate medication-driven weight gain in
BD [181,182]. Increased pain reporting is evident in BD, especially
migraine and chronic pain [183]. With peripheral [184], and proba-
bly central [185], TRYCATs contributing to this, along with raised
levels of immune-inflammatory cytokines and O&NS [186], it is
likely that alterations in nociception are another clinical aspect of
BD in which the melatonergic and TRYCAT pathways are inti-
mately involved.
The Aryl Hydrocarbon Receptor
Recent work has shown an important role for the aryl hydrocar-
bon receptor (AhR), classically referred to as the dioxin receptor, in
the regulation of a number of key processes known to be altered in
BD. The AhR is a ligand-activated member of the Per-Arnt-Sim
(PAS) family of basic helix-loop-helix (HLH) transcription factors.
The AhR is activated by a number of endogenous ligands, including
formylindolo[3,2-b]carbazole (FICZ) and kynurenine, as well as by
a wide variety of ubiquitous environmental contaminants such as
2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). The AhR complexes
with chaperone proteins in the cytoplasm, including heat shock
protein 90 (HSP90) and the AhR-interacting protein (AIP) [187].
Increased serum anti-HSP90 level is significantly raised in BD
patients during acute mania versus in remission [188], suggesting
that factors linked to AhR activation show alterations in different
poles of BD. When ligand-bound, the AhR complex undergoes
nuclear translocation, where it binds the AhR nuclear translocator
(ARNT). This AhR-ARNT heterodimer binds to specific motifs,
namely xenobiotic-responsive elements (XREs), in target gene
promoters. Such targets genes, include the cytochrome P450 family
(such as CYP1A1, CYP1A2, CYP1B1) thereby impacting on the
metabolism of melatonin and circadian regulation more widely
[189]. This also links the AhR to BD, where circadian rhythm dys-
regulation is common [190], as well as to psychosis more widely
As hinted at above, a number of pathways can regulate the
AhR, including via its proteasomal degradation and CYP1A1-
driven decreases in ligand availability, as well as the many factors
that can influence AhR-ARNT complex formation [192]. An impor-
tant aspect of AhR control is the induction of the AhR repressor
(AhRR), which is a polypeptide that competes with the AhR for
binding to ARNT, being induced by the AHR/ARNT heterodimer
via its binding to XRE in the AhRR promoter. As such, the AhR
negatively feeds back on itself via AhRR induction. There are no
direct studies on the role of the AhRR in BD, nor has any circadian
regulation of the AhRR been investigated. However, recent data
shows that there is a prenatal epigenetic influence on the AhRR that
is driven by increased levels of maternal obesity [193]. Burris and
colleagues showed that AhRR DNA methylation is increased by
over 2% in the offspring of obese compared to normal weight
mothers, with AhRR DNA methylation also being significantly
increased in pretermers comparted to full-term gestations [193].
Given the long suspected role of prenatal factors in BD [194,195],
as in many other psychiatric conditions, any changes in the AhR
and how it is regulated is likely to impact on the prenatal immune
influence on BD development [195]. Interestingly, given the activa-
tion of the AhR by TCDD, among other AhR activators in cigarette
smoke, maternal smoking may significantly increase offspring BD
risk [196].
A number of transcription factors can also interact with, and
regulate, AhR signaling, including signal transducer and activator
of transcription (STAT)-1,-3 and STAT5, as well as Pai-2, Sp1, c-
maf, and Bach2 as shown in a range of different cell types
[197,198]. Different STAT transcription factors are involved in
immune system regulation, including in determining dendritic cell
responses via the regulation of IDO [199]. STAT3 is also an impor-
tant regulator of the serotonin transporter, suggesting that this tran-
scription factor could have a role in the co-ordination of AhR regu-
lation with serotonin availability as a precursor for the melatonergic
pathways [200].
Central TRYCATs and AhR
As such, a number of ligands, including within the TRYCAT
pathway, activate the AhR, in turn regulating a number of key proc-
esses, including driving the patterning of the immune response
[201]. Kynurenine activation of the AhR can mediate immune cell
activation, including within central astroyctes [202] and microglia
[203], as well as murine central endothelial cells [202]. As such, the
AhR is highly expressed in humans centrally [204], with its activa-
tion effects in microglia involving the GSK-3/Akt/PI3K pathways
[203]. However, it is of note that AhR activation in microglia can
have pro-inflammatory and some anti-inflammatory effects [205].
The deletion of the AhR in rat cerebellar granule cell precursors
decreases neurogenesis [206], suggesting that the interaction of
specific TRYCATs with the AhR could have some impact on the
996 Current Pharmaceutical Design, 2016, Vol. 22, No. 8 Anderson et al.
cerebellar changes known to be evident in BD [207]. Calcium and
zinc homeostasis, both showing alterations in BD, may be also
dysregulated following toxin activation of the AhR [208], with
possible consequences for the differential regulation of the mela-
tonergic pathways, as indicated below. Given the increased
kynurenine pathway activity that arises from pro-inflammatory
cytokine induced IDO and cortisol induced TDO, it is likely that the
TRYCATs induced, particularly kynurenine, will be mediating
some effects, including in BD, via AhR activation of immune and
glia responses. Also some on the protection afforded by decreased
O&NS and increased KYNA in the brain when the AhR is KO'd
centrally, suggests that the presence of brain AhR is of importance
to wider O&NS and TRYCATs regulation [209]. There is currently
no data on how the AhR will interact with the important role of
high mobility group box 1 (HMGB1) in mediating the effects of
stress on microglia reactivity and HMGB1 release [210,211]. Ani-
mal studies suggest that this could be a two-way interaction with
the AhR being shown to regulate the levels of cortisol response
[212], as well as cortisol induced TDO contributing to kynurenine
activation of the AhR and to cortisol mediated priming of microglia
for HMGB1 release and NLRP3 inflammasome activation [213].
Such data on kynurenine activation of the AhR suggests that the
activity of kynurenine aminotransferase (KAT) will be of some
relevance to the differential regulation of CNS processes. By cata-
lyzing KYNA from kynurenine, KAT will act to increase the
KYNA inhibition of the 7nAChr, with consequences not only for
cognition, but also for the effects of melatonin, given that melatonin
effects may be intimately interwoven with those of the 7nAChr
[165]. As such, alterations in specific TRYCATs may regulate dif-
ferent central processes that are associated with alterations in BD,
including the melatonergic pathways, but also O&NS and immune
inflammatory responses. The interactions of these pathways in BD,
as well as other psychiatric and neurodegenerative conditions, will
be important to determine, including any differential expression in
different CNS areas, given that such changes are likely to signifi-
cantly impact on the patterning of brain activity, such as amygdala-
cortex interactions that are altered in schizophrenia and BD [54].
Overall, the AhR is highly expressed centrally, with significant
impacts on levels of glia activation, and interacts with factors
known to be altered in BD.
Systemic TRYCATs and AhR
However, although expressed at high levels in the brain, the
AhR has been more extensively investigated as an important regula-
tor of systemic immune responses [192], including in the regulation
of the Th1, Th17 and regulatory T cells [214], partly driven by its
regulation of dendritic cell responses [215]. The AhR is present in,
and is a significant modulator of, a wide spectrum of immune cells,
including T cells, B cells, macrophages and NK cells, as well as in
dendritic cells [215]. The AhR is also present in epithelial cells,
Langerhans cells and innate lymphoid cells, as well as in astrocytes
and microglia [192]. Recent work suggests that the alterations in the
immune system may be significant to the etiology, course and
treatment of BD, including as a consequences of increased levels of
autoimmunity [216-218]. As such, the presence of the AhR in such
an array of immune cells and the AhR interactions with biological
factors long associated with BD, including TRYCATs and circadian
regulation, suggests that O&NS and inflammation-driven activity
that may drive auto-immune responses is relevant in BD, as in
MDD [10]. The AhR is a significant regulator of Th17 cells and the
synthesis of IL-17, which is associated with a more prolonged in-
flammatory response that may underpin the development of auto-
immune aspects of BD [219].
