Early diagnosis, treatment and prevention of mood disorders targetting the activated inflammatory response system

Final Report Summary - MOODINFLAME (Early diagnosis, treatment and prevention of mood disorders targetting the activated inflammatory response system)

Executive Summary:

Executive summary:

A consortium of 15 European Universities/Research Institutes and 4 SME’s harboring expert psychiatrists, neuroscientists, immunologists, cell biologists, biochemists, pharmacologists and bio-informaticians has

1. Developed and studied in depth in a 4 year project 3 animal model systems of depressive-like behavior: the NOD mouse model, the Maternal Inflammation (MIA) model and the olfactory bulbectomized (OBX) rat model. All models are characterized by an activated inflammatory response system and/or an abnormal tryptophan metabolism. Using these animal models the consortium has been able to identify 3 major pathogenic pathways, which lead to aberrant immune-endocrine homeostatic set points and on the basis thereof in aberrant microglia-astrocyte-neuron interactions, particularly during the development of the brain. These aberrant interactions result in structural and functional deficits of important brain regions such as the hippocampus, hypothalamus and white matter tracts, e.g. in the corpus callosum.
2. Developed a series of assay systems to detect the aberrant immune-endocrine set point of the microglia and of peripheral immune cells in patients and individuals at risk for major mood disorders.
3. Developed a series of assay systems to detect abnormally regulated tryptophan breakdown products in patients and individuals at risk for major mood disorders.

Integration of the data of the animal model systems and the patients/individuals at risk has allowed the construction of a staging model of major pathogenic routes leading to severe mood disorders. We identify 2 stages in the development of a severe mood disorder

The first stage is a high risk pre-stage without the actual signs and symptoms of the mood disorder. In case of a high family load this stage is genetically determined and characterized by various genetic polymorphisms in immune (e.g. HLA-DR, MDL), endocrine (e.g. glucocorticoid receptor) and tryptophan catabolic (e.g. KMO) genes. In case of a major role for an early toxic event in the development of the mood disorder, epigenetic alterations of immune genes (due to early inflammatory conditions) or epigenetic alterations of glucocorticoid receptor genes (due to stress full life conditions) play a major role. The genetic and/or epigenetic high risk state is biologically characterized by an abnormal set point of the macrophage - T cell system/glucocorticoid-axis in general, which not only influences the set point of circulating monocytes and T cells, but more relevant to mood disorders also the set point of microglia-astrocyte-neuron interaction. This latter disturbed interaction leads to a defective production of neuronal growth and support factors by microglia, an altered purine signaling between microglia and astrocytes, a raised production of pro-inflammatory cytokines by microglia and an altered tryptophan metabolism with a high production of toxic metabolites in critical brain areas. This results in an abnormal structural and functional development of the brain, in particular in areas important for mood regulation, learning and an adequate responsiveness to stress.

The second stage is the stage of clinical signs and symptoms of the severe mood disorder. In children of a bipolar parent stress full live events are the most important eliciting factor. In this stage the underlying aberrant immune, endocrine and metabolic set points of the pre-stage can still be present or re-occur. We found them linked to particular signs and symptoms of the disorders, such as melancholia (within the depression spectrum) or a postpartum occurrence of the disorder (psychotic episodes) or mania (within the bipolar spectrum). The presence of a high inflammatory state also negatively determines the outcome of regular anti-depressive therapy.

The developed series of immune, endocrine and metabolic tests can in principle be used to predict and stage the mood disorders and predict treatment outcome and/or correct the abnormal immune-endocrine-metabolic set points. There are various potential drugs available that counteract the abnormal immune, endocrine and metabolic state and a variety of these drugs have been explored in the animal models and in vitro systems of MOODINFLAME. A COX-2 inhibitor study in patients is still in progress and will be continued in the PSYCH-AID project, which is an EU funded continuation IAPP project of MOODINFLAME. PSYCH-AID also further develops the test systems and will extract more data from the large bio-bank and database created in MOODINFLAME. Patents will be filed via PSYCH-AID. Also local projects (e.g. NL funded NWO-TOP grant) continue the therapeutic approach using anti-inflammatory drugs (statins) in psychiatric (schizophrenia) patients tested by the assay systems developed in MOODINFLAME.

The use and exploitation of the assay system developed in MOODINFLAME will have a significant future impact on the general health status of the EU population, particularly in relation to prediction of response to regular treatments or to novel immune modulating treatments. This will have an extraordinarily positive effect on health care and costs in Europe.

Project Context and Objectives:

A summary description of the project context and the main objectives. Concept

The concept of the project is that (at least part of) bipolar disorder, major depressive disorder and postpartum psychoses (collectively called the major mood disorders) are the result of a systemic abnormality, namely a proneness to inflammation. In our view these disorders are thus largely caused by a sort of mild “chronic inflammation” of certain brain areas important for mood regulation, such as the limbic system.

Objectives

1. The development of blood tests and brain scans to determine the proneness to inflammation in patients with a major mood disorder and its consequence for brain function.
2. The exploration of animal models with a depressive-like behavior and/or a proneness to inflammation to study the consequences and underlying mechanisms of such a “chronic inflammation” of the brain.
3. The exploration of in vitro monocyte/macrophage and microglia-astrocyte-neuron systems to study the basic mechanisms under-lying the discovered immune activations in psychiatric patients.
4. To study therapeutic correction possibilities using immune modulating medicines, such as anti-inflammatory drugs (e.g. COX-2 inhibitors) and to study the anti-inflammatory actions of regular psychiatric medications.

Achievements

Patient cohorts

The consortium (19 partners) has presently in store frozen leukocytes and serum of 734 major depressed patients (and 292 matched controls), 390 bipolar patients (and 264 matched controls), 229 postpartum psychosis patients at follow up after childbirth (and 161 matched controls), 80 offspring of a bipolar parent (in 4 collections ranging from 16 to 28 years of age with matched healthy controls and their parents), 30 bipolar twin pairs collected 2 times with an interval of around 5 years (and 20 control twin pairs, similarly collected) and 30 IFN-treated patients. The internet database has been constructed and clinical and laboratory data of all of these patients have been uploaded. Frozen leukocytes, serum and plasma of these patients have been stored (and in many cases already used).

Objective 1. The development of blood tests and brain scans to determine the proneness to inflammation in psychiatric patients and its consequence for brain function.

High through put systems (TLDA gene arrays, multicolor FACS systems, multiplex cytokine arrays and microRNA multiplex assays) to determine the immune state of patients have been developed and published.

Data show in individuals at risk for a major mood disorder and in individuals with a major mood disorder

1. An in the time fluctuating pattern of the presence in the circulation of inflammatory activated monocytes, the cells showing a characteristic fingerprint of both pro- and anti-inflammatory genes, including microRNA-146a - a known microRNA related to auto-inflammation- and a reduced ratio of the alpha/beta glucocorticoid receptor gene expression. Specific risk times for the activated monocyte profile are teenage, the postpartum period and older age. High monocyte inflammation correlated in bipolar disorder to mania symptoms and somatic gastro-intestinal symptoms, in major depressive disorder to feelings of guilt and somatic gastro-intestinal symptoms
2. A proneness of the monocytes to overreact with an enhanced secretion of inflammatory cytokines (IL-12 and TNF) to triggering of their TLRs by inflammatory stimuli. This proneness looks time-independent in preliminary observations.
3. An in the time fluctuating activation set point of the T cell system, with both anti- (T regulatory cells, Th2 cells) and pro-inflammatory forces (T cytotoxic cells, Th1 and Th17 cells) activated. Data analysis is in progress to study specific patterns related to age and confounders such as smoking.
4. Variable rises in pro-inflammatory cytokines in serum, particularly of IL-1 beta, IL-6, PTX3, TNF-alpha, CCL2 and sCD25, not only depending on disease phenomena, but also on gender, age, body mass index and serum lipid spectrum.

In patients with a major mood disorder an activated state of circulating monocytes/T cells determined the outcome of venlafaxine/imipramine treatment. Response to treatment was only seen in melancholic patients without monocyte inflammatory activation, a reduced level of cytotoxic T cells and high NK cells.

HPLC techniques to determine tryptophan catabolic pathway metabolites in patient blood have been validated and serum determination for virtually all the collected patients have been completed. Analysis of the data is in progress.

Antibodies to kynurenine break down products have been developed and an antibody to quinolate has been used in a morphological brain study of depressed suicide victims showing positivity of microglia in frontal brain areas.

MRS and BOLD fMRI studies in 30 MDD patients have been completed and blood samples await analysis for immune activation and tryptophan break down analysis.

A PET scan for activated microglia using the peripheral benzodiazepine receptor (PBR) as a microglial activation marker has been tested in 15 BD patients. The PBR is also part of the monocyte activation gene fingerprint and in vitro studies have revealed that activation of the PBR has anti-inflammatory properties as well as induction properties for the monocytes to develop into dendritic cells.

The saliva dexamethasone suppression test (DST) has been carried out in 409 mood disorder patients and 196 healthy controls. DST values are presently analyzed for the various patient groups. BD patients with a negative DST (thus with an HPA axis resistant to DEX) had the same monocyte alpha/beta glucocorticoid receptor gene expression as those with a positive DST (preliminary data), indicating that the steroid resistance at the level of the brain is not mimicked by a steroid resistance at the level of the circulating immune cells. EMC has developed a functional in vitro DST using leukocytes of patients and measuring the DEX suppressed cytokine production of monocytes and T cells to further study DEX insensitivity in relation to immune activation.

Objective 2. The exploration of animal models with a depressive-like behavior and/or a proneness to inflammation to study the mechanisms and consequences of such a “chronic inflammation” of the brain.

An LPS-induced prenatal inflammation mouse model has further deepened our understanding of how peripheral pro-inflammatory activation of peripheral monocytes/macrophages by a microbial component, i.e. LPS, gives rise to an immune activation of microglia (the monocyte/macrophage of the brain) in the limbic system during development leading to an abnormal hippocampus development/abnormal neuronal function, with later abnormal depressive-like behavior.

The major achievement has been that it appeared in whole genome analysis of the microglia that in the process many immune genes (both pro- and anti-inflammatory) were turned on, while many genes for axon guidance were turned off. This points in the direction that immune activated microglia neglects axonal growth/sprouting, explaining the abnormal hippocampus development in the fetal phase.

The newly developed NOD mouse model has also brought new exciting insights. This established model for autoimmune thyroiditis and autoimmune diabetes (developing after 12 weeks of age) has mild behavioral abnormalities before 10 weeks of age (hyperactivity), which can be severely aggravated to a strong depressive-like behavior by intra-peritoneal LPS injection. While the normal NOD mouse has a mild pro-inflammatory state of monocytes and microglia, LPS treatment induces an interferon-related inflammatory signature in its microglia. This points in the direction that a predisposition to autoimmunity (and a priming of microglia) can lead to strong inflammatory activation of microglia (and a severe depression-like syndrome) after a systemic microbial trigger. The NOD mouse also has a depressive-like behavior and prolonged HPA-axis response to immobilization stress accompanied by a higher expression of FKBP5 in the brain.

The OBX rat model (a model of local inflammation of the brain) and the rat colitis models (gut inflammatory models of systemic chronic inflammation) have also been developed. MRI found a reduced blood flow in the hippocampus of the OBX rat, while both systemic and brain signs of inflammation were found in the colitis models (raised iNOS expression and expression of inflammatory cytokines). In the colitis models also a higher TDO activity in the liver was detected explaining the reduced trytophan levels in the animals.

The chronic restraint stress animal models (both in normal and gestational rats) have yielded data supporting the concept that these models are reliable models of depression and particularly impact the expression of TDO in the liver (thus affecting the tryptophan break down pathways) and do not particularly impact the inflammatory state of the animals.

Objective 3. The exploration of in vitro monocyte/macrophage and microglia-astrocyte-neuron systems to study the basic mechanisms underlying the discovered immune activations in psychiatric patients.

In vitro monocyte culture systems did show that

1. Bacterial products (LPS), and the neurotransmitters adrenalin and ouabain elicit part of the gene expression signature in the cells. Ouabain and LPS induced the most complete and clear signatures, also involving the so-called cluster 2 genes (the adhesion/motility/chemotaxis genes). HCG, prolactin and oxytocin had no clear effects.
2. Over expression of PDE4B and ATF3 in the cells only leads to minimal changes, indicating the minor importance of these genes under steady state conditions.
3. An important series of experiments starts to elucidate the epigenetic regulation of the immune activation signature. Ouabain-induced-signature genes show strong overlap with a particular profile induced with a particular polycomb protein (a protein that epigenetically regulates gene transcription via chromatin remodeling). This polycomb protein (as well as particularly linked miRNA’s) is indeed induced by ouabain and present experiments suggest an involvement in the expression of the monocyte signature. Experiments are ongoing to unravel this type of inflammatory epigenetic regulation.
4. Triggering of the peripheral benzodiazepine receptor (PBR or TSPO) on monocytes leads to reduced production of pro-inflammatory cytokines and stimulates the monocytes to develop into dendritic cells.

The culture systems of microglia-astrocyte-neuron interactions yielded the following results:

1. Mouse microglia activated by LPS impairs neurite outgrowth in co-cultured neurons supporting the gene array data that immune activated microglia indeed neglects neuronal sprouting.
2. A microglial cell line BV-2 stimulated by various immune stimuli (e.g.. LPS, IFN-gamma) expresses the signature transcription factors ATF3, EGR3 and PU.1
3. The microglial cell line BV-2 stimulated by IFN-gamma expresses IDO and induces neurotoxic effects via the kynurenine pathway.
4. IFN-gamma stimulation induced in dorsal-raphe-slice-cultures IDO up regulation (but not in microglia) and serotonergic cell death, again suggesting neurotoxic effects via the kynurenine pathway
5. The important role of purine signaling in these interactions between activated microglia-astrocyte-neurons has been further illustrated in that

a. In both astrocytes and microglia the regulation of IL-6 and CCL2 release is mediated by P2X7 receptors.
b. Activation of microglial P2X7 receptors is implicated in the neuro-protection upon NMDA-induced neurotoxicity.

Objective 4. To study therapeutic correction possibilities using immune modulating medicines, such as anti-inflammatory drugs (e.g. COX-2 inhibitors) and to study the anti-inflammatory actions of regular psychiatric medications.

With regard to studies aiming at correcting the activation state of monocytes/microglia:

a) COX-2 inhibition had no effects on the behavior of the non-LPS stimulated NOD mouse (this regimen should now be tested in the LPS stimulated NOD mouse).
b) Imipramine treatment of the OBX rat attenuates local brain inflammation and improves behavior
c) Imipramine treatment is robust in the treatment of the GS rat model (unlike sertraline)
d) A trial with Cimicoxib has been finalized in patients (SECIM trial). Data show (near significant) beneficial effects in patients with a severe major depressive disorder.
e) The CIMICOXIB trial with Celecoxib is delayed but started in Germany (inclusion 32 patients). The IRBs of the other non-German centers have not been given (NL, IT, FBP) or have been given, but patients have not yet been included (IMU).

The immune model for mood disorders emerging from the Moodinflame studies

The concept of the project has been strengthened and refined by the combined study on animal models and patients. A new model for the pathogenesis of mood disorders can be constructed:

An abnormal immune activation set point of microglia in the brain (hippocampus) at critical stages in the development (early age) is emerging as central to the development of mood disorders. This abnormal immune activation set point of microglia impacts neuronal sprouting, axon guidance, synaptic function and toxicity via neglect of growth factors and/or production of toxic kynurenine products and/or via abnormal purine signaling (the latter often via astrocytes).

