Major depression and bipolar disorder are common clinical conditions with substantial consequences for both, the personal suffering of the patient and adjustment to the disease by friends and family as well as for the society. It had been recognized that mood disorders are a leading cause of disability posing a huge toll upon health care systems. In fact, mood and anxiety disorders are not only second to cardiovascular disease as cause of illness-related disability, they are also significant risk factors for a number of illnesses that include heart-attack, high blood pressure, stroke and diabetes.
The size of the problem is reflected by epidemiological studies which currently estimate that almost 19% of women and 11% of men will suffer from major depression at some point of their lifetime. The figures for bipolar disorder are one order of magnitude lower but tend to increase as people get older because the annual diagnostic conversion from major unipolar disorder to bipolar disorder is constantly 1.25% per year throughout life. Another factor deserving attention in an aging population is the difficulty to achieve remission from depression and the tendency to become chronically ill.
One cardinal symptom of depression is anxiety which often dominates the clinical picture and presents in many clinical facets, ranging from panic attacks to social phobia. For some practical reasons these anxiety disorders have been defined as clinical entities separate from depression. Whether this separation is justified by causal mechanisms of these disorders is yet not determined. The distinction has no clinical implication, as for both, depression and anxiety disorders, antidepressant drugs are the first line treatment.
Scientific advances in the neurosciences focusing upon depression and anxiety continue to be made at a rapid pace but still failed to implement major improvements in the drug development process. In fact, the current antidepressants are modifications of drugs serendipitously discovered half a century ago. Genomics has provided huge amounts of data but its complexity and lacking ability to allow for gene-environment interactions has prevented it from reaching its full potential as source of drug targets.
The Munich Antidepressant Response Signature Project (MARS) is devoted to identify gene variants predicting treatment outcome and to set-up a biomarker portfolio that helps to optimize genomic information. We believe that the combination of genotypes with differentially weighed information of biomarkers will ultimately result in a formula that is predictive for antidepressant treatment response. Solving this task is feasible if techniques needed are mastered and if access to patients under carefully monitored clinical conditions is possible at a large scale. The ultimate goal is the delivery of personalized medicines for treatment of depression – a task that poses major challenges to basic researchers, clinician scientists, healthcare insurances, public research organisations and industry – for the sake of the patients which justly expect better therapies.
Our research aims at designing a navigation chart for future personalized medicine and explores the potential of various technologies to generate biomarkers that are useful for clinical practice as well as for drug discovery and development.
The MARS-project will provide information which combination of biomarkers together with genotypes delivers the optimal information for the above goals.
Gene Expression Profiles
Profiling gene expression with DNA (or synthetic oligonucleotide) microarrays allows to measure the transcriptional activity of thousands of genes. Clearly, profiling gene activity in lymphocytes does not allow for studies of causality of signalling pathways involved in antidepressant treatment. However, gene arrays can produce a molecular fingerprint how all the changes of humoral homeostais – well documented among depressives – impacts upon gene activity. These molecular signatures will vary according to genetic, epigenetic and humoral activity of the individual patient and will also reflect the effect of a given drug upon individual transcription. Preliminary results from the Max Planck Institute of Psychiatry show that this technology is feasible and reliable and that it carries the potential to differentiate subgroups of affected patients. We currently test the hypothesis that gene expression in blood cells of mice could reflect respective changes also in the mouse brain. If this holds true it seems justified to make such extrapolations also in humans. In particular, the gene expression profile of an individual prior to and following drug exposure will be indicative in which way a patient may respond to the pharmaceutical intervention.
Proteomics allows the study of genome products and their interactions at any given point in time. It complements the transcriptome, available through gene expression profiles, which essentially quantifies mRNA, as there is no direct relationship between the concentration of mRNA and its encoded protein. Protein degradation, differential rates of translation of mRNA into proteins are among the reasons of poor correlation between mRNA and protein levels. Moreover, while genomics has to deal with around 25.000 genes (admittedly complicated by the current upsurge of non-gene sequences impacting upon gene regulation) the proteome represents around 600.000 to 800.000 individual proteins, each of which can undergo further chemical specification.