However, in the context of a role for autoimmune activity in
BD, it should be noted that the consequences of AhR activation on
types of immune responses, including auto-immune, may be partly
determined by specific AhR ligands. One of the problems in inves-
tigating the AhR, is the variations in responses to different AhR
ligands. This considerably complicates an understanding of a role
for the AhR in many medical conditions, which may be exemplified
in autoimmune research by the effects of different AhR ligands in
experimental autoimmune encephalomyelitis (EAE), which is a
rodent multiple sclerosis model. TCDD and 2-(1'H-indole-3'-
carbonyl)-thiazole-4-carboxylic acid methyl ester (ITE) suppress
EAE, whereas FICZ exacerbates the development of EAE [220]. A
number of different factors may underlie this, including: variations
in ligand stability and/or structure; variations in ligand AhR affin-
ity; variable concurrent pathway activations from these ligands that
may act to modulate AhR pathways; alterations in concurrently
induced microRNA; and the differential activation of the DRE/XRE
AhR in Wider Aspects of BD
The work by Chinen and colleagues shows the AhR to interact
with specific microRNAs in the regulation of gut homeostasis and
IL-10 production [221]. This may have some relevance to the puta-
tive alterations in gut microbiota and gut permeability in BD, given
that AhR activation profoundly regulates the immunological status
of the gastrointestinal tract. In the gut, the AhR establishes and
maintains a set of signaling networks that facilitate host-microbe
homeostasis at the mucosal interface. These sets of interactions
arise, in part, as a consequence of bacteria utilizing indole and in-
dolyl compounds as intra- and inter-species signaling molecules
[223]. Indoles are involved in a whole host of bacterial communica-
tions, including biofilm formation, bacterial motility, plasmid sta-
bility, virulence and antibiotic resistance [224]. However, although
highly likely, it has yet to be shown as to whether gut bacteria pro-
duce melatonin, an indole derivative, and, if so, as to whether this is
relevant to the positive gut effects of a subset of bacteria. The gas-
trointestinal tract is an area of cutting edge investigation across a
host of medial conditions. It can contain over 1012 enteric bacteria
of various species that have the capacity to synthesize indoles
[223]. Bacteria-derived indoles can act as signaling communica-
tions to the host, in turn modulating epithelial cell gene expression
as well as contributing to the maintenance of the gut-barrier and
decreasing gut permeability as well as lowering indices of inflam-
mation [225]. Recent work shows that the indoles act on the AhR to
regulate many gut processes, including gut-barrier integrity [224],
supporting previous work showing the importance of the AhR to
gut homeostasis and the maintenance of the gut epithelial cell bar-
rier integrity [226].
AhR signaling is also crucial in the development of the mam-
malian heart, being central to the regulation of pathways crucial for
cellular metabolism, cardiogenesis and cardiac function [227],
which are potential targets of environmental factors associated with
CVD, with CVD being significantly increased in BD [228]. Envi-
ronmental toxin activation of the AhR suppresses gluconeogenesis
and glycogenolysis, whilst stimulating lipogenesis and triggering
inflammatory gene expression in an AhR-dependent manner [229].
These authors also showed that such AhR activation altered the
ratio of unsaturated/saturated fatty acids [229], which is often found
in BD [230] and known to contribute to CVD susceptibility [231].
As such, available data would suggest that the alterations in AhR
ligands that are found in BD are likely to impact on processes un-
derlying increased CVD risk. Prior to the development of mood
stabilizers, the natural course of BD often resulted in death as a
result of CVD. It is not unlikely that the AhR played a role in this.
Obesity and wider metabolic dysregulation are common in BD,
especially as a consequence of the effects of prescribed mood stabi-
lizers and antipsychotics [182,232]. It is of note that AhR activation
is associated with increased high fat/sugar-induced adiposity and
wider metabolic dysregulation in rodents [233]. Melatonin has been
widely shown to decrease such diet-induced obesity and metabolic
dysregulation [234], including when as a consequence of antipsy-
Bipolar Disorder: The Role of the Kynurenine a nd Melatonergic Pathways Current Pharmaceutical Design, 2016, Vol. 22, No. 8 997
chotic medication in BD patients [235]. As to whether enhanced
AhR activation may be mediating such metabolic dysregulation, at
least in part, via the downregulation of melatonin and 7nAChr
effects has still to be determined.
Overall, the AhR may be intimately associated with many of
the processes thought to form the biological underpinnings of BD.
However, as can be gathered from the above, the variety of ligands
and their variable effects on the AhR in many cells, tissues and
organs considerably complicates its study in BD, as in other medi-
cal conditions.
AhR Cellular Effects
Given its wide array of effects and the differential impact of
different AhR ligands, it is not possible to describe a simple se-
quential set of changes and processes that drive AhR signalling.
However, an important aspect of AhR signalling seems to be the
activation of small GTPases, classically linked to mechanisms that
drive cellular plasticity. In this vein, it is interesting to note that
AhR activation can also increase levels of the spingosine-1-
phosphate (S1P)1r [236] and its associated effector small GTPase,
Rac1 [237]. The S1P1r is an important regulator of vascular barrier
integrity [238] and BBB permeability [239], as well as the migra-
tion of immune cells [240]. As well as chemoattracting immune
cells, regulating barrier permeability and modulating immuno-
inflammatory processes, activated S1Pr subtypes may differentially
interact with TLR4 induced NF-kB, which the S1P1r inhibits, and
which the S1P3r potentiates, as evidenced by TLR4 induced NF-kB
in different cell types [241,242]. S1Pr subtype interactions with
TLR4 requires testing as to their impact on the role of TLR4 activa-
tion in BD, including when arising from increased gut permeability
and from stress-induced HMGB1 from glia as well as its interac-
tions with the AhR. Prolonged activation of the S1P1r or the pro-
longed production of S1P increases levels of the S1P3r in the
plasma membrane, with the S1P3r have differential, and sometimes
opposing, effects compared to the S1P1r [238]. As such, some of
the differential effects of the AhR may arise from variations in the
levels of activity of different S1Pr subtypes.
The S1P1r induction of Rac1 leads to NADPH Oxidase activa-
tion, in turn driving an increase in the ROS production that fre-
quently underpins cellular reactivation, including in glia and im-
mune cells. AhR activation is a significant driver of NADPH Oxi-
dase activity [243], with NADPH Oxidase being the main contribu-
tor of HMGB1 effects in the brain [244] and other cell types [245].
This suggesting that the AhR may parallel the effects of chronic
stress in activating microglia via HMGB1 release and TLR activa-
tion [113], with the effects of TLR4 activation being determined by
the levels of S1P released and the S1Pr subtypes activated. AhR
activation may also interact via epigenetic mechanisms, a process
that may involve the activation of another small GTPase, RhoA
[246]. AhR/RhoA activation also increases BBB permeability
[247]. It is of note that RhoA is commonly induced by S1P2/3 re-
ceptors shifting into plasma membrane lipid rafts, in response to
initial S1P1r activation, should S1P release be prolonged [238].
Given that S1Pr subtype activation may be an intimate aspect of
lipid raft reorganization, and associated alterations in the clustering
of receptors and signalling pathways, it is not unlikely that the AhR
is having differential effects, in part driven by different S1Pr sub-
types having opposing effects on TLR4 induction of NF-kB. Con-
sequently, variations in AhR effects may be intimately associated
with NADPH Oxidase derived ROS that drives plasma membrane
lipid raft reorganization, coupled to variations in the levels of S1Pr
subtype activation. Such cellular plasticity processes require inves-
tigation in different cell types in BD, including as to how mem-
brane plasticity regulation drives the AhR effects that are linked to
TRYCAT and melatonergic pathway regulation.
Melatonin and AhR
As indicated throughout the above, the ultimate effects of in-
creased O&NS, immune-inflammatory cytokines and AhR activa-
tion may be a driving down of the melatonergic pathways in differ-
ent cell types and tissue, with consequences for increased glia and
immune cell reactivity that contribute to chronicity, neuroprogres-
sion and autoimmune processes in BD. Importantly, melatonin
inhibits the expression of many AhR-induced genes, including the
cytochrome P450 (CYP)1A1, CYP1A2 and CYP1B1 [248]. As
such, melatonin inhibits its own metabolism by the AhR-induced
CYP1B1 [249]. This may be important, as pro-inflammatory cyto-
kine induced IDO and stress-induced TDO that drive increases in
kynurenine, and therefore AhR activation, will also be enhancing
the breakdown of melatonin, in turn lowing melatonin effects, in-
cluding in the inhibition of AhR-induced genes. Given the high, but
relatively uninvestigated, levels of AhR in the brain and the raised
levels of kynurenine in BD, it is not unlikely that this will contrib-
ute to a lowering of brain melatonin, including when induced at
non-pineal sites. This would also be expected to alter the levels of
the 7nAChr, thereby impacting on mitochondria and cognitive
functioning as well as glia and immune cell reactivity [170,250]. As
such, the regulation of the TRYCAT pathways, may be intimately
associated with AhR activity, in turn modulating melatonin and
7nAChr levels and effects. Any loss of the antioxidant effects of
melatonin, including via its induction of endogenous antioxidants
will increase levels of ROS, thereby contributing to changes in
reactivity, HMGB1 release and TLR4 activation, in turn modulating
AhR effects.