The abnormal immune activation set point of the microglia is probably due to an inherited (genetic) proneness of the microglia (and other cells of the monocyte/macrophage system) to hyper react to inflammatory stimuli and can be induced by a local (e.g. apoptosis, infection) or systemic trigger (e.g. infection, hormones, obesity). When there is a systemic trigger it can also be detected as an abnormal activation set point of the circulating monocytes (which also show the inherited proneness to inflammatory hyper reactivity). Alterations in the T cell system and higher levels of cytokines parallel these monocyte abnormalities. The systemic immune abnormalities are evident in large proportions of patients and are linked to certain symptom profiles (mania, melancholia, somatic signs and symptoms) and are also important determinants for treatment response (poor reaction to anti-depressants).

Exploitable fore ground

In the end this project has generated in particular a deepened knowledge on the multiple immune mechanisms which play a role in disease development in sub sets of mood disorder patients. Complex in-time-changing disturbed interactions between microglia-astrocytes-neurons become evident as cornerstones in the gradual development of mood disorders. Simple assays on peripheral blood cells and serum to detect these complex multiple and changing interactions cannot be developed and patented, since many immune genes and compounds are simultaneously changed.

Nevertheless it might be possible to identify in our rostrum of data genes and compounds of key importance in the prediction of disease development and therapy outcome. Particularly in the follow-up project PSYCH-AID we want to identify these molecules and file these for a patent. We think e.g. of monocyte JUN expression/percentages of CD8+T cells to predict responsiveness to anti-depressants and antibodies to tryptophan break-down products to be used in ELISA’s to detect abnormalities in these products in individuals at risk for mood disorders.

Project Results:

Main S & T results/foregrounds.

WP 1: ANIMAL MODELS OF MOOD DISORDERS

Objectives:

We will further develop the NOD mouse, the OBX rat and the GS rat as models of pro-inflammatory depressive-like behavior. More specifically the objectives are:

A To further test the NOD mouse via various behavioral tests for a depressive like behavior (as has extensively been done for the two rat models)
B. To investigate whether the genes of the pro-inflammatory PDE4B associated signature (as found in patients) are also aberrantly expressed in the monocytes/macrophages of the NOD mouse, the OBX rat and the GS rat and if not, to find the aberrant molecular signature of the pro-inflammatory monocytes and macrophages of these animal models.
C. To further investigate the tryptophan catabolic pathway in the animal models in relation to depressive-like behavior.
D. To further investigate the HPA-axis of the animal models

2. Achievements:

The NOD mouse model has been established as an important model of mood disorders analogous to the human situation of mood disorders in combination with local and systemic immune disturbances (EMC, AF, TCD):

Hyper-reactive behavior and development of depressive like behavior were found in the first and second year by AF in the NOD mouse model.

The behavioral experiments on the NOD mouse model have in later years been fully expanded by TCD:

1. When compared in steady state to the CD1 control strain, naive NOD mice display increased anxiety like behaviors in the elevated plus maze and the open field, as well as depressive-like behaviors in the tail suspension test. In the forced swim test the NOD mouse shows high activity on the exposure day (Day 1), but the highest increase in immobilization on the test day, again indicating an increased development of depressive-like behavior (particularly after stress).
2. When exposed to acute immobilization stress, the NOD mice display reduced locomotor activity and rearing in the home cage when compared to their non-stressed controls. No stress-induced change in home cage behavior was observed in the CD1 strain.

A). This altered behavior observed in the NOD mice was accompanied by a prolonged HPA axis response; while corticosterone levels returned to control levels in the CD1 mice at 2 hr post stress, they remained elevated in the NOD mice at this time point.
B) To determine whether these findings were associated with an impaired glucocorticoid feedback mechanism in the NOD mice the expression of the glucocorticoid receptor (GR) and the GR co-chaperone FKBP5 were analyzed in the hippocampus and hypothalamus. We found that while stress-induced expression of GR was comparable in both the CD1 and NOD mice, the expression of the GR co-chaperone protein FKBP5 was significantly increased in the NOD mice when compared to their CD1 counterparts.
C) The effect of stress on plasma glucose levels was also determined in the NOD and CD1 mice. It has been demonstrated that while CD1 mice show an elevated glucose response to stress, this response is absent in the NOD mice. This finding may be further indicative of an impaired glucocorticoid feedback in these mice.
D) NOD mice display an altered baseline and stress-induced 5-HT metabolism that may be related to their aberrant behavior profile.

3. In collaboration with Erasmus MC a study has been carried out to determine whether NOD mice show an altered behavioral response to an acute systemic challenge with LPS. Importantly in response to this challenge, NOD mice display an exaggerated and prolonged LPS induced sickness behavior in the home cage. In fact there was a severe collapse of the active behavior of the NOD mouse pre-challenge and locomotor activity and rearing were suppressed for a longer period. Since this observation has been done previously (Amrani te al, 1994, Bluthe et al, 1999) this strengthens the view that the NOD mouse is very sensitive in its behavior to an acute systemic inflammatory challenge.

Associated changes to central and peripheral cytokine expression in the NOD mouse were determined following systemic LPS challenge. Expression of IDO and a range of microglial markers and pro-inflammatory cytokines were determined in the hypothalamus. There was no differences observed between NOD and CD-1 mice for the inflammatory markers but there was a differential expression of IDO and the microglial markers CD40 and CD68.

4. The molecular gene expression signature in microglia and monocytes of the NOD mouse

EMC found microglia to be set at an abnormal immune set point (gene make-up pattern: CCR6, IL-7R, S100-8A, TNF, Prostaglandin-R raised), but what was striking that the most important gene networks that were suppressed were the proliferation networks of the microglia and the neuronal growth support networks. Indeed BRDU experiments revealed a reduced cell renewal capacity of the NOD microglia. Besides these signs of a defective function of NOD microglia under steady state, the cells were over reactive to LPS stimulation, and Affymetrix analyses show a strong up regulation of inflammatory genes, particularly of the IFN-induced pathway.

EMC found monocytes in the NOD mouse also to have another gene make up as control mice under steady state: The Affymetrix analysis showed over expression of COX-2, IL-1B, PDE4B, THBS and TREM-1, also in IL-10 -/- and Casp8 -/- NOD mice. Thus the pattern of monocyte activation under steady state is similar but certainly not identical to the patient situation. However after LPS stimulation also the peripheral monocytes were hyper reactive to the inflammatory stimulus and classical pro-inflammatory genes were raised.

The NOD experiments also tell us that the gene make-ups of the peripheral monocytes and microglia, even under inflammatory conditions, are not identical, though being both pro-inflammatory.

This means that perhaps also for the human conclusions drawn from the inflammatory state of monocytes cannot simply be transferred to the microglia.

5. To date no evidence in support of an excessive activation of the kynurenine pathway in NOD mice compared to the CD-1 strain has been obtained. LPS provokes a reduction in circulating tryptophan 4 hrs and an increase in circulating kynurenine concentrations 24 hrs post challenge in NOD mice but there was no difference in the response obtained when compared to the CD-1 strain. While increased expression of the kynurenine pathway enzymes IDO and KMO were found in the brain of the NOD mouse, no changes in cortical kynurenine concentrations were obtained in either strain following LPS administration.

The OBX rat model has been established as a model of mood disorders due to local brain inflammation without systemic immune alterations (TCD)

This model has been confirmed as a model of mood disorders, yet turned out to be a difficult to handle model. Nevertheless it could be established that the model is really a model of local brain inflammation: Inflammatory markers in the brain were found abnormally expressed, while circulating pro-inflammatory cytokines were normal, there was a normal production of cytokines from stimulated splenocytes/blood cells, and normal tryptophan and serotonin levels in brain, though there were some IDO/TDO abnormalities in liver and spleen

The further objective was to analyze brain structural changes in the OBX rat using magnetic resonance imaging (MRI) that may associate with the altered behavioral and inflammatory profile established.

A detailed MRI study was undertaken in the OBX rat with a view to developing neuroimaging markers associated with the model. T1 and T2 relaxometry was undertaken as these parameters are proposed to relate to the presence of inflammation in body tissue. The total ventricular and hippocampal volume was assessed through manual tracing of these regions along the coronal plane to form a three dimensional volume which was quantified using MIPAV (Medical image processing, Analysis and visualization). Regional volumes were normalized to total intracranial volume to calculate the relative percentage volume of each region of interest. Bolus tracking arterial spin labeling (bt-ASL) was employed to obtain regional perfusion parameters including regional cerebral blood flow and volume.

The OBX rat exhibits a significant increase in total ventricular volume without any change in hippocampal volume. The structural changes however were accompanied by a reduction in perfusion and blood volume (flow) in the hippocampus. Changes in brain structure or cerebral perfusion were not attenuated by chronic imipramine treatment.

The OBX rat also exhibits a decrease in T2 relaxation time in a number of cortical and limbic regions analysed with no change in T1 relaxation time. Regional changes in T2 relaxation times have previously been reported to be associated with the presence of inflammation. Reduced T2 relaxation times observed in the hippocampal region were attenuated following chronic imipramine treatment.

In collaboration with EMC, microglia and monocytes have been collected for microarray analysis of the inflammatory signature of the OBX rat.

In addition to the OBX rat model TCD has also studied 3 other rat models of immune-related depression: The colitis model, the chronic immobilization stress model and the Poly I:C-induced depression model.

Gestational stress (GS) rat model (Bradford)

To further characterize the responses of the rat dams to gestational stress, Bradford investigated as to whether or not chronic stress induced depressive-like behavior prior to parturition. Using the same stress paradigm as their usual approach (daily restraint for one hour) they measured immobility scores in rat dams in the forced swim test on day 17 (exposure day) and 18 (test day) of gestation. This assessment comes following restraint episodes which begin on gestational day 10 for all rats (except non-stressed controls). They found that 6-7 episodes of daily restraint prior to the forced swim exposure and test day led to increased immobility times of the stressed dams on the test day, but no obvious response on the exposure day. Thus, perhaps surprisingly, chronic restraint stress leads to depressive-like behavior in rat dams PRIOR to them giving birth, and precedes the appearance of the litters. It is unclear if this matches human data on postnatal depression, but some evidence would suggest that mothers suffering postnatal depression probably have some mood disturbances prior to their infants’ deliveries.

Bradford has included in their analyses studies on the maternal behavior of the control and stressed dams. Their initial findings indicate that gestationally-stressed (GS) dams exhibit heightened anxiety/vigilance such that they transfer their nests and pups around the housing cages a great deal more than non-stressed dams. With frequent transferring they also observed the GS dams building less complex and resourced nests (e.g. they tend to utilize less of the available nesting materials provided for them to build nests); in contrast to the control dams who consistently employ 100% of the available nesting materials. Other measures of maternal behavior are still undergoing analyses. Because the heightened nest transfer and rebuilding behaviors appeared to be anxiety-like behavior, we tested groups of rat dams in an elevated plus maze for anxiety. We found no compelling evidence that the GS dams were more anxious than non-stressed dams, and found no effect of either sertraline or imipramine on anxiety measures.

Michael Meaney has suggested that stress-induced changes in rats in general may relate to alterations in core body temperatures of the animals. If our GS dams had difficulties regulating body temperatures this could explain, in part, alterations in nesting behaviors. If the GS dams were overheated they would nest less and interact with their infants in a more irregular fashion. We have monitored nest temperatures of all our rat dams and find no difference in the ambient temperatures of the nests regardless of whether the dam is stressed or not. In agreement with that finding, our data on core body temperatures do not reveal any gestational-stress effect either. Thus we do not believe that changes in dam body temperatures or corresponding nest temperatures play any significant part in the altered maternal depressive-like behavior or nurturance that we observe in our studies.

In terms of physiological parameters, Bradford has examined neurogenesis in their dams and offspring since this has been reported to change with gestational stress. Their preliminary findings indicate that parts of the brain (hippocampus and ento-rhinal cortex) display reduced neurogenesis in the GS dams compared to control dams, while the opposite holds true for the offspring whose neurogenesis appears to be enhanced in frontal cortex and hippocampal areas of the stress-exposed pups.

The Bradford group collected serum and PBMCs from GS rats and pups and are sending it to Rotterdam.

Reached Objectives:

- We have developed the NOD mouse, the OBX rat and the GS rat as models of pro-inflammatory depressive-like behavior. More specifically:
- We have established in particular the NOD mouse via various behavioral tests as a new model of spontaneously depressive like behavior, excessively aggravated by LPS stimulation. The OBX and GS rat have been further and firmly established as models of depressive-like behavior.
- We have established that the genes of the pro-inflammatory (PDE4B associated) signature (as found in patients) are also aberrantly expressed in the monocytes of the NOD mouse, however not spontaneously, but only – and excessively –after LPS treatment. Monocytes of the non-LPS stimulated NOD mouse do however show intrinsic abnormalities making the cells vulnerable for high inflammatory responses to LPS. We thus found the NOD mouse as the most appropriate model for the monocyte activation in patients.
- We have established HPA-axis abnormalities in the NOD mouse model of depressive like behavior (particularly after acute immobilization stress).
- We have established that the OBX rat has a local inflammatory brain response and we have therefore not investigated the model for its circulating monocytes and peripheral immune stimulation. The OBX rat turned out to be the least appropriate model for the monocyte activation in patients.
- We have established abnormalities in the tryptophan catabolic pathway, in the NOD mouse model (though minor), but also (and in particular) in the colitis model and the chronic stress model. These models have been developed extra in the program.

Not yet reached objectives (reasons)

Although we clearly established the GS rat as a model of depressive like behavior, we have not yet been able to study the materials taken from the animals. We plan to work out this material in the coming years depending also on the outcomes of our further studies in the NOD mouse model and patients.

Future plans:

We continue the work in PSYCH-AID with the core groups of MOODINFLAME and specifically focus on a further exploration of the NOD mouse model, since the model resembles the syndrome we study in psychiatric patients characterized by a familial burden of autoimmunity and mood disorders the most. In particular we have started to study the relationship between the monocyte and microglia inflammatory state.

WP 2: INFLAMMATION AND IN VITRO BRAIN CELL CULTURE STUDIES

Objectives:

Effects of the activated inflammatory response system on synaptic function:

- Identification of the alteration of synaptic transmission upon pharmacological induction of neuro-inflammation in a mouse model.
- Identification of synaptic alterations in the NOD mouse model and the OBX rat.

Rescue of synaptic dysfunction in inflammatory mouse models of mood disorders.

Effects of the activated inflammatory response system/ the kynurenine pathway abnormalities on neurodegeneration

- To determine the role of kynurenine pathway activation in inflammation-related neurodegeneration
- To dissect the precise role of microglia and astrocytes in neurotoxicity induced by kynurenine pathway activation
- To determine if microglia prepared from NOD mice have increased inflammatory reactivity and neurodegenerative potential

The role of purinergic signaling and S100B in the activated inflammatory response system in the brain.

- Assessment of the expression of cytokines (PTX3, IL-1, IL-6, TNF-α), purine (adenosine, ATP) receptor subtypes and of S100B
- Assessment of the effects of purines on adult neurogenesis/neurotoxicity/ synaptic plasticity in brain culture systems

Achievements:

LPS-induced prenatal inflammation mouse model (INSERM/EMC).