We expect that proteomics may ultimately deliver protein patterns that amplify the information from genotyping and gene-expression profiles. The ultimate goal however would be the identification of a biomarker that can be derived from blood. The manifold of blood components that are present at extremely diverse concentrations makes the discovery of such a biomarker a daunting task. The proteomic platform at the Max Planck Institute of Psychiatry is well equipped and in the light of the animal models that can be taken as reference, the huge repository of blood and cerebrospinal fluid samples at the Institute provide a good opportunity to solve this task. Preliminary results underscore that we are aiming not too high.
The study of all small molecule metabolites in blood serum, cerebrospinal fluid or urine, called metabolomics holds promise to predict an individual’s disease risk. From plant physiology it is known that metabolite profiling allows predictions regarding a plant’s future development, e.g. growth rate. To date, cerebrospinal fluid can be considered the best compartment to discover biomarker fingerprints for early diagnosis of brain disease. The potential of cerebrospinal fluid metabolomics for monitoring drug effects in the brain is yet unexplored. Currently we are designing a study that interrogates the effect of an antidepressant given orally to patients and healthy controls. Once entering in the brain, the mode of action should be reflected by changes of the metabolite profile in the cerebrospinal fluid and blood. We postulate that antidepressants share a metabolomic signature. A testable hypothesis is that the changes in the metabolite pattern depend on the patient’s genotype and proteome. A combination of genomics, proteomics and metabolomics is therefore likely to increase our understanding of the biology behind depression and mode of action of drugs relieving this clinical condition.
We have currently established a collaboration between the Max Planck Institute for Molecular Plant Physiology (Professor L. Willmitzer) in order to explore whether changes in metabolite patterns in the cerebrospinal fluid of patients from the MARS study may be predictive for the way patients respond to the antidepressive treatment. Once established we also postulate that measuring changes in the metabolome following drug exposure will serve two purposes: (1) it will hint whether the drug is a potential antidepressant at all; and (2) whether it will have beneficial effects in the specific patient under study. The latter aspect includes adverse effects, which may be helpful since many newer antidepressants interfere with metabolic pathways.
The most vigorously validated neuroendocrine test that allows characterisation of stress hormone function with high sensitivity is the so called dexamethasone/CRH (corticotrophin releasing hormone)-test, discovered and developed by researchers at the Max Planck Institute of Psychiatry. In this test patients have to ingest orally a low dose of dexamethasone at 11 p.m. and receive at 3 p.m. the next day a test dose of 100 µg CRH intravenously. Subsequently, secretion of the stress hormones corticotrophin and cortisol is monitored. Normal controls exhibit only small increases of their hormone secretion. In contrast, patients with depression show excessive corticotrophin and cortisol secretions, which during successful treatment return to almost normal values. Those patients who fail to normalize responding are less well or prone to relapse after intermediate amelioration of symptoms. These findings, along with molecular studies of glucocorticoid receptor mediated cellular signalling and mouse models carrying modified genes involved in stress regulation, were the main pillars on which the corticosteroid receptor hypothesis was built.
Studies at the Max Planck Institute of Psychiatry and elsewhere have repeatedly demonstrated that changes in the dex/CRH test meet the criteria for a clinical biomarker predicting drug response prior to changes of clinically overt symptoms. Today a dex/CRH test result is considered a viable endophenotype constructing subgroups and thus supporting genetic studies.
Changes in brain activity among patients with depression resulted in a number of findings which may have biomarker potential. One advantage of functional magnetic resonance imaging (fMRI) is that it responds to cognitive stimulation in a way that reflects depressive mood. For example, a well established neuropsychological impairment in depression is a mood-congruent processing bias such that ambiguous or positive events tend to be perceived as negative. Such impairment is reflected in the fMRI as reduced capacity for activation in several brain areas deemed to be implicated in the neuropathology of depression. Such impairments tend to resolve under successful antidepressant treatment posing the question if such imaging techniques may an efficient way of screening candidate drugs in small populations.
Research at the Max Planck Institute of Psychiatry has used a paradigm that focuses upon the resting state of networks that implicates brain structures relevant for depression and anxiety. Abnormal activation patterns were found under these experimental conditions. This experimental set-up can be kept constant across centres and is independent from actual neuropsychological test performance. If we can demonstrate that these changes are reversible prior to symptomatic relief from depression, our fMRI paradigm might be able to develop into a bona fide biomarker.