We now focus on the role of the melatonergic pathways in BD,
having placed this pathway in the wider context of O&NS, im-
mune-inflammation and TRYCAT changes above.
Pineal and Circadian Melatonin in BD
Melatonin secretion abnormalities have been proposed to be
both “state” and “trait” markers of BD [5]. Indeed, as a state
marker, it has been observed that BD patients during episodes of
mania, versus BD depression and controls, appear to present with
elevated daytime melatonin levels [251]. The most important
changes seem to be during the afternoon and evening hours and did
not appear to be due to a phase advance [251]. Elevated levels of
melatonin during mania have been replicated [252], some authors
have found decreased serum melatonin levels compared with con-
trols and no significant differences between manic and depressive
episodes, compared to euthymic phases [253-257]. These latter
findings may indicate a disruption of temporal serum melatonin
regulation in BD compared to controls. Although such finding may
represent the heterogeneity of BD presentations and/or be the result
of neuroprogressive processes [22], should this data be replicated in
future studies, it may also provide an explanation as to the disrup-
tions evident in the diurnal expression patterns of several clock
genes in BD [251-254]. It should also be remembered that there are
many sources of melatonin, including the gut where melatonin ex-
pression levels can be over a hundred-fold greater than that pro-
duced by night-time pineal gland production. As such, the cellular
and tissue sources of any increased daytime melatonin in BD will
be important to determine, including as to the effects of medication
on this, with valproate known to upregulate the melatonergic path-
ways and melatonin receptors [83,93].
BD patients during remitted phases are supersensitive to mela-
tonin suppression by light [253-257], which has been proposed as a
“trait” biomarker of BD, given its independence of patients’ mood
state, strong heritability and increased prevalence with increased
familial genetic load [253,254]. Confirming this genetic influence,
monozygotic and dizygotic twins studies have shown that both
998 Current Pharmaceutical Design, 2016, Vol. 22, No. 8 Anderson et al.
overall melatonin levels and suppression by light at 500-lux have a
strong genetic component [257]. Consistent with the existence of an
abnormal light suppression of melatonin function, some data indi-
cate the existence of abnormal concentration and periodicity of
nocturnal melatonin secretion in BD, especially in euthymic BD I
subtype patients (significantly lower melatonin levels at baseline
and following light exposure of 500-lux light administered between
2 and 4 AM, as well as a later peak time for melatonin on the dark
night [253]. Furthermore, these abnormalities appear more promi-
nent in patients with BD, when compared to unipolar MDD [258].
Although melatonin onset abnormalities are associated with BD,
some discrepancies are observed in the literature regarding the ex-
act nature of the change [259]. Also, a few studies found no mela-
tonin suppression by light in BD patients versus healthy controls
Melatonin plays a role in the therapeutic actions of mood stabi-
lizers in BD [262]. Indeed, the two main mood-stabilizers, lithium
and valproate, reduce melatonin suppression in response to light in
healthy volunteers [262,263]. Moreover, lithium has been demon-
strated to act on the nocturnal melatonin peak and to modify the
function of the retinal-hypothalamic pineal pathway [264]. Val-
proate increases melatonin receptors in C6 glioma cells and in the
hippocampus of the rat brain [83,93]. Exogenous melatonin demon-
strates efficacy with both objective (polysomnography or actigra-
phy) and subjective measures (scales, questionnaires, sleep logs) in
treating several comorbidities associated with BD [265], such as
insomnia [266], delayed sleep phase disorder [267], and sleep dis-
turbances, such as sleep onset latencies and total sleep time [268].
Finally, preliminary data demonstrate that exogenous melatonin and
melatonin agonists may be useful for improvement of sleep quality
and relapse prevention in BD [1].
Anatomic studies of the pineal gland have shown similar re-
sults, with two studies having been performed. The first study
showed no difference in the total pineal gland volume between BD
patients and healthy subjects [269]. This is supported by a subse-
quent study of BD and schizophrenic patients, versus controls, with
results indicating an absence of any significant deviation in pineal
gland morphology [270].
Biochemically, one hypothesis of melatonin effect in BD is the
stabilization of the circadian rhythm of neurotransmitters systems
involved in the pathophysiology of BD, with a number of neuro-
transmitters being associated with rhythmic deregulations in BD
patients [1]. Indeed, it has been extensively shown that serotoner-
gic, dopaminergic and noradrenergic neurotransmissions play an
important role in the regulation of mood as well as following a
circadian variations as to their levels of secretion, catabolism and
receptors [271]. Given that serotonin acts as the immediate precur-
sor for NAS, and therefore for melatonin synthesis, there exists a
close connection between the pineal gland and serotonin [34].
Measures of CSF levels of serotonin and its metabolites show tryp-
tophan and 5-hydroxy indoleacetic acid to be associated with sig-
nificant biological variability over a 24 hours period. Furthermore,
data shows tryptophan levels to be relatively increased at night and
diminished at mid-day [272]. Moreover, pharmacological, although
not physiological, doses of melatonin can inhibit serotonin uptake
by platelets [273]. In the retina, dopamine has multiple roles related
to circadian rhythmicity, including as a chemical messenger for
light adaptation [274]. Amacrine cells release dopamine that acti-
vates dopamine receptors distributed throughout the retina and
modulate melatonin secretion via the retinal-hypothalamic-pineal
pathway [274]. Moreover, in the ventral tegmental area and sub-
stantia nigra, dopamine signaling is a key component of the
sleep/wake cycle and REM sleep [271,272]. In the SCN, the activ-
ity of the CLOCK-BMAL1 complex is regulated by dopaminergic
D2 receptors [271]. Another neurotransmitter, noradrenalin acts to
regulate melatonin synthesis in the pineal gland [270]. It has been
demonstrated that maximal concentrations of pineal AANAT and
melatonin were related to changes in the activity of pineal beta
adrenergic receptors, with their stimulation by noradrenaline being
necessary for pineal melatonin synthesis [275]. Taking together,
these findings suggest that melatonin levels and circadian rhythmic-
ity of melatonin secretion are closely interconnected with seroton-
ergic, dopaminergic and noradrenergic neurotransmitters that are
involved in the regulation of mood in BD, including along the reti-
nal-hypothalamic-pineal pathway. It is of note that melatonin recep-
tors are also evident in the retina in animal studies [276], suggesting
that alterations in melatonin levels and its receptors may be another
site for influencing retinal-hypothalamic-pineal pathway activity.
Genetic studies indicate that a common polymorphism
(rs4446909) of the ASMT promoter, encoding one of the two en-
zymes involved in melatonin biosynthesis from serotonin, is associ-
ated with BD and has weaker transcription effects, leading to lower
levels of ASMT activity [277]. Interestingly this BD-associated
functional variant has been examined along with actigraphy meas-
ures in remitted BD patients and healthy controls, where it was
shown to influence sleep and circadian rhythms [278].
In summary, several pineal gland melatonin secretion abnor-
malities are evident in BD, apparently during all phases of the dis-
order. These alterations are probably better explained by functional
abnormalities and changes in the regulation of melatonin, rather
than abnormal pineal gland volume. These functional alterations
may include gene polymorphisms in melatonin biosynthesis key
enzymes; serotonergic, dopaminergic and noradrenergic neuro-
transmissions abnormalities; supersensitivity to melatonin suppres-
sion (trait marker); and temporal dysregulation of melatonin pro-
files during mood states, perhaps especially during manic episodes
(state marker).