This model as studied by INSERM/EMC bridges the gap between peripheral microbial inflammatory stimulation- monocyte activation – inflammatory microglia activation- abnormal hippocampus development/abnormal neuronal function – abnormal behaviour.

We showed in the first two years of the project that LPS (i.p.) injection in dams resulted in

1. Strong inflammatory monocyte activation, with up regulation of classical inflammatory genes such as IL1B, TNF, IL-6, PTGS2 (thus very similar to the patient situation)
2. Similar pro-inflammatory microglia activation in the developing fetal hippocampus. A technique was developed to obtain pure microglia.

The group size for the microarray experiments on microglia was increased in the 3rd year of MOODINFLAME. Data were re-analyzed by robust statistics and tested a new service of ingenuity called iReport. Aberrantly down-regulated pathways in the microglia of the fetal hippocampus after maternally induced inflammation include: axonal guidance, reeling and ephrin signaling. Up regulated are the immune pathways including IL-22 signaling, GM-CSF signaling; in addition we found a particular set of genes involved in microglia activation up-regulated. IL-16 is significantly up-regulated after inflammation. Embryonic hippocampal tissue for protein analysis has now been collected to validate the microarray data.

Inflammatory microglia of the fetal hippocampus induced in vitro:

a) Changes in synaptic transmission. During the previous period, we showed that inflammation rapidly increased the frequency of excitatory post-synaptic currents (EPSC) with no change in their amplitude. We have now identified the cellular and molecular factors of this microglia to neuron communication. We have shown that upon inflammation, microglia recruit astrocytes by purinergic signaling. Astrocytes then produce glutamate that will favor the presynaptic release of glutamate. Finally, INSERM showed that they could rescue the neuronal activity by selectively blocking this pathway using purinergic agonists.
b) Impaired outgrowth of neurites from cultured neurons. To identify the mechanism responsible for the inhibition of axonal growth after LPS-induced microglia activation INSERM further studied an in-vitro model. Using this in vitro culture system, INSERM confirmed that activated microglia actually prevent the growth of neurites. They further demonstrated that in agreement with the down regulation of trophic genes by microglia, prenatal inflammation impacts the development of axon bundles such as the corpus callosum in vivo. During the last period, INSERM demonstrated, in collaboration with EMC that mutation of microglial protein DAP12, which activate microglia induced the same alteration than prenatal inflammation.

This study has therefore established a link between a collection of microglial genes expressed during inflammation and neuronal dysfunction.

Inflammation, tryptophan breakdown and neurodegeneration using a microglial cell line (TCD)

Following on from the establishment of glial-neuron co-culture models at the start of the project, the objective of TCD in later years was to determine if kynurenine pathway activation plays a role in inflammation-related neuro-degeneration.

Results to date indicate that conditioned media from IFN-gamma treated BV-2 microglia induces neurotoxicity in primary cortical neurons. Expression of kynurenine pathway enzymes IDO, KMO, HAAO and kynureninase but not kynurenine aminotransferase (KAT II) is increased in the microglial BV-2 cells following IFN-gamma exposure. Pre-treatment of the BV-2 microglia with the IDO inhibitor 1-methyl-L-tryptophan and the KMO inhibitor Ro 61-8048 prevented IFNgamma-induced neurotoxicity and support a role of the kynurenine pathway in inflammation-induced neuro-degeneration.

The objectives for later years were also to determine whether the transcription factors PU.1 ATF3 and EGR3 are up regulated in activated BV-2 microglia. These transcription factors are of importance as they have been shown to be up-regulated in the monocytes of several psychiatric conditions including schizophrenia and bipolar disorder (these genes are important transcription factors of the activation signature). The expression of the transcription factors ATF3, PU.1 and EGR3 in activated BV-2 microglia was analysed using real-time PCR.

Results to date show with regard to

ATF3: Treatment with IFN-gamma, LPS, IFN-gamma and LPS in combination, IFN-alpha, TNF-alpha, IL-6 and IL-1beta all induced significant increases in ATF3 mRNA in BV-2 microglia.

PU.1: Treatment with IFN-gamma, LPS, IFN-gamma and LPS in combination, TNF-alpha, and IL-6 induced significant increases in PU.1 mRNA in BV-2 microglia.

EGR3: Treatment with IFN-gamma induces significant increases in EGR3 mRNA in BV-2 microglia.

Following on from the initial characterisation of the expression of these transcription factors in BV-2 cells in response to pro-inflammatory cytokines the objective was to determine the effect of anti-depressants, anti-psychotics and lithium chloride. Characterisation of the BV-2 cell line showed that these cells express the serotonin transporter (SERT) and MAO-A but no other synthetic, degradation enzymes, or transporters for the mono-aminergic neurotransmitter systems. In addition the following results were obtained:

(1) Inhibition of the serotonin transporter (SERT, with fluoxetine) and MAO-A (with moclobemide) inhibits the response of microglia to LPS, measured by iNOS mRNA expression and TNF-alpha production, classical markers of microglial activation.
(2) Preliminary data shows when BV-2 microglia are incubated with serotonin (5-HT), they have the ability to take up and degrade 5-HT, indicated by a reduction in 5-HT concentration in the culture medium and an increase in 5-HIAA. When cells are treated with IFN-gamma or LPS, their ability to metabolise serotonin is reduced significantly.
(3) Pre-treatment with 1 µM of fluoxetine, citalopram, moclobemide, clorgyline and lithium chloride had no effect on the induction of ATF3, EGR3 and PU.1 by IFN-alpha or IFN-gamma
(4) Pre-treatment with 1 µM of haloperidol and clozapine had no effect on the induction of ATF3 and PU.1 by IFN-alpha or IFN-gamma, but did potentiate the induction of EGR3 mRNA by IFN-alpha and IFN-gamma

Inflammation, serotonergic cell death and IDO expression in dorsal raphe slices (IMU)

Inflammation of the serotonergic system in the dorsal raphe nucleus (doR) is considered to play an important role in the pathophysiology of depression. L-tryptophan can be metabolized to the neurotransmitter serotonin (5-HT), but also to kynurenine and its metabolites, by the enzyme indoleamine-2,3-dioxygenase (IDO). IDO is the first rate limiting enzyme of the pathway and it is highly inducible by the cytokine interferon gamma (IFNgamma).

The aim of the study of Innsbruck was to investigate doR serotonergic neurons and IDO expression in doR organotypic brain slices. The effect of pro-inflammatory mediators, especially IFNgamma and neurotrophic factors on death or survival of serotonergic neurons and on IDO expression was revealed. Their data show a functional serotonergic system in the organotypic brain slice, where 5-HT and its metabolite 5-hydroxyindole acetic acid (5-HIAA) were detected. They found a significant decrease of serotonergic neurons and enhanced IDO expression after inflammation in the doR as well as IDO co-localized with neurons and to a lesser extent with serotonergic neurons, but not with microglia or astrocytes. In conclusion their data show that inflammation in the doR leads to serotonergic cell death and IDO up regulation (not in microglia), which might play a role in depression.

IMU in addition studied whether beta-amyloid induces inflammation of serotonergic raphes neurons.

Beta-amyloid (Abeta) is a small peptide that plays a potent role in synaptic plasticity, but also forms amyloid plaques in Alzheimer’s disease (AD). Recent studies suggest that Abeta deposition is deleterious not only in AD, but also in depression. This Abeta effect is associated with inflammatory processes. However, further evaluation is needed to understand how Abeta and inflammation interact and contribute to the regulation of the cholinergic, serotonergic, and dopaminergic neuronal populations. The aim of the study of IMU was to investigate the effects of Abeta1-42 on cholinergic neurons of the nucleus basalis of Meynert (which degenerate in AD), on serotonergic neurons of the dorsal raphe nucleus (which play a role in depression) and on dopaminergic neurons of the ventral mesencephalon (which degenerate in Parkinson’s disease) in rat organotypic brain slices. Furthermore, IMU aimed at investigating whether anti-inflammatory drugs (celecoxib, citalopram, cyclooxygenase-2-inhibitor, ibuprofen, indomethacin, piclamilast) modulate or counteract Abeta-induced effects. Two week-old organotypic brain slices of the nucleus basalis of Meynert, dorsal raphe nucleus and ventral mesencephalon were incubated with 50 ng/ml Abeta1-42 with or without anti-inflammatory agents for three days. The results reveal that Abeta significantly decreased the number of choline acetyltransferase-positive cholinergic, tryptophan hydroxylase-positive serotonergic and tyrosine hydroxylase-positive dopaminergic neurons. Anti-inflammatory drugs partially counteracted the Abeta-induced neuronal decline. This decline was not due to apoptotic processes (as evaluated by TUNEL, propidium iodide, caspase) or oxidative stress (as measured by nitrite, catalase or superoxiddismutase-2) or inflammation, but most likely caused by a down regulation of mentioned key enzymes.

Inflammation and purine signaling (Freiburg)

In order to study neuronal-glia interactions several culture systems have been set-up in Freiburg. Pure astrocytes, microglia and neuronal cultures are up and running. Moreover, organotypic slice cultures are used to analyse glial reactions, neuronal survival and neurogenesis. Methods to deplete and replenish microglia in these slices were established and are typically used in excitotoxicity experiments.

Purinergic regulation of cytokines in cultured astrocytes and microglia:

Freiburg found that ATP dose-dependently induced the production of cytokine IL-6 and CCL2, but not TNF-alpha in primary mouse astrocytes. However, the specific P2X7 receptor agonist BzATP induced not only IL-6, CCL2 but also TNF-alpha production in astrocytes. The effects of ATP and BzATP in astrocytes were inhibited by non-selective P2 receptor antagonist suramin, suggesting the potential role of P2X7 in cytokine release in astrocytes.

In purified microglia, only BzATP, but not ATP, induced the robust production of cytokines while P2 receptor antagonist suramin inhibited the BzATP effects. Interestingly, the release of IL-6 and CCL2, but not of TNF-alpha, were inhibited by P2X7 receptor antagonists BBG and A438079. These results suggested that P2X7 receptor regulates IL-6 and CCL2 in cultured microglia and that TNF-alpha release is regulated differently. To explain this difference Freiburg first investigated the potential involvement of other P2 receptors such as P2X4, P2Y2/4 and P2Y6. However, none of the other P2 receptor agonists (2-MeSATP, UTP and UDP) with respective specificity induced cytokine release in cultured microglia. Co-stimulation with the non-selective P1 receptor agonist NECA inhibited the release of IL-6 and CCL2 evoked by BzATP in microglia, but not the BzATP-induced synthesis of TNF-alpha.

In order to further clarify the receptor that is involved in TNF-alpha release in cultured microglia knockout animals were used. Interestingly, the effects of ATP and BzATP in the release of IL-6, CCL2 and TNF-alpha were completely abolished in microglia cultured from P2X7 receptor knockouts, clearly showing the vital importance of this receptor for the control of cytokine release. No effect was observed in microglia lacking P2X4. These results would exclude that the unusual pharmacology that we observed for the release of TNF-alpha is due to receptor dimerization of P2X7 and P2X4. We therefore speculated that CCL2 and IL-6 were released via a different mechanism compared to TNF-alpha release in response to P2X7 activation. Therefore experiments with an inhibitor of pannexin-1 channels (that is opened after P2X7 stimulation) were performed. These experiments revealed that pannexin-1 is not required for ATP and BzATP induced cytokine release in cultured microglia.

Thus, whereas the pharmacological data indicate that in both astrocytes and microglia the regulation of IL-6 and CCL2 release is most likely mediated by P2X7 receptors, it is not yet known how purine signaling causes the release of TNF-alpha. Although P2X7 is also required for TNF-alpha release most likely other signaling molecules are in addition needed.

Organotypic hippocampal slice cultures (OHSCs): Activation of microglial P2X7 purine receptors is vital for neuroprotection upon NMDA-induced neurotoxicity

Pre-treatment of OHSCs with the P1 adenosine receptor agonist N-ethylcarboxamidoadenosine (NECA) protected neurons from NMDA-induced cell death in a concentration-dependent manner. This effect was blocked by the A1-receptor specific antagonist Cyclopentyl-1,3-dipropylxanthine (DPCPX) indicating it is mediated by A1 receptors. Using the macrophage toxin clodronate, we generated OHSCs devoid of microglia. These cultures survive normally but show increased susceptibility to NMDA-induced neurotoxicity. Importantly, activation of A1 receptors by NECA was neuroprotective also in the absence of microglia. To examine the effects of purines on P2 receptors we used ATP as treatment for OHSCs. In the presence of microglia 500 µM ATP significantly reduced NMDA-induced neuronal cell death. However, in the absence of microglia the neuroprotective effect of ATP was completely abolished.

In conclusion, these data support the notion that adenosine A1 receptors possess neuroprotective functions. Since depletion of microglia had no impact on the neuroprotective effect of A1 receptor activation, we conclude that receptors expressed on neurons and/or astrocytes account for this neuroprotection. In contrast, the protective effects of ATP were lost upon depletion of microglia, arguing for the importance of this cell type for the effect to occur. To examine a potential role of TNFalpha in the ATP-mediated neuroprotection, OHSCs were treated with a TNFalpha-neutralizing antibody to eliminate basal as well as stimulation-induced levels of the cytokine. Under these conditions, neuronal cell death induced by 25 µM NMDA could not be reduced anymore by pre-treatment with 500 µM ATP. In slice cultures prepared from TNFalpha deficient mice, neurotoxicity upon treatment with 25 µM NMDA was much more pronounced as compared to wild type OHSCs, and no neuroprotection was observed upon treatment with ATP or BzATP. These data show that TNFalpha is involved in the neuroprotective effects of purines acting on microglia.

Reached Objectives:

Effects of the activated inflammatory response system on synaptic function:

- We have clearly identified alterations of synaptic transmission, microglial activation and microglial deficiencies to provide neuronal and axonal regeneration factors upon LPS stimulation in the maternal neuro-inflammation mouse model.

Effects of the activated inflammatory response system/the kynurenine pathway abnormalities on neurodegeneration

- We have identified toxic kynurenine pathway activation in microglia culture systems upon IFN-γ stimulation, this was not the case in brain-organ slice culture systems. Also in this latter culture system astrocytes played no role, neurons were mainly involved in tryptophan metabolism.
- In WP1 we have already described that NOD microglia is spontaneously deficient in neuronal and axonal regeneration factors and responds to LPS treatment in a peculiar way: With a strong IFN- driven inflammatory response. This led to an even stronger deficiency in neuronal and axonal regeneration factors

The role of purine signaling and S100B in the activated inflammatory response system in the brain.

- In these studies co-culture systems have been used of astrocytes, microglia and neurons, as well as brain-organ slice cultures. We have shown a role of adenosine and the microglial P2X7 receptor in the production of CCL2, IL-6 and TNFα by microglia and astrocytes. The P2X7 receptor induced production by microglia was vital in preventing NMDA induced neurotoxicity in the culture systems, again underlining that microglia is not only involved in neurotoxicity, but also in neuroprotection and axonal/neuronal growth.

Not yet reached objectives (reasons)

- We have not applied above described culture systems (cell culture systems, brain-organ culture systems) fully on the NOD mouse. Also the role of S100B has not been studied in the brain cell culture systems. Time restraints are the reason for this, as well as the necessity to first fully identify the behavioral and immune abnormalities in the NOD mouse model.