Non-pineal Sources of Melatonin and BD
As indicated above, there are many non-pineal sources of mela-
tonin, including the gut [279], astrocytes [280], retina [281],
macrophages [76,78] and fibroblasts [282]. Given that the former
three tissue and cellular sources of melatonin could plausibly be
linked to the etiology, course and/or treatment of BD, the data per-
taining to the genetic susceptibility factors in the melatonergic
pathways that associated with BD, may be well see as a conse-
quence of altered melatonin at these non-pineal sites [93]. In fact,
recent work suggests that all mitochondria containing cells may
produce melatonin [283], implicating wider cells and tissues that
genetic and epigenetic alterations in the melatonergic pathways
may regulate.
An important aspect of this could be the role of alterations in
melatonin production by glia and immune cells. Given the data
showing that O&NS and immune-inflammatory cytokine levels are
increased in BD, the role of non-pineal melatonin, as an antioxidant
and anti-inflammatory, may be important in the regulation of reac-
tive oxygen species (ROS) and immune-derived pro-inflammatory
cytokines. Melatonin is a known inhibitor of astrocyte [284] and
microglia reactivity [285], as well as of immune cell reactivity
[286], with the autocrine effects of melatonin in macrophages lead-
ing to an M2-type, anti-inflammatory, pro-phagocytic phenotype
[76]. With such factors known to be altered in BD [4], this would
suggest that changes in levels of melatonin are likely to have a
shotgun effect across a number of cells, tissues and processes in
It is likely that these non-pineal sources of melatonin may act to
regulate pineal melatonin synthesis. Recent work by Regina Markus
and colleagues [77] has shown that an increase in pro-inflammatory
cytokines, perhaps especially tumor necrosis factor-alpha (TNF-)
[78], will act to switch off pineal melatonin synthesis. As such, any
increase in O&NS and immune-inflammatory cytokines may act to
switch off pineal melatonin synthesis, indicating that the variations
in melatonin levels in BD may be intimately intertwined with al-
terations in oxidant regulation and immune-inflammation.
Bipolar Disorder: The Role of the Kynurenine a nd Melatonergic Pathways Current Pharmaceutical Design, 2016, Vol. 22, No. 8 999
Gut permeability and alterations in gut microbiota are at the
cutting edge of research in a number of conditions, including unipo-
lar depression [112]. Increased gut permeability allows tiny frag-
ments of partially digested food and gut bacteria to cross over the
gut barrier, triggering an immune response. Recent work suggests
that this phenomena could be a means by which unipolar depression
links to the neurodegenerative disorders, such as Alzheimer's dis-
ease, Parkinson’s disease and multiple sclerosis [137,287]. Mela-
tonin is a significant inhibitor of gut permeability [114,288], in part
by decreasing the activity of the NLRP3 inflammasome [289], as
well as acting via the 7nAChR [114,288]. The crossing of LPS
over the compromised gut barrier, may lead to the activation of
TLR4, which lead s to the release of TNF- in pineal microglia, and
the suppression of pineal melatonin synthesis. This is another
means by which alterations in melatonin levels at different periph-
eral sites may impact on levels of pineal melatonin synthesis and
thereby with the alterations in circadian regulation that is evident in
BD, as well as in many other depression-associated conditions
The role of these non-pineal sources of melatonin in BD is
looked at in more detail below, in the section 'Integrating TRY-
CATs, Melatonergic and wider pathways in BD'. Before this is
looked at, it is important to review some of the data on NAS, in-
cluding as to the possible relevance of the NAS/melatonin ratio in
N-acetylserotonin (NAS)
NAS is relatively unexplored in comparison to melatonin.
However, this may have important implications as NAS, although
also a powerful antioxidant, does not have the same effects as mela-
tonin, suggesting that variations in the NAS/melatonin ratio may be
of some importance across a wide array of medical conditions, in-
cluding BD [4]. NAS is released by cells, including pinealocytes in
the pineal gland, along with melatonin. Like melatonin, NAS is
amphiphilic, being able to readily cross membranes. NAS is also a
brain-derived neurotrophic factor (BDNF) mimic [290]. As such,
NAS is likely to have local and systemic effects that arise from its
activation of the BDNF receptor, TrkB, including on their effects
on immune cell regulation. Alterations in circulating and central
BDNF have been found in BD, with correlations between circulat-
ing BDNF with amygdala [291] and hippocampal volumes [292],
suggesting that the BDNF-mimicking effects of NAS may be of
some relevance in BD.
Role for NAS/Melatonin Ratio
In comparison to data on the role of melatonin in BD and many
other medical conditions, there has been a relative paucity of stud-
ies on NAS and very little on any role for variations in the
NAS/melatonin ratio. This is partly due to the difficulty in the
measurement of these compounds, due to their low circulating lev-
els, especially that of NAS. Few methods offer the sensitivity,
specificity and dynamic range needed to monitor NAS and mela-
tonin, as well as their metabolites in small blood samples, such as
those that can be obtained from early developmental studies and
paediatric studies of BD [293].
The few studies that have looked at alterations in NAS/ mela-
tonin ratio have not looked at its role in BD and have only looked at
plasma measures. One such study in first episode neuroleptic-naive
schizophrenic patients found an increase in the NAS/melatonin
ratio, when compared to healthy controls [294]. Yao and col-
leagues, in noting that early studies in schizophrenia found both
hyper- and hypo-serotonemia that were variously associated with
the longitudinal course of this disorder, led these authors to investi-
gate 13 plasma TRYCATs in 25 first-episode neuroleptic-naive
schizophrenia patients, in comparison to 30 healthy controls [294].
This study also compared TRYCATs and melatonin metabolites at
baseline and at 4 weeks following antipsychotic treatment. Yao and
colleagues showed that NAS was increased in the patient sample in
comparison to healthy controls. As well as an increase in the
NAS/melatonin ratio, these authors also found that the
NAS/tryptophan and melatonin/serotonin ratios were also increased
in patients versus controls. Such alterations were not evident fol-
lowing treatment with antipsychotics, suggesting that such ratios
are significant treatment targets of antipsychotic medications. An-
other indicant of the relevance of alterations in these pathways in
psychosis, is the lack of significant correlations of serotonin with
tryptophan, melatonin, kynurenine or tryptamine, which showed
strong correlations in the control group. As such, alterations in the
TRYCAT and melatonergic pathways are evident in first-episode
schizophrenia patients, suggesting, given the often shared vulner-
ability and pathophysiology of schizophrenia with BD, that changes
in the relative proportion of these factors may be of relevance in
BD. Given that NAS mimics BDNF, by activating TrkB [290],
which may be associated with antidepressant induced switching to
mania [295], it is important to look at the factors that may regulate
the NAS/melatonin ratio.
Potential Modulators of the NAS/Melatonin Ratio in BD
Although most studies looking at the melatonergic pathways
look to melatonin as the final product, melatonin is in fact metabo-
lized into a number of other factors, such as AMK and AFMK, that
also have significant antioxidant effects. As such, an increase in the
metabolism of melatonin could in principle influence the
NAS/melatonin ratio. It should also be noted that melatonin can be
converted back to NAS by a process of O-demethylation, which
may be driven by CYP2C19 [296,297]. CYP2C19, belonging to the
cytochrome P450 (CYP450) metabolizing-enzyme family, is impli-
cated in the metabolism of many psychiatric drugs, including anti-
depressants. CYP2C19 gene is highly polymorphic with the exis-
tence of different alleles resulting in modified enzyme activity. In
addition, drug-drug interactions have been described leading to
either induction or inhibition of CYP2C19 activity. As such, factors
that modulate CYP2C19 activity would be expected to impact on
the NAS/melatonin ratio [297]. Different SSRIs may differentially
impact on such regulation of the NAS/melatonin ratio [296,298].
As has been noted throughout the previous sections, there are a
number of factors that may be associated with the biological under-
pinnings of BD, including the AhR, gut permeability, stress-
induced cortisol, O&NS and pro-inflammatory cytokines as well as
variations in glia and immune cell reactivity. As noted throughout
the sections above, these factors are all linked to the TRYCAT and
melatonergic pathways, with relevance for the etiology, course and
treatment of BD. In this section, we try to link such diverse and
previously disparate data on the biology of BD within a model that
emphasizes the role of TRYCAT and melatonergic pathway
changes, which is illustrated in Fig. 2. Other data pertaining to the
biology of BD may also be integrated within this model.