Future plans:

We will continue the work in PSYCH-AID with the core groups of MOODINFLAME and in the Dutch-funded NWO projects and will specifically focus on microglial culture systems, also from the NOD mouse model.

WP 3: PATIENT SELECTION, CHARACTERIZATION AND BANKING

1. Objectives:

To select and clinically characterize patients with a mood disorder and to select individuals at risk for development of a mood disorder, all to be used for monocyte signature testing and tryptophan catabolism as well as for recruiting patients for the clinical trial (Work package number XI). For this purpose we will establish serum, leukocyte and DNA banks. The patients will also be used for the nested studies indicated in work packages 6, 7 and 11.

2. Achievements:

Number of collected patients and samples: See Table. The programs of immune tests carried out on these patient samples (See Table in attached PDF document):

Program A: Gene and miRNA expression in monocytes, full lymphocyte subset determination, stimulated in vitro production of IL-12 and TNF by monocytes, DEX sensitivity of this stimulated production, and the measurement of 10 plasma cytokines/growth factors

Program B: Due to reduced quality of the collected samples from some centres only full lymphocyte subset determination and the measurement of 10 plasma cytokines/growth factors could be measured

Program C: Due to reduced quantities of material KCL has developed its own program for measurement of a few genes in monocyte preparations for IFN-alpha treated patients.

In sum: We have been collecting the number of samples as expected, in fact more patients and their respective samples have been collected as envisaged (only IFN-treated patients stay somewhat behind), biobanking is adequate, as is the redistribution of samples to the partners (see scheme).

WWU has monitored in the past years established procedures for the training of clinical raters and the certification of new raters for all clinical centres.

WWU has also collected in addition to the above listed and envisaged groups of patients

4 MDD patients after ECT treatment
24 MDD patients that were recruited a second time (follow-up study)
16 MDD patients of the atypical subtype
10 MDD patients of the melancholic subtype

The construction of a web interface database (CRL)

This web based database has been constructed, trained, and is being filled in by the various partners and is largely complete.

Diagnostic Biomarker Platform:

- Clinical upload portal COMPLETED 2011
- Assay upload information analysis COMPLETED 2012
- Spotfire – production database connection COMPLETED 2012/2013
- Spotfire analysis pipeline COMPLETED 2012/2013

Database:

- Production database COMPLETED 2011
- Adjusted for laboratory data COMPLETED 2012

Visualisations:

- Web application integrated basic info COMPLETED 2011
- Spotfire visualizations client portal COMPLETED 2012 / 2013
- Spotfire web visualization / collaborators portal COMPLETED 2012 / 2013
- MoodInFlame website COMPLETED 2011

http://www.moodinflame.eu

At present the database with clinical and assay related data is accessible by all collaborators through the usage of Spotfire, the built in security layer supports the data integrity and makes sure collaborators only have access to their own data or data from collaborators who agreed to share their data. The Spotfire tool is connected to the statistical package R making it feasible to run complex data analysis immediately on the production database, the results are visualized via Spotfire.

Reached Objectives:

We have – as planned - selected and clinically characterized large groups of patients with a mood disorder and we have selected individuals at risk for development of a mood disorder, all to be used for monocyte signature testing and tryptophan catabolism. For this purpose we have established serum, leukocyte and DNA banks. We have a large database under care of CRL.

Not yet reached objectives (reasons)

We have insufficiently recruited patients for the clinical trial (see WP11, see there for more detail).

Future plans:

We continue the work in PSYCH-AID with the core groups of MOODINFLAME and specifically focus on selecting enough patients for the clinical trial (see WP11) and also collecting more MDD patients via Partner UM.

WP 4 AND 8: INFLAMMATION, ACTIVATED MONOCYTES, ACTIVATED T CELLS, INFLAMMATORY CYTOKINES, AUTOIMMUNITY AND HLA STATUS

1. Objectives:

WP4: To generate/validate and register high throughput systems (automized MACS combined with custom-made mRNA arrays) using minimal quantities of blood to determine abnormal molecular signatures in monocytes of patients with a major mood disorder and individuals at risk to develop a major mood disorder.

WP8: To know the relation of mRNA signature positivity (or its individual molecules), concomitant autoimmune phenomena and known autoimmune related HLA-DR/Q polymorphisms.

2. Achievements:

Immune parameter analysis in particular the development of an assay measuring monocyte activation via gene expression (EMC).

1. Due to additional Affymetrix analyses on monocytes of patients with the various mood disorders we have expanded our panel of monocyte signature genes to a number of 70-80 genes with 4 reference genes. A quicker, but far more expensive array system (TLDA) has been introduced by EMC with success for the Q-PCR of the 70-80 monocyte signature genes, as well as a more rapid custom made statistical program specifically developed for TLDA (Data Assist software). Abl turned out to be the best reference gene and at present EMC calculates all its values in relation to abl with healthy control values run in the same assay, which are set to 1. This gives us reliable and consistent data in fold change expression over that of healthy control monocytes. EMC will nevertheless validate all values against the entire panel of reference genes as well as carrying out other ways of calculation. This will be possible since all raw data of the Q-PCR (the original ct values) are stored in the CRL database.
2. Lymphocyte subsets and T cell subset numbers are evaluated with an advanced multicolor FACS system enabling the measurements of many CD markers in small quantities of sample. In this way EMC enumerates in the circulation of patients monocytes, NK cells, B cells, T cells, activated T cells (CD3+CD25+T cells), T cytotoxic cells, T-helper cells (and within this population IFN-gamma+Th1, IL-4+Th2, IL-17A+Th17 CD4+ lymphocytes), and natural T regulatory cells (CD4+CD25highFoxP3+).
3. Cytokines are being measured with a multi-analyte system, i.e. the cytometric bead array, enabling the measurement of many cytokines in a small sample of serum. Other multi-analyte systems, such as the Raybiotech array, have been tested, but appeared not reliable. The cytokines presently measured by multi-analyte are IL-1beta, IL-6, TNFalpha, CCL2, VEGF, IL-8.

A detailed analysis of peripheral immune activation (monocyte gene signature with TLDA, FACS analysis of lymphocytes and cytokines) had been carried out at EMC for the various collected groups of

a. Patients with bipolar disorder: i) UMCU, Twin study; ii)EMC/PSYQ; iii)UMCG/KU-Leuven
b. Children of a bipolar parent (UMCG/UU)
c. Women with postpartum psychosis (EMC)
d. Patients with MDD: i)EMC; ii) WWU; ii)KU-Leuven

The BD/MDD cohorts of Muenchen, FBP and HSR will be tested from June until December 2013.

Conclusion of monocyte, T cell and cytokine studies per June 2013 using above described assay systems, in particular TLDA: Strong monocyte inflammatory immune activation was found in bipolar disorder (particularly those with mania and appetite disturbances), in melancholic MDD (not in atypical MDD), in the acute phase of postpartum psychosis and in not yet psychiatrically affected children of a bipolar parent, particularly at teenage time (16/17 yrs).

This immune activation was composed of higher expression of the selected signature genes in monocytes (which include both pro- and anti-inflammatory genes arranged in largely two clusters of genes), altered numbers of NK cells, B cells, T regulatory and T effector cells, as well as variable increases in serum cytokines, such as PTX3, CCL2, IL-1beta, IL-6, IL-8 and sCD25. Patterns of these immune aberrancies were different for the different psychiatric conditions, their activity state and for different points in the natural histories of the diseases. Underneath a synopsis can be found for the outcomes thus far.

Ad a: Bipolar disorder

EMC detected in the study of the initial cohort of UMCU for monocytes a high inflammatory set point in bipolar disorder (BD). The on-going study visualized in the recent year the following interesting findings in this cohort of BD patients:

(1) A relation between pro-inflammatory gene expression and manic symptomatology (increased motor activity and speech disorganization),
(2) A pro-inflammatory gene expression was found particularly in patients with an earlier age of onset (< 20 years) and a longer duration of disease, and
(3) Support for the concept of an immune suppressive action of some of the mood regulating medications, in particular lithium, anti-psychotics and anti-epileptics.

EMC detected for T cells in this cohort of BD patients:

a) That percentages of anti-inflammatory CD4+CD25highFoxP3+ regulatory T cells were higher, the latter only in BD patients < 40 years of age. Percentages of Th1, Th2 and Th17 cells were normal.
b) The monocyte pro-inflammatory state and the raised percentages of CD4+CD25highFoxP3+ regulatory T cells occurred independently from each other.
c) In BD patients positive for thyroid autoimmune disease (AITD) a significantly reduced percentage of CD4+CD25highFoxP3+ regulatory T cells was found as compared to BD patients without AITD.

EMC detected for the cytokines only an over expression in serum for IL-1beta and PTX3, serum levels for IL-6, TNFalpha and CCL2 were normal.

Conclusion for BD:

The data show

1. That monocyte and T-cell networks are activated in a proportion of BD patients,
2. That these immune networks are activated independently from each other,
3. That the activation of the immune networks involve both pro- and anti-inflammatory forces,
4. That the activation of the monocyte system is linked to an earlier age of onset of BD (< 20 years) and more severe disease,
5. That T cell activation, particularly of T regulator cells, probably occurs particularly in younger patients and that the lack of such anti-inflammatory regulatory T cell forces in BD patients is a determinant in the development of AITD in the patients.

At present the BD cohorts of EMC-PSYQ and UMCG are worked out (most of the assays have been done) and data will be analyzed in a similar fashion to enable studies on similarities and dissimilarities with the initial cohort UMCU cohort.

Ad b: Children of a bipolar parent

Offspring (n=140) of a bipolar (BD) parent were examined in 1997-1999 (Time 1, mean age 16 years), 132 children were available for reassessment after 14 months (Time 2, mean age 17 years), after another 41 months (Time 3, mean age 21 years) and at present we have also completed the 4th collection (Time 4, mean age 28 years), including the parents (see WP3). At Time 3, 30% of the offspring had developed a major mood disorder, at Time point 4 this figure has risen to 54%, of which 13% is a bipolar spectrum disorder (only 3% met DSM-IV criteria for bipolar I disorder). Overall, 72% of the bipolar offspring developed a lifetime DSM-IV axis I disorder. In 88% of the offspring with a bipolar spectrum disorder, the illness started with a depressive episode. In total, 24% of offspring with a unipolar mood disorder developed a bipolar spectrum disorder over time. Mood disorders were often recurrent (31%), complex (comorbidity rate, 67%), and started before the age of 25.

Aim: To determine immune activation (the monocyte, T cell arm and some cytokines) in the offspring at Times 1 and 2 (teenage), Time 3 (young adulthood) and Time 4 (adulthood). Of Time 1 only serum is available.

Results:

1) Monocyte gene fingerprinting has been done in all children over the time of follow-up. Offspring were strongly positive for the fingerprint, particularly at teenage. The pattern of the fingerprint changed over the years of follow up: at young adulthood (21 years) there was a weaker pro-inflammatory state of monocytes; at adulthood (28 years) the monocyte gene activation had vanished and was normal as in healthy controls. Noteworthy is that the monocyte activation at teenage and young adulthood was irrespective of later development of a mood disorder or the actual presence of a mood disorder, 45% of life time euthymic offspring showed a monocyte inflammatory gene fingerprint at those ages.

Serum PTX3 and CCL2 (both monocyte fingerprint genes) levels were higher at all measured time points (also at the later age, when the circulating monocytes did not show an activation anymore), but levels of these cytokines did not predict later development of a major mood disorder.

2) FACS evaluations of the circulating lymphocytes showed at all the time points studied a reduced number of CD3+CD4+ T cells, while there were variable numbers of T regulatory, Th1, Th2 and Th17 cells over time.
3) We previously reported a higher prevalence of AITD in children of a bipolar parent, irrespective of the presence of mood symptoms. We found that the lack of anti-inflammatory regulatory T cell forces in BD patients is important for a development of AITD (see above). The immune evaluation in offspring developing or having an AITD shows a similar picture as in bipolar patients with AITD: The difference in children with and without autoimmune thyroid disease lays entirely in the level of the T regulatory cells: Reduced T regulatory cell levels were found in the offspring with an AITD, i.e. with positive TPO-Abs.

Conclusion for offspring of a BD parent:

It is clear that a strong pro-inflammatory activation of circulating monocytes is present in the offspring of a parent with BD, irrespective of the development of a mood disorder. This monocyte activation precedes in time the onset of mood disorders and shows a changing pattern over time with teenage time showing the strongest activation and adulthood the weakest (in fact monocyte activation is at the level of healthy controls, yet CCL2 levels are raised). A reduction in T cell numbers can be found at all the time points. The lack of T regulatory T cell forces is correlated to the development of thyroid autoimmunity in the cohort.

Ad 3: Postpartum psychosis

Postpartum psychosis is a life-threatening psychiatric emergency, which often occurs without significant premorbid symptoms. Although many studies have postulated an involvement of the immune and endocrine systems in the onset of postpartum psychosis, the specific aetiological factors have remained unknown.

EMC examined the hypothesis that autoimmune thyroid dysfunction is associated with the onset of postpartum psychosis and have found that women with postpartum psychosis are at higher risk not only of developing thyroid autoimmunity but also of clinical thyroid failure. These data implicate thyroid function as an important clinical outcome in patients with postpartum psychosis and warrant screening for TPO-Abs in patients with postpartum psychosis.

Since the postpartum period is a period of immune activation, EMC investigated whether an up regulation of monocyte inflammatory genes and of T cells is part of the physiological immune activation postpartum and whether such up regulation is more pronounced in patients with postpartum psychosis. EMC found that the physiological immune activation postpartum comprised predominantly an up regulation of the T cells system and to a much lesser extent of the monocyte system.

Compared to healthy non-postpartum women, monocyte gene expression of women with postpartum psychosis (PP) were significantly elevated 4 weeks postpartum: Among the postpartum women, PP patients had a significant up-regulation of monocyte genes (in the two typical clusters as known from bipolar disorder patients) not otherwise elevated in the normal postpartum period. Further, the GR-beta/alpha gene expression ratio was increased in monocytes of PP patients, strongly correlating with their immune activation. Remarkably however, PP patients had significantly reduced levels of total T-cells and total T helper cells, as well as Th1 and Th17 subsets, compared to healthy postpartum controls. With regards to cytokines only CCL2 (and to some extent also IL-6) were raised.

Conclusion: The EMC study thus demonstrates a robust deregulation of the immuno-neuro-endocrine set point in PP, with a notable over-activation of the monocyte/macrophage arm of the immune system 4 weeks postpartum. At present data 6 months postpartum are analysed as are data 1 year postpartum. Preliminary data show a gradual normalization over time.

Ad 4: Major Depressive Disorder (MDD)

Currently, there are no accurate predictors of response to antidepressants in MDD, and successful treatment relies greatly on ‘trial and error’. The choice of what drug to prescribe for optimal response is largely a matter of taking an educated guess. Obviously, this “trial and error” strategy contributes to serious treatment delay as clinical response can only be reliably evaluated after 4-6 weeks of treatment. Therefore there is an urgent need for valid predictors of effective treatment response.

EMC treatment trial

EMC has carried out a trial in which melancholic MDD patients participated in a double blind, randomized trial of venlafaxine and imipramine. Material of 47 HC and of 49 relatively old melancholic MDD patients (median age 54 years) has been studied before (after 1 week washout period, Time point T1) and after 7 weeks of treatment with venlafaxine (n=24) or imipramine (n=25) in each group (T2). The rate of responders and non-responders for the venlafaxine arm is 13/24 (52%), that of the imipramine arm is 7/25 (28%).

Outcomes of all melancholic MDD patients before therapy

Monocytes showed a high immune activation set point. This set point was composed of the two sub-sets of strongly correlating genes as described above with only minor differences.

Numbers of circulating T cells were reduced in melancholic MDD patients in particular the Th1, Th17 and natural T regulator cells.

With regard to cytokines: CCL2, IL-1beta, IL-6 and IL-8 were raised, while TNFalpha was normal.

Conclusion: Older melancholic MDD patients show a strong monocyte inflammatory activation with a reduced T inflammatory activation. Natural T regulator cells are reduced as well.

Prediction of treatment outcome:

EMC and CRL have evaluated the “monocytes signatures” in the responders and non-responders to venlafaxine and imipramine and they have found that:

1. Patients who responded to venlafaxine/imipramine had a low monocyte inflammatory activation, a reduced number of CD8+ cytotoxic T cells and a raised number of NK cells and actually turned to a “normal” non-activated monocyte activation state after treatment (CD8+ T cells and NK cells stayed abnormal)
2. Patients who did not respond to venlafaxine/imipramine had the opposite and had a significantly increased monocyte inflammatory gene expression and the highest CD8+ T cells and lowest NK cells. The non-responders did virtually not change their immune state during therapy.
3. A formula has been developed in which monocyte JUN expression can predict non-responsiveness to venlafaxine/imipramine with sensitivities and specificities between 70-80%, showing proof of principle of the strength of our immune approach for clinical outcome prediction.
5. The raised cytokine levels (CCL2, IL-1beta, IL-6 and IL-8) were not predictive, we are not informed on the effect of treatment on the cytokine levels.

KCL treatment trial

KCL has performed a similar trial as EMC also showing that the inflammatory state of depressed patients predicts treatment outcome:

• Non-responders patients to 8 weeks of treatment with escitalopram or nortriptyline had higher baseline mRNA levels of IL-1beta, MIF and TNF-alpha in PBMC.
• Antidepressants reduced the levels of IL-1beta and MIF, and increased the levels of the GR and p11, but these changes were not associated with treatment response.
• Successful antidepressant response was associated with a reduction in the levels of IL-6 and of FKBP5, and with an increase in the levels of BDNF and VEGF-A
• Patients with old-age depression and hypothalamus-pituitary-adrenal (HPA) axis hyperactivity showed a 23% increase in SGK1 expression.

Muenster MDD cohort

The Muenster cohort studied for above described immune parameters comprises at present 56 patients who are considerably younger than the EMC cohort (median age 28 yrs) and only counts 4 patients with melancholia.

Monocyte gene activation was very modest in the cohort and was highest in the older patients and those with feelings of guilt, insomnia and somatic gastro-intestinal complaints (all signs and symptoms reminiscent of melancholia).

With regard to FACS analysis preliminary data show that Th2 cells are increased in non-melancholic depression. NK cells are reduced. Further analyses are carried out.

Conclusion MMD:

Measurement of the inflammatory state as performed in MOODINFLAME shows strong activation of the monocyte system particularly at older age and in melancholic patients. In younger non-melancholic patients such monocyte activations are minor, while lymphocyte abnormalities predominate: Increases are found in Th2 cells, while NK cells are reduced.

The assessment of the inflammatory state in melancholic older MDD patients opens new avenues for the prediction of treatment response to classical anti-depressive drugs.

Nested studies on immune activation and various co-morbidities of mood disorders

IMU (Innsbruck) has set up patient and material collections in the last two years of the project to study immune activation in targeted patient groups with various comorbidities, such as suicidal behavior, bone metabolism and cognitive vulnerability. The material is of sufficient quality to perform FACS analyses and cytokine analyses, but the quality is insufficient to perform gene expression studies in monocytes:

Interferon-α Study (KCL)

A total of 70 patients were recruited and completed at least one follow-up assessment during IFN-α treatment. Saliva cortisol samples have been analyzed, PUFA levels have been measured in plasma for 48 patients at all the time points of the study, kynurenine and tryptophan pathway metabolites have been measured in plasma for 38 patients at 3 study time points. Data mining of all these variables is in progress.

Material has been sampled for Moodinflame cellular immune analysis, however qualities and quantities have been insufficient to determine monocyte activation. KCL has extracted some mRNA from PBMC and has conducted microarray gene expression analyses in 48 patients at 3 study time points.

Reached Objectives:

WP4: We have generated/validated high throughput systems (Automized MACS combined with custom-made mRNA arrays, FACS analysis systems and multi-analyte serum analysis systems) using minimal quantities of blood to determine abnormal molecular and immune signatures in leukocytes and serum of patients with a major mood disorder and individuals at risk to develop a major mood disorder.

WP8: We have established that mood disorder patients with thyroid autoimmunity have a normal (= not inflammatory) activated monocyte profile, but considerably reduced percentages of natural T regulator cells, showing the tolerance failures in these mood disorder patients with thyroid autoimmunity.

Not yet reached objectives (reasons)

At present we have not registered any of the systems yet for a patent. Reason is the complexity and costs to register multi-compound/analyte systems for the indicated broad range of disorders. We plan to register patents for single genes/leukocyte populations/serum compounds for very specific indications within the frame work of PSYCH-AID.

We have not yet measured the HLA-DR/Q profiles in the mood disorder patients with a concomitant thyroid autoimmunity (see also WP 6).

Future plans:

We will continue the work in PSYCH-AID with the core groups of MOODINFLAME and will specifically focus on finding single genes/leukocyte populations/serum compounds for very specific indications to be patented.

WP 5: INFLAMMATION AND TRYPTOPHAN CATABOLIC PATHWAY DETERMINATIONS

1. Objectives:
1. To validate the tryptophan catabolic pathway biomarkers using standardized sensitive high performance liquid chromatography (HPLC) analyses on the patient serum samples of the multicentre longitudinal collection for early diagnostic and prognostic use as described in WP 3.
2. To develop the diagnostic reagents for the detection of tryptophan catabolic pathway biomarkers in human blood,
3. To generate and validate a simple and fast assay test kit for a primary-health-care level as well as for specialized hospitals.
2. Achievements:

Establishment of an external quality assurance system (EQAS) for standardization of HLPC analyses of KP metabolites

APD has organized and led the External Quality Assurance Scheme to standardize the laboratory analyses of tryptophan metabolites and to develop the diagnostic tools for analyses of tryptophan metabolites.

APD has organized up to 10th EQAS rounds within 6 centres (LMU, UIAntwerp, TCD, UMCG, UNSW and UIllinois). Spiked human serum has also been included in the EQAS samples. Technology transfer between the centres has been realized.

TCD is another important partner to participate in the EQAS consortium. New laboratory protocols have been introduced by TCD for the determination of KP metabolites by LC-MS in collaboration with LMU. Lower limits of detection are possible with LC-MS over HPLC which should enable the detection of lower concentrations of metabolites in plasma and tissue samples. GCMS has also been set up to further improve analysis of the KP metabolites.

Kynurenine measurement in patients

Based on the already validated extraction method LMU has optimized the efficiency of the solid phase extraction in order to increase the sensitivity of the analysis. They established a method, which they have called ‘dual load SPE’, characterized by repeated loading of the sample under different pH conditions onto the extraction column. This method development resulted in an increase of the sensitivity by 237% for 3-HAA, 221% for TRP, 172% for KYN, 99% for 3-HK, 27% for KYNA and 21% for 5-HIAA. The coefficients of variation were 2% for TRP and KYN, 3% for 3-HK, 3-HAA and KYNA, and 6% for 5-HIAA. LMU therefore successfully optimized the extraction method for the analysis by HPLC.

A further aim was the development of an analytical method for the simultaneous measurement of TRP, 5-hydroxy-TRP, 5-HT, 5-HIAA, melatonin, 6-hydroxymelatonin, KYN, 3-HK, KYNA, AA, 3-HAA, quinaldic acid, xanthurenic acid, picolinic acid, and quinolinic acid using an ultra-HPLC (Acquity, Waters) coupled with tandem mass spectrometry (Xevo TQ, Waters). The optimization of the chromatographic conditions (analytical column, injection volume, temperature, mobile phases, flow rate, gradient shape) resulted in a total run time of 10 min. Finally the optimal MS conditions were tested and fixed with a source temperature of 150°C, a desolvation temperature of 650°C and a desolvation flow of 1000 l/hr. Also the extraction method has been optimized.

After these optimizations the actual patient sample analysis has been carried out from end 2011 till spring 2013.

All redistributed samples (see Table on page 11) collected by the universities of Münster, Leuven, Innsbruck, München and of the interferon study of King’s College, London have been analyzed for TRP, Kynurenine, 3-hydroxykynurenine, kynurenic acid and 3-hydroxyanthranillic acid using HPLC. Samples from postpartum psychosis study in Rotterdam were analyzed for all metabolites using HPLC and UPLC/MS.

The scientific analysis of the samples in relation to immune activation is presently in progress.

Antibody development for ELISA’s

Regarding the development of diagnostic tools, APD has established the mouse monoclonal antibodies against tryptophan, kynurenic acid and quinolinic acid. The characterisation of the antibodies in ELISA setting showed promising results and detailed characterization using immunohistochemistry and western blot are in progress. The purification and detailed characterization of previously developed polyclonal antibodies against tryptophan, kynurenine, kynurenic acid and quinolinic acid are in progress.

The development of ELISAs using polyclonal antibodies are in progress. For measurement of kynurenic acid, the lowest detection limit reached to 10 ng/ml and for quinolinic acid, 100 ng/ml. It still needs to be improved to reach lower detection level.

Morphological studies on human post-mortem brain tissues using the antibody against quinolate has been carried out and published. These studies show quinolate to be over expressed in the microglia of suicide victims.

Reached Objectives:

- We have measured various tryptophan catabolic pathway biomarkers using standardized sensitive HPLC analyses on patient serum samples as collected in WP 3.
- We have developed antibodies for the detection of the tryptophan catabolic pathway biomarkers Quin and KynA in human blood,

Not yet reached objectives (reasons)

We have not yet been able to apply our progress in this field to generate and validate a simple and fast assay test kit for a primary-health-care level as well as for specialized hospitals for tryptophan metabolites.

Future plans:

We continue the work in PSYCH-AID with the core groups of MOODINFLAME and specifically focus on the validation of the tryptophan metabolites in relation to the abnormal immune parameters in the studied patient groups and on the generation and validation of a simple and fast assay test kit for a primary-health-care level as well as for specialized hospitals for tryptophan metabolites.

WP 6: ACTIVATED MONOCYTES, TRYPTOPHAN CATABOLISM, HPA-AXIS AND GENES

1. Objectives:

In nested-case control studies on mood disorder patients we will study interrelations between the presence of the pro-inflammatory signature/ biochemical signs of an abnormal tryptophan catabolism and HPA-axis abnormalities and relevant gene polymorphisms

Regarding HPA-axis abnormalities: To determine the relationship between the outcomes of the dexamethason-suppression test (DST) and the presence of the aberrant monocyte signature, the aberrant tryptophan metabolism and abnormal levels of pro-inflammatory IL-1, IL-6, TNF, IL-10 and PTX3 in the circulation

Regarding gene polymorphisms: To determine the genetic polymorphisms that underlie the activated inflammatory response system in patients by studying the association between specific alleles or haplotypes in functional candidate genes (PDE4B, IL-6, COX-2, CCL2, IL-10, TPH II, IDO, TDO, KAT II, KMO) and the presence of the aberrant monocyte signature, an aberrant tryptophan metabolism and abnormal levels of pro-inflammatory IL-1, IL-6, TNF and PTX3 in the circulation.

2. Achievements:

a. Regarding HPA-axis abnormalities

Assessment of dexamethasone-suppression test (DST) cortisol levels in patients and control individuals (KU-Leuven)

In accordance with the DOW, 8 saliva samples, before and after dexamethasone administration, were collected from all patients for assaying cortisol levels. These samples were sent to KU Leuven for cortisol level determination.

In the BD cohort, in total 2464 samples from 308 patients have been sent to KU Leuven. In 204 patients (1632 samples), saliva cortisol levels have been analysed. Most patients in this group were included at the UMCG and KU-Leuven. Additionally, saliva samples from 104 patients from the UMC Utrecht twin sample are awaiting analysis of cortisol levels.

In the MDD group, Leuven has received 1640 saliva samples from 205 Patients (68 from KU-Leuven, 96 from Münster, 41 from München). Of these samples, 768 have been analysed for cortisol levels.

In the control group finally, Leuven has received 3080 saliva samples, stemming from 385 individuals. In 196 individuals (1568 samples), cortisol levels have been determined.

KU-Leuven has set up a joint analysis plan together with the UMCG and EMC to analyse the DST data in BD patients and controls, and correlate those to the immune function data. Mrs. E. Gellens set up the primary analysis plan, and in January 2013, all data were transferred to the UMCG (prof. A. Schoevers) to prepare for a joint publication.

Further, KU Leuven is planning a similar analysis in MDD patients in a joint analysis together with WWU Münster University.

GRα/β gene expression in monocytes (EMC, KULeuven)

As part of the immune activation signature GRalpha/beta gene expression in patient monocytes have been performed using the new TLDA systems (see WP 4).

Decreases in the monocyte GRalpha/beta expression ratio have been found in all cases where a pro-inflammatory monocyte signature was detected (children of a bipolar parent, PPP, melancholic MDD and acute SCZ), mainly due to an increase in the gene expression of the inhibitory GRbeta receptor (but in MDD also due to a decreased GRalpha expression). The decreases in the GRalpha/beta ratio strongly correlated to the higher expression of many of the activation genes in the monocytes. These gene data suggest that steroid resistance at the level of the monocytes is part of the immune activation signature.

Saliva sample taking for cortisol, analysis of cortisol data, consequences of non-suppressor state (cortisol resistance) for immune activation (KU-Leuven).

In order to determine the relationship between the outcomes of the DST and the presence of the aberrant monocyte signature in UMCG BD patients, the aberrant tryptophan metabolism and abnormal levels of pro-inflammatory IL-1, IL-6, TNF, IL-10 and PTX3 in the circulation, K.U. Leuven has provided EMC with a list of 20 strong suppressors and 20 clear non-suppressors on the DST. In these BD cases the monocyte signature has been established.

Preliminary data show that there is no correlation between the GRalpha/beta ratio, the inflammatory monocyte signature and the suppressor/non-suppressor status of the BD patients as measured via the HPA-axis.

In a similar study we found also no correlation between the GRalpha/beta ratio and the inflammatory monocyte signature in BD cases of EMC-PSYQ and the hair cortisol measurements (a newly developed measurement to establish hyper-cortisolism over a longer period of time)

To further explore the relationship between hypercortisolism, steroid resistance, immune function EMC has developed further functional assays to be carried out on the collected patient PBMCs (see WP10).

b. Regarding gene polymorphisms

In agreement with all Moodinflame partners, EMC has sent cell samples of all the patients to KU-Leuven, from which DNA has been extracted and banked.

As stated in the mid-term report, there will be less emphasis on studies on polymorphisms in the signature inflammatory genes (and more on epigenetics).