Switching Poles in BD
Understanding of the switch from depression to mania is often
described as the 'holy grail' of BD research. In a recent large-scale
study of switching within four months of a depressive phase in
1720 BD patients, Niitsu and colleagues investigated 8403 episodes
without switch as well as 512 episodes where a period of depression
was followed by a switch [299]. Several baseline factors associated
with switching included: younger age, history of rapid cycling,
severe manic symptoms in previous episodes, previous suicide at-
tempts and a history of amphetamine use, as well as some pharma-
cological and psychotherapeutic treatments. In the immediately
prior depressive episode, risk factors included: possible mood ele-
vation, at least moderate severity multiple mania-associated symp-
toms and co-morbid panic attacks.
1000 Current Pharmaceutical Design, 2016, Vol. 22, No. 8 Anderson et al.
Switching can be induced by antidepressants, which is a signifi-
cant concern when treating BD depression that requires careful
monitoring. Generally, the rate of switching is significantly higher
for venlafaxine versus SSRIs or third generation antidepressants,
such as bupropion, perhaps especially when rapid-cycling is evident
[300,301], with the general consensus being that tricyclics and ven-
lafaxine, are more likely to induce hypomania/mania versus SSRIs
[302]. However, Eppel and colleagues have raised a number of
concerns in regard to such studies, including the relatively short
follow up period that would miss any cycle acceleration [303,304].
Antidepressant discontinuation-related mania may also contribute,
including when antidepressants are changed [305].
Another factor associated with switching is a history of am-
phetamine use. Amphetamine has long been used to manage de-
pression, especially when other pharmaceuticals have failed in the
treatment of suicidal patients [306]. Amphetamine is also widely
known among BD patients to be a means to trigger a manic episode,
contributing to its misuse. Amphetamine administration is also a
commonly used rodent model of mania [295]. Both amphetamine
and antidepressants, including venlafaxine, are associated with
alterations in the regulation of BDNF [307,308].
BDNF has been suggested to play a key role in BD [309]. Anti-
depressant-induced switching may be linked to an increase in
BDNF, including by venlafaxine [308], with amphetamine also
having differential effects on BDNF mRNA and protein in different
CNS regions [307]. Interestingly, BDNF shows significant altera-
tions in different brain regions in animal models of mania [295].
However, when measured peripherally in BD patients BDNF is
generally decreased, although this effect is confounded by age and
illness duration [310]. Increased levels of peripheral BDNF have
been found in rapid-cycling BD presentations [311], with medicated
BD patients in mania having higher BDNF levels versus medicated
BD patients in the depressive pole [312]. Importantly, NAS can
increase BDNF, including in the murine hippocampal dentate gyrus
[313], suggesting that variations in local NAS synthesis in different
CNS regions may not only activate TrkB, but also increase BDNF
production, with relevance to the different poles of BD presenta-
A number of factors altered in BD may act to differentially
regulate NAS and melatonin via the variable regulation of AANAT
and HIOMT or the levels of reconversion to serotonin, including by
CYP2C19/CYP2D6 [314], or the approximately 15% of melatonin
that is 'back-converted' to NAS [315]. The differential regulation of
AANAT and HIOMT may be associated with 14-3-3, Ca2+ and
glutamate [316-318], which are all subject to alteration in BD
[191,319,320]. Following the NE induced cAMP/PKA-driven
phosphorylation of AANAT at its N- (Thr31) and C-(Ser205) ter-
minal region, phosphorylated AANAT binds to 14-3-3 proteins,
forming a pAANAT/14-3-3 complex that stabilizes the AANAT
enzyme and protects it against proteolytic destruction. The
pAANAT/14-3-3 complex also induces allosteric changes in the
AANAT molecule that increase its enzymatic activity. AANAT and
HIOMT may therefore be differentially regulated, although in-
creased AANAT and NAS levels have classically been seen as a
Fig. (2). Shows how O&NS, mitochondria dysfunction, immune-inflammation and cortisol/stress, as well as possibly increased gut permeability may interact
to drive a decrease in the tryptophan/kynurenine ratio via the induction of IDO and TDO. Activation of the kynurenine pathways can increase AhR activity,
further contributing to IDO activation as well as leading to AHRE-III up-regulation of dopamine, which may also alter the ACh/dopamine ratio, which is pro-
posed to drive changes in inter-area neuronal patterning that underlies manic and depressive symptoms. An increase in the NAS/melatonin ratio is proposed to
underlie the shift to the manic pole, which may be driven, in part, by changes in glia reactivity, as well as genetic and epigenetic factors, that contribute to
alterations in the regulation of the NAS/melatonin ratio. A decrease in melatonin will also impact on a7nAChR levels and activity, with consequences for
cognition, as well as ACh/dopamine patterning.
Mito O&NS
? Genetic
Altered Neuronal
Manic Poles
Bipolar Disorder: The Role of the Kynurenine a nd Melatonergic Pathways Current Pharmaceutical Design, 2016, Vol. 22, No. 8 1001
simple relay station to increased melatonin. This is unlikely to be
the case. As such, the regulation of the NAS/melatonin ratio, in
different cells and brain regions, will be interesting to determine
over the course of BD.
Role of AhR in BD Switching
The AhR activation by TCDD increases the AhR response ele-
ment-III (AHRE-III), which is a significant regulator of tyrosine
hydroxylase, and therefore of dopamine synthesis [321]. An in-
crease in dopamine is associated with enhanced arousal and motiva-
tional behaviours, primarily via its release by the ventral tegmental
area (VTA) into many other brain regions, especially the striatum,
but also the frontal cortex and amygdala [322]. In BD models, the
two BD poles are thought to be driven by alterations in ACh in the
depressive pole, with increased dopamine thought to drive manic
presentations [323]. As such, should the activation of the AhR by
kynurenine parallel its activation by TCDD, kynurenine is likely to
impact on the levels of AhR induced AHRE-III and the levels of
VTA dopamine synthesis. KYNA activation of the 7nAChR will
also act to regulate PFC levels of dopamine release [324], as well as
glutamate and ACh itself [324,325]. As such, variations in the lev-
els of specific TRYCAT products, kynurenine and KYNA via AhR
and 7nAChR activation, are likely to be intimately associated with
the regulation of important aspects that differentiate the two BD
poles, and could be a key component of models of the catechola-
mine-ACh balance as underpinning the two poles of BD [326]. It
will be interesting to determine as to whether the AhR and/or the
7nAChR have any differential regulatory impact on AANAT and
HIOMT, and therefore on the NAS/melatonin ratio. A role for the
AhR in BD suggests that exposure to cigarette smoke or environ-
mental toxins, via AhR activation, could have a role in the etiology
and course of BD, as well as being a significant treatment target.
Relevance to Inter-area Brain Connectivity in BD
Genetic alterations in the dopaminergic pathways and their
alteration by methylation regulate the risk of BD [327]. Also SNPs
in the dopamine transporter, DAT1, increase BD risk [328]. Some
of the effects of stress and HPA axis dysregulation are thought to be
mediated by alterations in dopaminergic signalling [329]. Dopa-
mine-replacement therapy in Parkin son's disease is associated with
mania and hypomania in a sizeable percentage of patients [330],
highlighting the relevance of dopamine to manic presentations.
However, dopamine is very diffusely increased during VTA activa-
tion, being associated with the regulation of many factors, including
general arousal levels, motivation and cognition, including working
memory [331]. An important, but under-investigated, aspect of
increased dopamine release may be its effects in the paracapsular
cells of the intercalated masses that surround much of the
amygdala. Classically, paracapsular cells have been viewed as relay
cells that mediate the negative feedback of the PFC to the
amygdala. However, data in rodents shows that increased dopamine
activation of the D1r in paracapsular cells can lead to G protein-
coupled inwardly-rectifying potassium channels (GIRK)-induced
hyperpolarization, in turn blocking PFC negative feedback to the
amygdala, thereby increasing and prolonging the activation of the
amygdala by a process of disinhibition [54,332]. This is one means
by which a heightened affective influence on thought and behav-
ioural outputs may occur, and would be parsimonious with the data
showing increased and/or prolonged amygdala activity in BD [333].