Association between kynurenine pathway and genotypes

SNP’s were selected from around 20 genes that were associated in the literature with HPA-axis abnormalities, the activated inflammatory response system, the ‘Moodinflame monocyte signature’ and the kynurenine pathway.

APD has assisted LMU and KULeuven to investigate the association between kynurenine pathway metabolites and related genes. In collaboration with KULeuven and APD and external collaborator St. George Hospital, London, LMU has analyzed the kynurenines from IFN-α treated patients and KULeuven/APD has analyzed the related genotypes. Analysis is in progress.

HLA-DR genotypes (FBP)

FBP took the lead in the analysis of HLA A, B, and DR/DQ genes of patients and correlating these with the other variables studied. FBP participated for that reason in the selection of patients with BD and MDD throughout the period. The final patient groups collected by FBP comprised 52 BD and 64 unipolar MDD Bulgarian patients (116 patients in total). The healthy Bulgarian control group reaches 44 subjects.

HLA-A,-B typing was performed by PCR-SSP (PCR-Sequence Specific Primers) method using Protrans kits. HLA-DRB1, -DQB1 typing was performed PCR-SBT (PCR-Sequenced Based Typing) method using Abbott kits. Allele and haplotype frequencies were estimated by Expectation Maximization algorithm (Program Arlequin v1.1). Comparisons between groups were done by Chi-square test. Several alleles have reached statistical significant differences between patient and control groups and between the two patient groups as well. The results suggest associations of HLA alleles, mainly class II, with both patient groups. HLA haplotype analysis clarifies further the role of HLA genes in the pathogenesis of affective disorders.

Besides the Bulgarian population group, the analysis of HLA subtypes was performed also for two large Dutch groups of patients and controls. Analysis awaits. Correlations of HLA subtypes and some clinical data were also performed. Correlations with other biological data still need to be performed.

MBL2 genotypes

Additional tests were performed on the two Bulgarian patients groups and the control groups for MBL2 genotyping using microsphere-based methodology. MBL analysis shows decreased frequencies of haplotypes associated with high expression in both patient groups and predominance of haplotypes determining low and intermediate expression.

All the Bulgarian patients and controls have full range of scales performed as well as blood collected, necessary cells and liquids isolated and frozen. Two shipments with frozen test-tubes were sent to Rotterdam. The material is of sufficient quality to perform FACS analyses and cytokine analyses, but the quality is insufficient to perform gene expression studies in monocytes. These assays will be performed in the coming year and data correlated to HLA-DR genotypes and MBL genotypes.

Epigenetics (KULeuven,EMC)

Analysis of methylation patterns in patients with mood-disorders/history of trauma (KU-Leuven/EMC).

During the last year, KU-Leuven has built up the expertise to measure methylation patterns at several promoter sites of the glucocorticoid receptor (GR) gene, using bisulfite conversion followed by analysis using Sequenom EpiTYPER technology. This method provides a sensitive way to measure even low levels of methylation with high reliability (Izzi B et al. A new approach to imprinting mutation detection in GNAS by Sequenom EpiTYPER system (Clin Chim Acta. 2010 Dec 14;411(23-24):2033-9).

Our primary objective is to investigate the methylation patterns in the CpG-rich region upstream of the human glucocorticoid receptor gene NR3C1, but other genes will be included in the study as well. This will allow relating methylation patterns to other features studied in Moodinflame, such as early trauma, pro-inflammatory signature, HPA-axis function and glucocorticoid receptor function.

In the last year, KU Leuven has developed assays to assess methylation status of the NR3C1 B, D and F promoter regions using sequenom MassARRAY methylation analysis. Recently, KU Leuven has published a first study using this methodology (Hompes et al., 2013). In a pilot study, methylation status has been evaluated in a number of MDD patients from the KU Leuven Moodinflame cohort (publication in preparation). Further, KU Leuven has set up a joint analysis plan with the University of Rotterdam (dr. V. Bergink) in order to assess methylation levels in women with postpartum psychosis. Subsequently, KU Leuven is establishing a joint analysis plan with the Institute of Psychiatry (Dr. L. Carvalho, dr. C. Pariante) in order to assess NR3C1 promoter methylation patterns in additional mood disorder patient groups. In these samples, GR promoter methylation in relation to clinical data, but also to GRalpha/GRbeta ratio’s and to immunological data is currently being studied.

Reached Objectives:

Objectives

Regarding HPA-axis abnormalities: We have measured all parameters to determine the relationship between the outcomes of the DST and the presence of the aberrant monocyte signature, the aberrant tryptophan metabolism and abnormal levels of pro-inflammatory IL-1, IL-6, TNF, IL-10 and PTX3 in the circulation. We are in the process of evaluating the data. In addition we have determined the GRα/β ratio in monocytes to be correlated to the DST and the inflammatory state of monocytes.

Regarding gene polymorphisms: We have determined genetic polymorphisms in HLA, the enzymes involved in the kynurenin pathway and in Mannose Binding Lectin (MBL) in groups of MDD patients. Analysis is in progress.

Not yet reached objectives (reasons)

We have not measured all originally indicated polymorphisms yet, although material is there to perform these analyses. We have decided to wait with these determinations to see which outcomes are relevant in the other WPs to update our list of relevant molecules to determine polymorphisms in (also related to the state of the art in the literature at that moment). In any case the consortium found it more relevant right from the start to focus more on epigenetics of the GR and so we have done.

Future plans:

We will continue the work in PSYCH-AID with the core groups of MOODINFLAME and specifically focus on relevant polymorphisms (HLA-DR, MBL) in compounds thought (via our research in MOODINFLAME) to play key roles in the abnormal set point of the innate immune system in mood disorders, as well as in the GR metabolism.

WP 7: INFLAMMATION, ACTIVATED MONOCYTES/MICROGLIA, ACTIVATED ASTROCYTES AND BRAIN SCANS

1. Objectives:

In nested-case control studies on mood disorder patients we will study interrelations between the presence of the pro-inflammatory signature/ biochemical signs of an abnormal tryptophan catabolism and HPA-axis abnormalities, relevant gene polymorphisms, astrocyte activation and outcomes of the brain scan

Regarding astrocyte activation: To determine the relationship between the serum level of S100B and BDNF and the presence of the aberrant monocyte signature, an aberrant tryptophan metabolism and abnormal levels of pro-inflammatory IL-1, IL-6, TNF and PTX3 in the circulation.

Regarding brain scans: 1. Clarify whether brain inflammation (as measured with MRSI or PET) occurs in mood disorder patients particularly in those with the aberrant monocyte signature and/or other signs of an activated peripheral IRS.

2. To correlate changes in BOLD fMRI neural activity of anterior cingulate cortex before/after a course of antidepressant treatment with changes in the activated IRS in MDD patients and with changes in the psychopathological conditions of patients.

2. Achievements:

a. Regarding neuronal growth activity and astrocyte activation S100B and BDNF measurements (WWU)

• A standardized procedure for S100B and BDNF analysis was set-up and validated.
• WWU prepared analysis plans together with other partners.
• WWU analyzed more than 1000 samples of patients and control subjects.
• WWU is presently calculating correlations between S100B/BDNF levels and the pro-inflammatory state of patient monocytes of MDD cases of Muenster, EMC, Innsbruck, the children of a BD parent (UMCU) and the BD cases of Milano.

Preliminary data show:

• For the children of a BD parent that they have reduced BDNF at teenage (also when not affected), but raised levels at adulthood. With regard to S100B only the kids who do not develop a mood disorder have reduced S100B at teenage, at adulthood all children have a raised S100B.
• For the melancholic MDD cases of EMC S100B correlated to treatment outcome and there was a strong correlation of S100B with the IL-1 cluster of genes in circulating monocytes.

Needless to say more evaluations now need to be done. They will be within PSYCH-AID.

b. Regarding brain scans MRS and BOLD fMRI (HSR)

1) HSR collected the whole requested sample (n=30/30) of well characterized MDD patients each one of them underwent a magnetic resonance Spectroscopic imaging (MRS) to explore changes of glutamate-glutamine.
2) HSR collected a second sample of MDD patients (n=20/20), each patient underwent a blood oxygen level dependent functional magnetic resonance imaging (BOLD fMRI) to explore changes in brain activity following a moral valence decision task and induced by a standardized protocol of sleep deprivation and light therapy. BOLD fMRI scans were performed before and after chronotherapeutic treatment.
3) HSR has collected also a sample of healthy controls (n=30/30), each patient underwent a blood oxygen level dependent functional magnetic resonance imaging (BOLD fMRI) to explore brain activity following a moral valence decision task and a magnetic resonance Spectroscopic imaging (MRS) to explore changes of glutamate-glutamine.
4) All the patients recruited at point 2) underwent also a T1 sequence to explore brain volumetric changes and a Diffusion Tensor Imaging sequence to explore connections among brain regions and integrity of white matter.
5) All patients and controls socio-demographical, clinical data were collected; also all questionnaire requested by the project were collected and inserted in a dedicated database.
6) Blood of all patients and controls participating to the study was collected, managed, and stored following the operating instruction and sent to EMC for analysis and redistribution. Blood of patients that underwent a chronotherapeutic treatment was collected twice, before and after therapy.

Unfortunately the material is of insufficient quality to perform gene expression studies in monocytes, but it is of sufficient quality to perform FACS analyses and cytokine analyses. The assays will be performed in the second half of 2013 and analysed thereafter.

c. PET (UMCG)

UMCG performs PET scans for activated microglia using a ligand for the peripheral benzodiazepine receptor (PBR), which is part of the immune activation signature in patient monocytes and a known activating factor of microglia. At July 2013 15 BD patients and 13 HC have been included at UMC Groningen. Blood of all patients and controls participating to the study was collected, managed, and stored following the operating instruction and has recently been sent to EMC for analysis and redistribution. The actual laboratory determinations and analyses will be performed in the fall of 2013.

Reached objectives:

Regarding astrocyte activation: We have measured the serum level of S100B and BDNF in the MDD groups of Muenster and Rotterdam and in children at risk for mood disorders and are in the process of determining the relationship between the levels of these compounds and the presence of the aberrant monocyte signature, an aberrant tryptophan metabolism and abnormal levels of pro-inflammatory cytokines.

Regarding brain scans: 1. We have measured brain inflammation via MRSI or PET in mood disorder patients and are correlating the outcomes with an aberrant monocyte signature and signs of an activated peripheral IRS, such as abnormal leukocyte populations as measured in FACS and abnormal cytokine levels.

2.We are in the process of measuring the parameters to correlate changes in BOLD fMRI neural activity of anterior cingulate cortex before/after a course of antidepressant treatment with changes in the activated IRS in MDD patients and with changes in the psychopathological conditions of patients.

Not yet reached objectives (reasons)

None

Future plans:

We continue the work in PSYCH-AID with the core groups of MOODINFLAME and the Dutch-funded NOW-TOP project specifically focusing on PET analysis of TSPO and peripheral parameters of an activated IRS.

WP 8: MOLECULAR DETERMINANTS OF ACTIVATED MONOCYTES

1. Objectives:

To know the consequences of the expression of the abnormal inflammatory mRNA signature in patient monocytes for their inflammatory state

1. Using siRNA knock down interventions
2. Using correlation studies between the signature expression and the actual pro-inflammatory functions of the monocytes.

2. Achievements:

Since it turned out to be impossible to culture ex vivo purified monocytes while stopping their further differentiation into MØ and DC, EMC, Humanitas and UU have decided (in a session of the in vitro culture group in Humanitas Milano, 2010)

a) To primarily use monocyte cell lines stimulated with inflammatory agents or transfected with EGR3/PU.1 as surrogate patient monocytes,
b) To use whole blood cell cultures, which behave more constantly or
c) To take the approach to culture from the monocytes macrophages (MØ) or dendritic cells (DC) via standard techniques (M-CSF and GM-CSF/IL-4 cultures respectively) in the assumption that from signature positive monocytes MØ and DC will arise with a different gene expression pattern as from healthy monocytes (to be tested)

Such cultures have now been used for the planned studies in this WP.

Moreover ATF3, EGR3 and PU.1 were found in correlation studies and ChIP to be important key regulating transcription factors for sub-cluster 1A and 1B gene expression in monocytes of patients and PDE4B expression became more and more of lesser importance as a key regulating factor in monocyte activation. EMC, Humanitas and UU have therefore also decided to put less emphasis on PDE4B experiments and more on experiments involving the transcription factors ATF3, EGR3 and PU.1.

Ad a: Induction of inflammatory signatures in human monocyte cell lines.

Gonadotropic hormones (EMC)

Since the studies in PPP patients (see WP4) delivered that the postpartum state is in particular a period in which monocyte activation takes place (in particular of cluster 1 genes), EMC has embarked on studies in which the monocyte cell lines MM6 and THP-1 are stimulated with graded dosages of HCG, Prolactin or Oxytocin to study the pro-inflammatory effect of these hormones on monocytes. It turned that these hormones had minimal effects, if any.

Ouabain (UU)

Previous bioinformatics analysis suggested digitalis-like compounds would induce a similar gene-expression signature as found in monocytes from patients with mood disorders. Ouabain was selected as a candidate to use for mimicking the disease-associated gene expression in normal monocytes.

Integrating Ouabain gene-expression data and publicly available ChIP-Seq data showed an overlap between Ouabain-induced genes and (1) Polycomb repressor complex (PRC) silenced genes, and (2) histone demethylase KDM6B/JMJD3 target genes. Notably we find that the “signature” genes PTGS2, EGR3, MAPK, STX1A, PDE4B, IL6, TNFAIP3, CCL2, CCL7, CXCL2, CDC42, DUSP2, ATF3, and ADORA2A are targets of KDM6B. Expression of micro-RNAs was also affected by Ouabain and the up regulation of two miR´s targeting subunits of the Polycomb repressor complex 2 (PRC2) was observed (miR-101 targeting EZH2 and miR-200 targeting Suz12).

Epigenetic changes in human monocytes treated with Ouabain (UU)

Our expression data above suggests a model in which changes in Polycomb repression complex could epigenetically influence the level of inflammatory gene expression, and long-term sensitize monocytes for the expression of these genes. Chromatin immunoprecipitation (ChIP) on normal monocytes treated with Ouabain show a reduction in the Polycomb repressive mark (H3K27me3) at promoters of inflammatory genes, supporting this model. Current work is focused on exploring the genome-wide epigenetic changes in response to pro-inflammatory signals by analyzing histone modifications (H3K27me3, H3K4me3) using ChIP-Sequencing, and DNA-methylation using illumina bead-chip technology.

Generation of cell lines expressing signature genes and si-RNA knock down studies. (UU)

Investigating the functional role of “signature” genes for the inflammatory state of monocytes is proceeding along two lines of investigation. First, the generation of human monocyte cell lines (over)-expressing signature genes. Second, siRNA/shRNA knock-down of signature gene expression. Stable cell lines expressing FLAG-tagged PDE4B and ATF3 have been generated. Constitutive expression of PDE4B leads to increased expression of DUSP2, which is reduced by siRNA against PDE4B. Constitutive expression of ATF3 showed no apparent effect on expression of the genes tested. Testing with pro-inflammatory stimulation to analyze the effects of ATF3 on expression of signature genes is on-going. To study the role of the Polycomb repressor complex on “signature” gene expression we have also generated a monocyte cell line with inducible shRNA for knock-down of the methyl transferase subunit (EZH2) expression that leads to a reduction of the repressive H3K27me3 histone modification.