For example, during mania, heightened dopamine inputs to the
nucleus accumbens (N.Acc) would be associated with increased
motivation, coupled to an increased amygdala and decreased PFC
and hippocampal influence on motivated outputs. Recent work
shows some differentiation of amygdala connectivity in BD manic
versus depressed poles [334]. Also, SNPs in the dopamine D1r are
associated with BD risk, as well as the cognitive symptomatology
of BD patients [335], with a recent meta-analysis showing that
SNPs in the dopamine D1r are more associated with BD than
schizophrenia [336].
This leads to a model suggesting that the increased HPA axis,
O&NS and pro-inflammatory cytokines evident in BD, induce TDO
and IDO, thereby increasing TRYCATs that may then act on the
AhR and 7nAChR, in turn modulating levels of ACh and dopa-
mine release in different CNS regions that contribute to the differ-
ential changes occurring in the two BD poles. The activation of
such pathways will also act to regulate the melatonergic pathways,
including by modulating the relative availability of tryptophan for
serotonin synthesis. The BD changes in 14-3-3, Ca2+ and gluta-
mate, as well as CYP2C19/CYP2D6, that may well arise in such a
model, including when driven by epigenetic and genetic variations,
could then impact on the relative levels of NAS and melatonin,
differentially in different CNS regions. Given that BDNF increases
VTA dopamine release [337], this would suggest that any relative
increase in NAS in the VTA may contribute to the role of height-
ened dopamine in BD symptomatology. Such rodent-derived data is
supported by data in humans showing that genetic variations in
BDNF modulate stress and reward processing via changes in VTA
dopamine [338].
Data by Anthony Grace and colleagues may be relevant to this
[339]. These authors showed that amygdala activation is able to
over-ride hippocampal and frontal cortex inputs into the VTA-
N.Acc junction [339], with differential effects of acute versus
chronic stress on this [340]. This would suggest that heightened
amygdala activity in BD is likely to drive an affective-mediated
regulation of motivated outputs, at the expense of higher-order
cognitive influences. Any concurrent increase in BDNF or NAS at
the VTA-N.Acc junction are likely to enhance affect-driven neu-
ronal plasticity and subsequent amygdala-associated influences on
motivated outputs.
It should be noted that such a model is not restricted to changes
occurring in the CNS, as changes in the gut and systemic immune
cells will also be modulated by the above processes. For example, a
genetic or inflammation driven decrease in gut melatonin, may
contribute to increased levels of gut permeability and inflammation
that raise levels of CRP [341]. Increased CRP is found in BD, as
well as many other medical conditions where raised levels of in-
flammation are evident [341]. A meta-analysis of studies looking at
CRP in BD found that it is increased in the manic phase as well as
in euthymic patients, but not in BD patients in the depressed phase
[51], irrespective of medication. CRP can increase BBB permeabil-
ity, leading to an enhanced extravasation of systemic immune cells
into the CNS, contributing to local inflammatory processes [52].
The driving down of melatonin levels by TRYCAT activation will
also lower melatonin's protection of BBB integrity [342]. As such,
the O&NS and immune-inflammatory processes, by increasing
TRYCATs will contribute to alterations in melatonin's regulatory
effects in the gut and BBB, in turn increasing immune-inflam-
matory processes and their effects centrally.
Stress and Central Inflammation
Enhanced levels of cortisol activity that may occur in stress,
including from the HPA axis, but also from the activation of central
11B-HSD1. 11B-HSD1 is differentially unregulated in the hippo-
campus, amygdala and cortex in socially-stressed rodents [343].
Such stress leads to the induction of HMGB1 in microglia [211,
213]. The release of HMGB1 activates TLR4, enhancing microglia
reactivity that contributes to central inflammatory processes. Any
loss of central melatonin is likely to heighten such TLR4 activity
[344] and levels of microglia reactivity [345]. As such, heightened
levels of chronic stress are likely to contribute to many aspects of
the TRYCAT and melatonergic pathway regulation in BD, includ-
ing by increasing TDO and MAO, as well as central microglia reac-
tivity and innate immune-type responses in the brain.
1002 Current Pharmaceutical Design, 2016, Vol. 22, No. 8 Anderson et al.
Another route through which chronic stress may modulate im-
mune associated changes in BD is via increased gut permeability
[346], including via increasing CRP. Alcohol addiction and abuse
are not uncommon in BD. Alcohol also increases gut permeability,
in part by driving down melatonin levels, thereby contributing to
increased levels of endotoxemia in alcoholics [347]. This suggests
that alcohol abuse may contribute to BD symptoms and processes
of neuroprogression in BD, in part, via the regulation of gut perme-
ability. Gut permeability will be important to determine in BD,
especially as the prevention of increased gut permeability and asso-
ciated immune-inflammatory activation, prevents rodents from
becoming alcohol addicted [348].
In the context of such stress effects arising from cortisol dys-
regulation, it is important to note that the two main mood stabiliz-
ers, lithium and valproate, increase bcl2-associated anthanogene-1
(BAG-1) levels [349]. Given that BAG-1 prevents the nuclear
translocation of the Gcr [350], some of the therapeutic efficacy of
these mood stabilizers may then be on such stress effects on TDO,
MAO and pro-inflammatory cytokine induced TRYCATs, with
consequences for kynurenine activation of the AhR and AHRE-III,
in turn regulating dopaminergic activity. Stress/cortisol induction of
HMGB1 and TLR4 activation is also likely to be inhibited by mood
stabilizer induced BAG-1. It is also important to note that valproate
increases melatonin receptors in astrocytes [83], which may be
driven by increased local melatonin synthesis, given that melatonin
increases its own receptors. In this context, it is also of note that
both valproate and lithium increase 4 isoforms of 14-3-3 [351],
suggesting that they could be acting on levels of AANAT stability
and NAS production.
Role of the Cereb ellum in BD
A plethora of imaging studies have showed cerebellar altera-
tions in BD [352-354], with decreased cerebellar white and grey
matter being associated with illness duration prior to treatment
[355]. Improvements in BD symptomatology have been shown
following transcranial direct stimulation current modulation of pre-
frontal-thalamo-cerebellar pathways in BD patients [356]. How-
ever, cerebellar neurological soft signs do not discriminate BD from
schizophrenia [357], which is also associated with alterations in
cerebellar functioning and structure [358]. This perhaps suggests a
role for alterations in the cerebellum in psychosis more widely.
Classically, the cerebellum has been associated with the regula-
tion of balance and fine motor movements. However, recent work
has shown cerebellar involvement in a wide range of processes,
including implicit motor learning [359] and verbal working mem-
ory [360], as well as wider cognitive, behavioural [361] and affec-
tive processing [362]. Implicit motor learning deficits are evident in
BD patients [359], whilst the stimulation of the rodent cerebellum
inhibits central amygdala activation in a fear-conditioning paradigm
[363]. Overall, changes in the cerebellum are evident in BD, modu-
lating much of the BD symptomatology.
Melatonin is synthesized locally in the cerebellum, where it
attenuates the effects of LPS [364]. Interestingly, alcohol increases
cerebellar HMGB1, TLR4 and NLRP3 inflammasome activation,
without having any significant impact on circulating LPS levels
[365]. It is also of note that the a7nAChR regulates sensory inputs
into the cerebellum, where it modulates the shift from cerebellar
LTD to LTP [366]. Given the role of melatonin regulatory interac-
tions with the 7nAChR, it is not unlikely that variations in mela-
tonin, genetic and inflammation-driven, will impact on cerebellar
functioning. With a growing appreciation of the cerebellum in a
diverse array of processes, including cognition [361], changes in
cerebellar functioning in BD are likely to be linked to alterations in
the TRYCAT and melatonergic pathways.
Both TDO and IDO are expressed in the rodent cerebellum
[367,368], with the activation of the latter evident in reactive mi-
croglia [368], in association with increased TRYCAT synthesis. As
well as the classical synthesis of KYNA from kynurenine, catalysed
by kynurenine aminotransferases (KATs), recent data shows that
there are alternative routes to KYNA synthesis in the rodent cere-
bellum [162]. Blanco Ayala and colleagues showed that KYNA can
be increased by D-amino acid oxidase (DAAO) and the direct trans-
formation of kynurenine to KYNA by ROS [162]. Given the role of
KYNA in the inhibition of the 7nAChR and its influence on sen-
sory inputs and neuroplasticity processes in the cerebellum [366],
the interaction of melatonin and O&NS may be of particular impor-
tance in the cerebellum. It is also of note that SNPs in the DAAO
are associated with BD risk, as shown in a recent meta-analysis
The AhR is also evident in the cerebellum, where it regulates
important processes, including levels of neurogenesis [206] and
neuronal apoptosis [370], suggesting that alterations in AhR activa-
tion may contribute to the changes in cerebellar grey and whiter
matter in BD. Given that amygdala inputs to the cerebellum are
important to learning and cerebellar neuronal plasticity [371], it is
likely that there are significant interactions of amygdala changes in
BD with alterations in cerebellar development and plasticity that
involve changes in TRYCATS, O&NS, AhR and the melatonergic
How the diverse array of biological factors associated with BD
can be linked in this model is shown in the summary diagram in
Fig. 2.
As indicated throughout the article, the roles and interactions of
O&NS, immune-inflammatory activity, IDO, TDO and stress/ cor-
tisol in the modulation of the serotonergic and melatonergic path-
ways generates a number of areas for investigation in BD, as well
as for the development of appropriate pharmaceuticals. Many spe-
cific investigations have been indicated in the appropriate sections
above. However, a number of general, as well as specific investiga-
tive directions are worth emphasizing.
It will be important to determine as to whether variations in the
NAS/melatonin ratio are relevant to polarity shifts in BD patients,
including as to whether the proven efficacy of mood-stabilizers and
anti-psychotics in BD is mediated via the modulation of factors that
act to regulate this ratio, such as 14-3-3 isoforms. Gut permeability
has been shown to be relevant in MDD and this is in urgent need of
investigation in BD, including as to the degree to which the genetic
associations of the melatonergic pathways in BD are linked to al-
terations in melatonin's inhibition of gut permeability. Other factors
shown to have efficacy in the management of BD, including
omega-3 PUFAs, may also be acting via their regulation of gut
permeability [372]. As to whether, the role of increased glia reactiv-
ity in BD is linked to alterations in the melatonergic pathways will
also be important to determine, including as to the relevance of
variations in microglia melatonin release and autocrine effect in the
modulation of microglia phenotypes and threshold for reactivation.
It requires urgent investigation as to what extent, if any, the pleth-
ora of melatonin effects and benefits are driven by its regulation of
the 7nAChR in BD, as well as in other medical conditions [373].
A wide array of research data indicates that the cerebellum
plays a significant part in mood regulation, with cerebellar altera-
tions differentiating BD from MDD [354]. The melatonergic path-
ways are relevant modulators of immune and stress driven changes
in the cerebellum, suggesting that alterations in cerebellar melaton-
ergic pathways may be of some importance to the role of the cere-
bellum in BD.
Other wider areas of research, such as alterations in epigenetic
regulation and changes in the levels and patterning of microRNAs,
will be important to investigate in relation to changes in the mela-
tonergic pathways, down stream of O&NS and immune-inflam-
matory cytokines. As to whether alterations in the NAS/melatonin
Bipolar Disorder: The Role of the Kynurenine a nd Melatonergic Pathways Current Pharmaceutical Design, 2016, Vol. 22, No. 8 1003
ratio have any impact on such epigenetic processes will also be
interesting to determine.
The etiology, course and management of BD is intimately in-
volved with the relative activities of immune-inflammation, O&NS,
mitochondrial and TRYCAT changes. The picture now emerging
on the biological underpinnings of BD should enhance the under-
standing and treatment of BD per se, with concurrent benefits due
to how these inter-related processes contribute to the many BD
comorbidities, including obesity and addiction, that drive down life-
expectancy in BD patients.
AhR = aryl hydrocarbon receptor
AhRR = aryl hydrocarbon receptor repressor
AANAT = arylalkylamine N-acetyl-transferase
ASMT = acetylserotonin O-methyltransferase
BAG-1 = bcl2-associated anthanogene-1
BBB = blood-brain barrier
BD = Bipolar disorder
BDNF = brain derived neurotrophic factor
BMI = body mass index
DAAO = D-amino acid oxidase
CNS = central nervous system
CSF = cerebrospinal fluid
CRP = C-reactive protein
CVD = cardiovascular disease
DISC1 = disrupted-In-schizophrenia 1
HDAC = histone deacetylase
H-FABP = heart-type fatty acid binding protein
HIOMT = hydroxyindole O-methyltransferase
HMGB(1,2) = high mobility group box protein 1,2
HPA = hypothalamic pituitary adrenal
IDO = indoleamine 2,3-dioxygenase
IFN- = interferon-gamma
KAT = Kynurenine aminotransferase
KYNA = kynurenic acid
LTD = long-term depression
LTP = long-term potentiation
MAO = monoamine oxidase
MBP = myelin basic protein
MCP1 = monocyte chemotactic protein-1
MDD = major depressive disorder
NAS = N-acetylserotonin
NF-L = neurofilament light chain
NLRP3 = nod-like receptor pyrin domain-containing 3
O&NS = oxidative and nitrosative stress
TDO = tryptophan 2,3-dioxygenase
TRYCAT = tryptophan catabolite
PFC = prefrontal cortex
ROS = reactive oxygen species
SCN = suprachiasmatic nuclei
TCDD = 2,3,7,8-Tetrachlorodibenzo-p-dioxin
TDO = Tryptophan 2,3-dioxygenase
TGF- = transforming growth factor-beta
TLR4 = toll-like receptor 4
TNF- = tumor necrosis factor alpha
TSPO = translocator protein (18kDa)
VTA = ventral tegmental area
XRE = xenobiotic-responsive element
G. Anderson and A. Jacob have no conflict of interest.
F. Bellivier has received honoraria and financial compensation
as independent symposium speakers from Sanofi-Aventis, Lund-
beck, AstraZeneca, Eli Lilly, Bristol-Myers Squibb and Servier.
P.A. Geoffroy has consulted for Menarini and received com-
Declared none.
[1] Merinkangas KR, Tohen M. Epidemiology of bipolar disorder in
adults and children. In: Tsuang MT, Tohen MT, Jones PB, Eds.
Textbook in psychiatric epidemiology. Chichester: John Wiley and
Sons 2011; pp. 329-42.
[2] Leussis MP, Madison JM, Petryshen TL. Ankyrin 3: genetic asso-
ciation with bipolar disorder and relevance to disease pathophysi-
ology. Biol Mood Anxiety Disord 2012; 2(1): 18.
[3] Seifuddin F, Mahon PB, Judy J, et al. Meta-analysis of genetic
association studies on bipolar disorder. Am J Med Genet B Neuro-
psychiatr Genet 2012; 159B(5): 508-18.
[4] Anderson G, Maes M. Bipolar disorder: role of immune-
inflammatory cytokines, oxidative and nitrosative stress and tryp-
tophan catabolites. Curr Psychiatry Rep 2015; 17(2): 8.
[5] Geoffroy PA, Etain B, Franchi JM, Bellivier F, Ritter P. Melatonin
and Melatonin agonists as adjunctive treatments in Bipolar Disor-
ders. Curr Pharm Des 2015; 21(23): 3352-8.
[6] Maes M, Bosmans E, Calabrese J, Smith R, Meltzer HY. Interleu-
kin-2 and interleukin-6 in schizophrenia and mania: effects of neu-
roleptics and mood stabilizers. J Psychiatr Res 1995; 29(2): 141-52.
[7] Maes M, Meltzer HY, Bosmans E, et al. Increased plasma concen-
trations of interleukin-6, soluble interleukin-6, soluble interleukin-2
and transferrin receptor in major depression. J Affect Disord 1995;
34(4): 301-9.
[8] Maes M, Delange J, Ranjan R, et al. Acute phase proteins in
schizophrenia, mania and major depression: modulation by psycho-
tropic drugs. Psychiatry Res 1997; 66(1): 1-11.
[9] Wadee AA, Kuschke RH, Wood LA, Berk M, Ichim L, Maes M.
Serological observations in patients suffering from acute manic
episodes. Hum Psychopharmacol 2002; 17(4): 175-9.
[10] Anderson G, Kubera M, Duda W, Laso W, Berk M, Maes M.