Triggering of the PBR on monocytes has an anti-inflammatory effect and induces the cells to develop into dendritic cells.

The peripheral benzodiazepine receptor (PBR) is part of the immune activation signature in patient monocytes and a known activation factor of microglia and used in the PET scans of UMCG. It has therefore been studied by Humanitas in its cell systems. High-affinity PBR ligands such as, Ro5-4864 and Diazepam were tested for their ability to modulate the functions of blood monocytes purified from healthy donors.

Previous studies have postulated a role for PBRs in the regulation of immune functions. In particular it was shown that PBR ligands can regulate neutrophil migration and pro-inflammatory cytokine production. Benzodiazepines are normally used in the treatment of mood disorders and anxiety syndromes, however, their impact on the monocyte inflammatory state is still unknown.

Resting monocytes expressed PBRs, as analyzed by RT-PCR, and this expression was modestly regulated by the exposure to pro-inflammatory stimuli, such as LPS, TNFalpha and IL-1, with no effect of the anti-inflammatory agonists, such as IL-10, IL-13 and dexamethasone. Regulation of PBR was also found in human macrophage by low density microarrays. In these experiments, PBR was found negatively regulated in M1 macrophage polarization (LPS+IFNgamma), whereas, M2 macrophages (IL-4) did not show any regulation of PBRs.

Monocytes stimulated with 1-100 µM Ro5-4864 showed an impaired ability to release IL-1beta and TNFalpha in response to LPS. Similarly Diazepam inhibited the LPS-induced secretion of TNFalpha and IL-10. This inhibition could be related to the decreased expression of PBR. The effect of Diazepam was also tested on dendritic cells. These cells expressed PBR, by RT-PCR, and no modulation of this receptor was observed during the process of monocyte differentiation into dendritic cells. However, these experiments revealed that Diazepam was apparently able to promote dendritic cell differentiation as evidenced by the increased mean fluorescence intensity of the dendritic cell marker CD1a.

As expected M1 macrophages express PDE4, a prototypical M1 marker, and activin A, a pro-inflammatory cytokine recently found to be associated to the pro-inflammatory M1 status of mononuclear phagocytes. Stimulation of human DC by PBR agonists induced a marked decrease of TNFalpha, IL-1beta and activin A production.

These data outline a possible role of PBR in the differentiation of antigen presenting cells and thus on the ability of the immune system to initiate immune responses. In addition, these data indicate that benzodiazepines and activin A have a reciprocal negative regulation and that benzodiazepines can influence the pro-inflammatory state of monocytes by inhibiting the release of pro-inflammatory cytokines.

MiRNA studies (EMC)

This study was started with focusing at first on PPP, since many flanking data in this group are already available. miRNAs were isolated from monocytes using the mirVana miRNA isolation kit and the TaqMan Array Human MicroRNA A Cards v2.0 from Applied Biosystems, containing 384 TaqMan MicroRNA assays enabling analysis of 377 human microRNA’s were used in a finding study to detect miRNA’s abnormally expressed in monocytes of PPP patients. Data were analyzed against monocytes of healthy postpartum women (CP). Nine miRNAs were selected and tested using individual Q-PCR in a larger validation study on monocytes of 20 PP patients, 20 CP and 20 healthy non-postpartum women (HC).

In the validation study miR-146a expression was significantly down-regulated in the monocytes of first onset PP patients as compared to CP and HC; miR-212 expression was significantly down-regulated in PP patients with prior bipolar disorder. In silico miR-146a targeted 4 genes of the previously described monocyte activation signature in bipolar disorder; miR-212 targeted 2 of such genes.

This study identified changes in miR-146a and -212 expression in PP. Since these miRNAs are linked to inflammation, the study strengthens the view that PP is an inflammation-like condition

Reached Objectives:

- We have mimicked the PDE4B related inflammatory signature with ouabaine (and also to a lesser extent with LPS). We found ouabaine to induce epigenetic changes, in particular changes in a histone demethylase which targets many of the inflammatory signature molecules.
- We have over expressed PDE4B in monocytes and found an up regulation of DUSP2. We found ATF3, EGR3 and PU.1 as important transcription factors for the inflammatory signature, yet over expression of ATF3 did not lead to alterations in the expression of signature genes.
- We found in particular miRNA-146a to be abnormally expressed in mood disorder patients, strengthening the view that inflammatory alterations are important phenomena in these patients.
- We found the TSPO molecule (important in PET scanning) to be involved in the inflammatory state of monocytes and the differentiation of monocytes into dendritic cells.

Not yet reached objectives (reasons)

None

Future plans:

We continue the work in PSYCH-AID and the Dutch-funded NWO- TOP grant with the core groups of MOODINFLAME and specifically focus on the role of TSPO in the differentiation of microglia cells towards (anti-)inflammatory type microglia.

WP 9: THERAPEUTICS IN ANIMAL MODELS

Objectives:

To investigate the effect of the PDE4 inhibitor piclamilast on behavior, inflammatory and tryptophan catabolic pathway changes in animal models of depression.

To investigate the effect of Cimicoxib (COX2 inhibitor of SME3) on behavior, inflammatory and tryptophan catabolic pathway changes in animal models of depression.

To investigate the effect of tryptophan catabolic pathway enzyme inhibitors (particularly of the KMO inhibitor NW-1349 of SME4) on behavior, inflammatory and tryptophan catabolic pathway changes in animal models of depression.

Achievements:

In the midterm review we have decided to put less emphasis on PDE4 and its inhibitors on our studies, but put more emphasis on general anti-inflammatory agents and the anti-inflammatory effects of regular psychiatric drugs, since PDE4 is not the prime key molecule (probably the transcription factors ATF3, EGR3 and PU.1 (found in MOODINFLAME, see WP8) are equally important. But also of importance is that PDE4 inhibitors are in first instances targeted by the pharmaceutical companies as drugs against COPD; the pharmaceutical companies thus hardly support any other work in other disorders(most notably in psychiatry)

Effects of Imipramine treatment on behavior, inflammatory and tryptophan catabolic pathway changes in the OBX-rat model (TCD)

TCD investigated the effects of the tricyclic antidepressant imipramine on behavior, inflammatory and tryptophan catabolic pathway changes in the OBX rat model of depression

1. Chronic imipramine treatment attenuates OBX hyperactivity in the open field.
2. Chronic imipramine treatment also attenuates the elevated GFAP and IL-1beta expression previously observed in the OBX rat.
3. The effect of imipramine treatment on structural and anatomical changes in the brain of the OBX rat is currently being assessed using MRI.

Effects of treatment with NOS inhibitors on behavior, inflammatory and tryptophan catabolic pathway changes in the OBX-rat model (TCD)

In addition to testing imipramine in the OBX rat, inhibitors of nitric oxide synthase were assessed in the OBX rat for novel antidepressant activity. The highlights are

1. L-NA has antidepressant properties in the OBX rat model of depression
2. L-NA attenuated OBX related changes in ventricular perfusion and T2 relaxation times.

Effects of Cimicoxib (COX2 inhibitor of SME3) treatment on behavior, inflammatory and tryptophan catabolic pathway changes in the NOD mouse model (AF/TCD/EMC)

AF started the NOD mouse studies in year 1 and 2, but TCD has taken over these studies, since TCD was better equipped to carry out the studies (see WP1 and 2). AF has only studied the NOD mouse under home-cage conditions and noted the hyper-activity of the mouse; this hyperactivity was not influenced by Cimicoxib, the COX-2 inhibitor of AF. Tissue samples of NOD mice treated with the Cox-2 inhibitor over a time period of 4 weeks have been sent by AF to partner EMC for RNA profiling and to partner TCD for determination of tryptophan catabolic pathway changes. A summary of the respective behavioral data, including the files with the original raw behavioral data, was delivered to the colleagues of EMC/TCD for supporting their future studies in the NOD mice involving stress exposure and exposure to LPS (in the latter instance the NOD mouse shows serious abnormal depressive-like behavior).

In addition, animal pills containing the Cox-2 inhibitor Cimicoxib were produced according to requests of partner EMC/TCD. The pills were delivered to EMC/TCD and will be used for a long-term treatment study in NOD mice in order to study the influence of a Cox-2 inhibitor on monocyte gene expression signature, cytokines and serotonin-tryptophan-kynurenin metabolites in various biological samples particularly in the LPS-exposed NOD mouse. Depending on resources available, an intervention study using the antidepressants imipramine and fluoxetine will be carried out in the NOD mice to determine whether treatment can attenuate the aberrant behavior and potentially altered inflammatory response system observed in this model in response to stress and an acute systemic challenge.

Effects of sertraline and imipramine treatment on behavior, inflammatory and tryptophan catabolic pathway changes in the GS rat model (Bradford)

Bradford has examined the effects of sertraline and imipramine on maternal depressive-like behavior and have found that both drugs have a tendency to reduce immobility times in the postnatal forced swim test (which were conducted on postnatal days 4 and 5). Responses to sertraline are quite variable however, while the responses to imipramine appear to be more robust and reliable. These data fit with Bradford’s previous examination of fluoxetine which similarly reduces maternal depression following gestational stress. In all instances, the 3 drugs have been administered each day following delivery of the litters; it may be useful to determine if there is a better response of the mothers to drug treatment that starts during the pregnancy itself, particularly given the evidence Bradford has shown (see WP1) that depressive-like disturbances are present prior to delivery. In agreement with the forced swim test findings above, imipramine, and to a lesser extent sertraline, reduce the nest transposition/building behavior.

The remit of Bradford has been to further characterize the rat gestational stress model of depression, and provide testing of the COX2 inhibitors, Bradford has stock animals prepared for this purpose.

NW-1349 for animal and in vitro use (Newron)

The Kynurenine 3-OHase (KMO) inhibitor NW-1349 was planned to be tested in the NOD mouse, OBX rat and in the GS mouse model of depression. Among proprietary KMO inhibitors, NW-1349 has been selected as one of suitable molecules to be used in vivo animal models. It shows a good in vitro inhibitory activity (IC50 = 484 nM) and has a favorable develop ability profile, such as no cytotoxicity and no interference with cytochrome P450 system. A part of the compound (500mg) was delivered to Dr. Livia Carvalho (Institute of Psychiatry, KCL) who has used it for in vitro testing. The remaining amount of NW-1349 (1.6gr) is available for in vivo and in vitro studies as planned.

Reached Objectives:

- We have investigated the effect of Cimicoxib (COX2 inhibitor of SME3) on behavior in the home-cage NOD mouse model of anxious behavior. Effects were not noted.
- We have investigated the effects of regular anti-depressants in the OBX-rat model of local inflammation and have found beneficial behavioral effects and indications for an anti-inflammatory effect.
- We have investigated the effects of regular anti-depressants in the GS model and found beneficial behavioral effects; material has been collected to study for inflammatory markers.

Not yet reached objectives (reasons)

- We have not yet tested Cimicoxib on the LPS stimulated behavior of the NOD mouse (see future plans)
- We have not investigated the effect of the PDE4 inhibitor piclamilast on the behavior, inflammatory and tryptophan catabolic pathway changes in our animal models of depression. Reason is that pharmaceutical companies are not enthusiastic up till now to develop this drug for psychiatric usage. Moreover in the in vitro models the effects of piclamilast are minimal (see below), we therefore prefer regular anti-inflammatory agents, such as COX-2 inhibitors.
- We have not yet investigated the effect of tryptophan catabolic pathway enzyme inhibitors (particularly of the KMO inhibitor NW-1349) on behavior, inflammatory and tryptophan catabolic pathway changes in our animal models of depression, although medication for animal work has been prepared.

Future plans:

We continue the work in PSYCH-AID with the core groups of MOODINFLAME and we will specifically focus 1) on a COX-2 inhibitor treatment of LPS exposed NOD mice, and 2) a NW-1319 treatment in the NOD mouse model.

WP 10: THERAPEUTICS IN IN VITRO MODELS

Objectives:

1. To know the in vitro effects of the PDE4 inhibitor Piclamilast, the COX-2 inhibitor Cimicoxib and of desipramine and fluoxetine on

a. The monocyte signature molecule expression and aberrant monocyte function of signature positive patients and patients taking IFN-α.
b. The GR function in monocytes of signature positive patients
c. The tryptophan catabolism of monocytes of signature positive patients and patients taking IFN-α.

2. To know the in vitro effects of inflammatory mediators (e.g cytokines and chemokines) and the IDO-, KMO- or COX2-inhibitors mentioned in WP 9 on neurotransmitter changes in organotypic brain slice cultures

Achievements: Monocyte-microglia in vitro therapeutic studies

a: Piclamilast (PDE4 inhibitor) studies (EMC)

- Piclamilast, a PDE4 inhibitor, only has a mild anti-inflammatory effect on patient monocytes and in fact only on the TNF expression and the IL-1 and IL-6 production of the cells (EMC).
- The effects of anti-depressants and Piclamilast have been studied in relation to GC insensitivity of patient leukocytes (whole blood cell culture, IL-6 as read-out) (KCL):
- Clomipramine corrects the GC insensitivity of MDD leukocytes, particularly in patients with a good clinical response to the anti-depressant, piclamilast potentiates this effect of clomipramine
- Unresponsiveness to anti-depressant therapy can be predicted in particular by the in vitro effects of anti-depressants on MR function.

We decided not to continue further in vitro experiments with PDE4 inhibitors, since PDE4B turned out not to be the key major inflammatory factor in the monocyte activation signature (see also previous WP’s)

b. Functional in vitro assays to back up the GRalpha/beta gene expression measurements in monocytes (EMC, KCL)

Apart from the existing in vitro whole blood assay of KCL (endpoint measurement IL-6) to study steroid resistance of immune cells a FACS system has been developed by EMC to simultaneously test the glucocorticoid sensitivity of monocytes and T cells in the patient samples. This assay now measures Ca-ionophore/TLR stimulated IL-12/TNF-alpha production in monocytes and IFN-gamma production in the T cells in the absence or presence of graded concentrations of Dexamethasone (DEX).

Preliminary data show

- A higher sensitivity to Ca-Ionophore/TLR of monocytes of patients with a psychiatric disorder as compared to HC as evidenced by a higher production of IL-12/ TNF-alpha
- An altered pattern of reactivity of patient monocytes to dexamethasone in this test: Earlier a higher sensitivity to DEX then a reduced sensitivity
- No correlation with the HPA axis disturbances as measured with the saliva test.

c: Drug studies on BV-2 microglia cultures (TCD)

TCD demonstrated that conditioned media from IFN-gamma treated BV-2 microglia induce neurotoxicity in primary cortical neurons (see WP2). Pre-treatment of the BV-2 microglia with the IDO inhibitor 1-methyl-L-tryptophan and the KMO inhibitor Ro 61-8048 prevented the IFNgamma induced neurotoxicity. These findings further support the role of the kynurenine pathway in inflammation-induced neurodegeneration.

Following on from the initial characterization of the expression of the transcription factors ATF3, EGR3 and PU.1 in BV-2 cells in response to pro-inflammatory cytokines the objective was to determine the effect of anti-depressants, anti-psychotics and lithium chloride on the expression of these factors. Characterization of the BV-2 cell line showed that these cells express SERT and MAO-A but no other synthetic, degradation enzymes, or transporters for the monoaminergic neurotransmitter systems. In addition the following results were obtained:

- Inhibition of the serotonin transporter (with fluoxetine) and MAO-A (with moclobemide) inhibits the response of microglia to LPS, measured by iNOS mRNA expression and TNF-alpha production, classical markers of microglial activation.
- Preliminary data shows when BV-2 microglia are incubated with serotonin (5-HT), they have the ability to take up and degrade 5-HT, indicated by a reduction in 5-HT concentration in the culture medium and an increase in 5-HIAA. When cells are treated with IFN-gamma or LPS, their ability to metabolize serotonin is reduced significantly.
- Pre-treatment with 1 µM of fluoxetine, citalopram, moclobemide, clorgyline and lithium chloride had no effect on the induction of ATF3, EGR3 and PU.1 by IFN-alpha or IFN-gamma
- Pre-treatment with 1 µM of haloperidol and clozapine had no effect on the induction of ATF3 and PU.1 by IFN-alpha or IFN-gamma, but did potentiate the induction of EGR3 mRNA by IFN-alpha and IFN-gamma

d: Effect of anti-inflammatory drugs on dorsal raphe slices (IMU)

The effect of pro-inflammatory mediators, especially IFNgamma on death or survival of serotonergic neurons and IDO expression has been studied by Innsbruck. Innsbruck found a significant decrease of serotonergic neurons and enhanced IDO expression after inflammation in the doR as well as IDO co-localized with neurons and to a lesser extent with serotonergic neurons, but not with microglia or astrocytes. Innsbruck will explore in the coming year whether different anti-inflammatory drugs (Ibuprofen, Citalopram, Minocycline, Norharmane, Indomethacine, Celecoxib) can counteract above described inflammatory reaction in the slice cultures.

e: Hippocampal progenitor cells (KCL)

KCL has used an in vitro model, using human hippocampal progenitor cells (HPC03A/07, ReNeuron, UK) to investigate several molecular mechanisms and counteraction of these mechanisms by various drugs:

The production of IL-6 induced by IL-1beta was reduced by the antidepressant venlafaxine in this cell line, while it was increased by sertraline. Both antidepressants activated NfkB.

DHA dose-dependently reduced levels of transcripts for IDO, the enzyme that metabolizes tryptophan to kynurenine. No changes were observed for EPA. No significant changes were observed in the levels of kynurenine-aminotransferase-1, that regulates production of the potentially neuroprotective kynurenic acid.

Reached Objectives:

- We know the in vitro effects of the PDE4 inhibitor Piclamilast and of anti-depressants on the monocyte signature molecule expression and aberrant monocyte function of signature positive patients, i.e. piclamilast and anti-depressants have mild anti-inflammatory effects on inflammatory gene expression in patient monocytes.
- We have developed an in vitro assay system in which GR function can be measured in patient monocytes and T cells; we found up till now not much evidence for a GR insensitivity of monocytes and T cells regarding inflammatory compound (TNF-α and IFN-γ) production, on the contrary in some patient groups we have found a hyper sensitivity for glucocorticoids.
- We have assessed the in vitro effects of inflammatory mediators (particularly IFN-γ), of IDO-, KMO-inhibitors, of cortisol and of various anti-depressants on neurotransmitter, inflammatory compound and tryptophan metabolites in organotypic brain slice and other brain cell culture systems

Not yet reached objectives (reasons)

We have not systematically tested COX-2 inhibitors in either PBMC or brain cell culture systems

We have not systematically studied the tryptophan catabolism of monocytes of signature positive patients.

Both time constraints play a role in not carrying out these parts of the objectives.

Future plans:

We continue the work in PSYCH-AID with the core groups of MOODINFLAME and will specifically focus on the role of anti-inflammatory agents (not necessarily only COX-2 inhibitors) on microglia function.

WP 11: COX-2 INHIBITOR TRIAL

1. Objectives:
1. To investigate the clinical effect of Cimicoxib as an add-on medication to standard antidepressant SSRI treatment in patients with major depressive disorder in terms of response of depressive symptoms.
2. To investigate a) the predictive value of immune parameters (in particular monocyte COX-2 expression) prior to treatment regarding outcomes under objective 1 and b) the effect of Cimicoxib treatment on immune parameter changes.
3. To investigate a) the predictive value of serotonergic and tryptophan catabolic pathway parameters prior to treatment for outcomes under objective 1 and b) the effect of Cimicoxib treatment on serotonergic and tryptophan catabolic pathway changes.
2. Achievements:

SECIM Trial

The analyses of the SECIM trial (Part I) run by Affectis were finalized and the results were reported to the consortium at the consortium meeting November 2010. There was a (not significant) successful outcome in that the COX-2 inhibitor Cimicoxib had a potentiating effect on the regular anti-depressant therapy with sertraline as measured in the MADRS and HamD scores. Significance was almost reached (p=0.08) for the MADRS when only severe patients with a HamD score of >22 were included.

Further analyses have been performed to identify sub-groups of patients which do or do not show remission after the add-on treatment with the Cox-2 inhibitor. Serum samples (collected at time points 0, 2 and 6 weeks of treatment) of age- and gender matched individuals of those subgroups were transferred to partner EMC to investigate:

a) The predictive value of immune parameters prior to treatment regarding outcomes under objective 1, and
b) The effect of Cimicoxib treatment on immune parameter changes.

Immune serum parameters have been tested using the Raybiochem system. However this system turned out not to be reliable. Plasma’s will be further stored till Moodinflame and Psych-Aid (the runner up of Moodinflame) have been fully analysed and we will be better informed on which plasma parameters are the best to evaluate.

CITICOX trial

The CITICOX trial has been severely delayed due to

1. A problem with the stability of Cimicoxib. It was decided after a year to switch over to celecoxib. Affectis supported the principal investigator in all regulatory issues. The blinded study medication (placebo, Celecoxib) is placed at the disposal and distributed upon request to the clinical centres involved in the study.
2. Delays with the ethical approval in countries like the NL, Italy and Bulgaria (FBP). The review boards are continuously asking improvements and are critical towards the use of new therapeutics in a vulnerable group as MDD patients. Finally the review board in Bulgaria has approved the study, but up till now the review board in the NL and Italy have delayed the process such that we are not able anymore to start the project, since clinicians do not want to support the study any further. Nevertheless approval has been given in Germany in 2011.

At present a total of 30 patients have been recruited by LMU group and 2 by WWU (which institute certified 5 clinical raters, set-up study workflow and procedures). IMU and FBP have ethical approval, but were unable to recruit patients who fulfill the criteria.

We have changed the COX-2 inhibitor Cimicoxib (unstable preparation) to Celecoxib in the CITICOX study. Study will be continued in PSYCH-AID, the runner up of Moodinflame

Due to the change in COX-2 inhibitor and other problems we are delayed with the CITICOX study. Still we hope to recruit in the coming years enough patients to complete the study. It must be noted that the future laboratory studies on the previous SECIM trial samples will probably also enable us to reach part of the goals as set in the objectives of this WP.

Not yet reached objectives (reasons):

We have neither been able yet to investigate the clinical effect of Cimicoxib as an add-on medication to standard antidepressant SSRI treatment in patients with major depressive disorder in terms of response of depressive symptoms, nor have we been able to investigate immune and tryptophan break-down parameters as predictors of such treatment.

The reason is that we have at the closing date of MOODINFLAME only recruited 32 patients (for the trial we need at least 50 patients). Reason for the low recruitment is the bankruptcy of the SME supplying the special COX-2 inhibitor for the trial in the 2nd year of the project and the changes which needed to be made in the trial due to this bankruptcy. Thereafter we experienced a long process of obtaining (new) ethical permission for the trial in some centers (Muenchen, Muenster). We finally also failed to get ethical permission (in time) for other centers (EUR, Milano, Sofia), since the Review Boards were super critical because the trial was not sponsored anymore by a pharmaceutical company. Furthermore selection criteria appeared to be too strict for the two recruiting centers to recruit sufficient numbers of patients within the resting time limits of the project (Muenster).

Future plans:

We continue the work in PSYCH-AID with the core groups of MOODINFLAME and specifically focus on selecting enough patients for the clinical trial via the Muenchen group (and the Muenster group) to be at least able to see whether we can predict the response to COX-2 inhibitors via immune or tryptophane break down products.

Potential Impact:

Strategic impact

European Health care

With regard to size and burden of mood disorders in Europe there are two excellent recent reviews (Paykel et al, European Neuropsychopharmacology 15: 411-423, 2005 and Pini et al, European Neuropsychopharmacology 15: 425-434, 2005). The large number of prevalence studies from the older EU allows clear conclusions. Unipolar depression (UD) is an important phenomenon throughout Europe with 12-month prevalence rates ranging from 2.0 to 8.3 %, thus affecting around 1 in 40 adults at any given time and 1 in 20 for at least a few weeks in any year. Females are about twice as much affected than males, particularly in the middle age groups. In those with a less privileged social position (e.g. low income, unemployed, etc.) prevalence rates of UD are higher. BD has a lower 12-month prevalence of approximately 1% with no evidence for major differences by age group or gender. Mood disorders produce substantial economic costs, and are particularly expensive because they are often life-long and cause major medical and societal problems.

The relationship of mood disorders to stressful life events is well established, both in epidemiological as well as clinical samples. Yet, these stressful events seem – in relation to prevention – inevitable consequences of the life cycle and the family relationships particularly in the affluent “older” EU states. Therefore primary prevention targeting these events should be approached cautiously. However our approach of drug-targeting an underlying biological predisposition in individuals at risk, i.e. an activated

Immune-endocrine-metabolic system, might be more feasible to make progress in preventing mood disorders.

We have now developed in MOODINFLAME a series of test systems to detect the immune activation state of microglia and of peripheral immune cells and linked endocrine and metabolic abnormalities (HPA-axis abnormalities, toxic tryptophan break down products) in patients, but more importantly also in individuals at risk for major mood disorders.

There are various drugs available that counteract this immune activation state and these drugs have been explored in MOODINFLAME both in vitro as well as in the animal models developed in MOODINFLAME. These drugs are or will be formally tested in larger clinical trials presently started to treat disease, but also in future trials which will be started to prevent disease in individuals at risk.

Therefore the subsequent exploitation of these tests and drugs will have a significant impact upon the general health status of the EU population and an extraordinarily positive effect on health costs in Europe. We predict that MOODINFLAME will lead to the use of clinically applicable assays to detect early and pre-stages of mood disorders and to intervention strategies to halt or slow the development of the disorders.

European Science

The MOODINFLAME project has met the strategic objectives of Integrating and Strengthening the European Research Area as defined by the European Commission in COM 2002 834 by bringing together 15 academic research groups and 4 SMEs from 10 different countries and by establishing a critical mass of expertise required. The project is now continued in an EU-funded IAPP project PSYCH-AID.

The MOODINFLAME project has provided an excellent framework for educating the next generation of young scientists in the field of the immunological aspects of mood disorders, because of its multidisciplinary nature. At present more than 10 PhD theses have been or are under development in the project.

MOODINFLAME has facilitated and still facilitates via the PSYCH-AID project the exchange of personnel within European countries and fosters contacts between different laboratories and clinics at all levels.

Since several of the senior investigators and younger staff are women, they have provided role models and support for younger women scientists.

Industry

The SMEs participating in this consortium have provided the knowledge to implement the data bases and assays into practice. The project has provided the ideal environment to achieve an integrated approach between scientists from academia and industry, and furthermore there are presently more intense collaborations between academia and industry in the runner up project PSYCH-AID. A successful program along the lines proposed herein will, therefore, confer important strategic advantages for European life science in academia, in healthcare, in the pharmaceutical industries, and other areas, and will generate valuable intellectual property. As life sciences and functional genomics are among the most promising frontier technologies, the MOODINFLAME and PSYCHAID projects have high potential to contribute to the aim of making Europe a leading force in this sector, and will help increase the number of highly skilled jobs in Europe.

Reinforcing European competitiveness and solving European societal problems.

European research in the field of mood disorders has an outstanding worldwide reputation and MOODINFLAME has contributed to that. The project has become well known in the field of psychiatry, also outside Europe.

The development of personalized diagnostic tests coupled to a drug treatment approach to prevent mood disorders is a formidable challenge. This can only be achieved by bringing together leading European centres of expertise in the fields of mood disorders in general, in the immune aspects of mood disorders, in immunology in general, in cell biology, in neuroscience, in genetics, in pharmacology and diagnostic development. Fortunately, the collective capabilities of the different European groups participating in the MOODINFLAME project has provided all the technical skills and innovative advances required to do so. Work in this cooperative approach has enabled European scientists and the biotechnology industry to maintain their leading position as compared to the research efforts ongoing in the US (financed by e.g. the Stanley Foundation and NIH) and in the Far East, where significant financial resources are invested into this type of research.

The project has set up a long lasting and strong interaction between clinicians and scientists based on current state-of-the-art expertise in cellular and molecular biological and immune techniques. Importantly this has resulted in several joint publications and will continue to do so. The present program has allowed a solidification of the collaborations, and their extension to novel partners.

The strategic importance of our research lies in the fact that there is an important demand – millions of patients suffering from mood disorders in Europe.

The key innovations brought by MOODINFLAME include:

1. The usage and targeting of monocytes and microglia and the tryptophan catabolic pathway for diagnostic, prognostic and therapeutic purposes.
2. The selection of subgroups of mood disorder patients characterized by a specific monocyte mRNA signature and/or a tryptophan catabolic disturbance. We are so bold to state that we are the first to be able to make such a distinction with the signature and tryptophan tests and with such distinction it is – in our opinion – possible to make a step forward towards a targeted treatment of mood disorders, since patients can be selected using the tests for a specific treatment targeting the immune and tryptophan metabolic aberrancies. Our program thus results in a personalized treatment of mood disorder patients on the basis of their specific immune abnormality.
3. The development of a novel animal model of mood disorders characterized by an inborn and intrinsically activated inflammatory system, the NOD mouse model. This model was already known for years as an excellent model of autoimmunity.

The model will enable the scientific community to study the inflammatory response characterisic of mood disorders at the level of the brain in neuron-astrocyte-microglial interactions and to study novel anti-inflammatory therapies for their efficacy in mood disorders.

Exploitable fore ground

In the end this project has generated in particular a deepened knowledge on the multiple immune mechanisms which play a role in disease development in sub sets of mood disorder patients. Complex in-time-changing disturbed interactions between microglia-astrocytes-neurons become evident as cornerstones in the gradual development of mood disorders. Simple assays on peripheral blood cells and serum to detect these complex multiple and changing interactions cannot be developed and patented, since many immune genes and compounds are simultaneously changed.

Nevertheless it might be possible to identify in our rostrum of data genes and compounds of key importance in the prediction of disease development and therapy outcome. Particularly in the follow-up project PSYCH-AID we want to identify these molecules and file these for a patent. We think e.g. of monocyte JUN expression/percentages of CD8+T cells to predict responsiveness to anti-depressants and antibodies to tryptophan break-down products to be used in ELISA’s to detect abnormalities in these products in individuals at risk for mood disorders.

List of Websites:

The Public website address for the Moodinflame project is http://moodinflame.eu. At this website one can find the details of the participants of the consortium of the Moodinflame project as well as the yearly Publishable Summaries.

Final Report Summary - MOODINFLAME (Early diagnosis, treatment and prevention of mood disorders targetting the activated inflammatory response system)

Verwandte Dokumente

Diese Seite teilen

Herunterladen