Increased IL-6 trans-signaling in depression: focus on the trypto-
phan catabolite pathway, melatonin and neuroprogression. Phar-
macol Rep 2013; 65(6): 1647-54.
[11] Maes M, Anderson G, Kubera M, Berk M. Targeting classical IL-6
signalling or IL-6 trans-signalling in depression? Expert Opin Ther
Targets 2014; 18(5): 495-512.
[12] Myint AM, Kim YK, Verkerk R, et al. Tryptophan breakdown
pathway in bipolar mania. J Affect Disord 2007; 102(1-3): 65-72.
[13] Moylan S, Berk M, Dean OM, et al. Oxidative & nitrosative stress
in depression: Why so much stress? Neurosci Biobehav Rev 2014;
45: 46-62.
[14] Kim HK, Chen W, Andreazza AC. The potential role of the NLRP3
inflammasome as a link between mitochondrial complex I dysfunc-
tion and inflammation in bipolar disorder. Neural Plast 2015; 2015:
[15] Duffy A, Lewitzka U, Doucette S, Andreazza A, Grof P. Biological
indicators of illness risk in offspring of bipolar parents: targeting
the hypothalamic-pituitary-adrenal axis and immune system. Early
Interv Psychiatry 2012; 6(2): 128-37.
[16] Nievergelt CM, Kripke DF, Barrett TB, et al. Suggestive evidence
for association of the circadian genes PERIOD3 and ARNTL with
1004 Current Pharmaceutical Design, 2016, Vol. 22, No. 8 Anderson et al.
bipolar disorder. Am J Med Genet B Neuropsychiatr Genet 2006;
141B(3): 234-41.
[17] Scola G, Andreazza AC. The role of neurotrophins in bipolar dis-
order. Prog Neuropsychopharmacol Biol Psychiatry 2015; 56: 122-
[18] Markham JA, Koenig JI. Prenatal stress: role in psychotic and
depressive diseases. Psychopharmacology (Berl) 2011; 214: 89-
[19] Hamdani N, Daban-Huard C, Lajnef M, et al. Relationship between
Toxoplasma gondii infection and bipolar disorder in a French sam-
ple. J Affect Disord 2013; 148(2-3): 444-8.
[20] Bellivier F, Geoffroy PA, Etain B, Scott J. Sleep- and circadian
rhythm-associated pathways as therapeutic targets in bipolar disor-
der. Expert Opin Ther Targets 2015; 19(6): 747-63.
[21] McClung CA. How might circadian rhythms control mood? Let me
count the ways. Biol Psychiatry 2013; 74(4): 242-9.
[22] Gama CS, Kunz M, Magalhães PV, Kapczinski F. Staging and
neuroprogression in bipolar disorder: a systematic review of the lit-
erature. Rev Bras Psiquiatr 2013; 35(1): 70-4.
[23] León-Caballero J, Pacchiarotti I, Murru A, et al. Bipolar disorder
and antibodies against the N-methyl-d-aspartate receptor: A gate to
the involvement of autoimmunity in the pathophysiology of bipolar
illness. Neurosci Biobehav Rev 2015; 55: 403-412.
[24] Prieto ML, McElroy SL, Hayes SN, et al. Association between
history of psychosis and cardiovascular disease in bipolar disorder.
Bipolar Disord 2015; 17(5): 518-27.
[25] Grant BF, Goldstein RB, Saha TD, et al. Epidemiology of DSM-5
alcohol use disorder: results from the national epidemiologic sur-
vey on alcohol and related conditions III. JAMA Psychiatry 2015;
72(8): 757-66.
[26] Geoffroy PA, Scott J, Boudebesse C, et al. Sleep in patients with
remitted bipolar disorders: a meta-analysis of actigraphy studies.
Acta Psychiatr Scand 2015; 131(2): 89-99.
[27] Miller C, Bauer MS. Excess mortality in bipolar disorders. Curr
Psychiat Rep 2014; 16(11): 499.
[28] Claustrat B, Brun J, Chazot G. The basic physiology and patho-
physiology of melatonin. Sleep Med Rev 2005; 9(1): 11-24.
[29] Anderson G, Maes M. Local melatonin regulates inflammation
resolution: a common factor in neurodegenerative, psychiatric and
systemic inflammatory disorders. CNS Neurol Disord Drug Targets
2014; 13(5): 817-27.
[30] Takahashi JS, Hong H-K, Ko CH, McDearmon EL. The genetics of
mammalian circadian order and disorder: implications for physiol-
ogy and disease. Nat Rev Genet 2008; 9(10): 764-75.
[31] Lewy AJ, Wehr TA, Goodwin FK, Newsome DA, Markey SP.
Light suppresses melatonin secretion in humans. Science 1980;
210(4475): 1267-9.
[32] Zeitzer JM, Dijk DJ, Kronauer R, Brown E, Czeisler C. Sensitivity
of the human circadian pacemaker to nocturnal light: melatonin
phase resetting and suppression. J Physiol 2000; 526 Pt 3: 695-702.
[33] Snyder SH, Borjigin J, Sassone-Corsi P. Discovering light effects
on the brain. Am J Psychiatry 2006; 163(5): 771.
[34] Hickman AB, Klein DC, Dyda F. Melatonin biosynthesis: the
structure of serotonin N-acetyltransferase at 2.5 A resolution sug-
gests a catalytic mechanism. Mol Cell 1999; 3(1): 23-32.
[35] do Prado CH, et al. Reduced regulatory T cells are associated with
higher levels of Th1/TH17 cytokines and activated MAPK in type
1 bipolar disorder. Psychoneuroendocrinology 2013; 38(5): 667-76.
[36] Panizzutti B, Gubert C, Schuh AL, et al. Increased serum levels of
eotaxin/CCL11 in late-stage patients with bipolar disorder: An ac-
celerated aging biomarker? J Affect Disord 2015 ; 182: 64-9.
[37] Kessing LV, Vradi E, McIntyre RS, Andersen PK. Causes of de-
creased life expectancy over the life span in bipolar disorder. J Af-
fect Disord 2015; 180: 142-7.
[38] Knijff EM, Breunis MN, van Geest MC, et al. A relative resistance
of T cells to dexamethasone in bipolar disorder. Bipolar Disord
2006; 8(6): 740-50.
[39] Drexhage RC, Hoogenboezem TH, Versnel MA, Berghout A,
Nolen WA, Drexhage HA. The activation of monocyte and T cell
networks in patients with bipolar disorder. Brain Behav Immun
2011; 25(6): 1206-13.
[40] Liu HC, Yang YY, Chou YM, Chen KP, Shen WW, Leu SJ. Im-
munologic variables in acute mania of bipolar disorder. J Neu-
roimmunol 2004; 150(1-2): 116-22.
[41] Torres KC, Souza BR, Miranda DM, et al. Expression of neuronal
calcium sensor-1 (NCS-1) is decreased in leukocytes of schizo-
phrenia and bipolar disorder patients. Prog Neuropsychopharmacol
Biol Psychiatry 2009; 33(2): 229-34.
[42] Abeer, El-Sayed A, Ramy HA. Immunological changes in patients
with mania: changes in cell mediated immunity in a sample from
Egyptian patients. Egypt J Immunol 2006; 13(1): 79-85.
[43] Munkholm K, Braüner JV, Kessing LV, Vinberg M. Cytokines in
bipolar disorder vs. healthy control subjects: a systematic review
and meta-analysis. J Psychiatr Res 2013; 47(9): 1119-33.
[44] Modabbernia A, Taslimi S, Brietzke E, Ashrafi M. Cytokine altera-
tions in bipolar disorder: a meta-analysis of 30 studies. Biol Psy-
chiatry 2013; 74(1): 15-25.
[45] Aguilar-Valles A, Inoue W, Rummel C, Luheshi GN. Obesity,
adipokines and neuroinflammation. Neuropharmacology 2015;
96(Pt A): 124-34.
[46] Coplan JD, Gopinath S, Abdallah CG, Berry BR. A neurobiologi-
cal hypothesis of treatment-resistant depression - mechanisms for
selective serotonin reuptake inhibitor non-efficacy. Front Behav
Neurosci 2014; 8: 189.
[47] Barbosa IG, Rocha NP, Bauer ME, et al. Chemokines in bipolar
disorder: trait or state? Eur Arch Psychiatry Clin Neurosci 2013;