Peripheral brain-derived neurotrophic factor (BDNF) as a biomarker for affective disorders?

After more than a decade of ‘a molecular and cellular theory of depression’ the neurotrophin brain-derived neurotrophic factor (BDNF) remains a central part of the concept for the pathophysiology of depression and antidepressant treatment (Altar, 1999; Duman et al.1997). Stress, an important precipitant of depression, reduces the expression of BDNF, especially in the hippocampus, while a plethora of antidepressant measures, e.g. antidepressant drugs, electroconvulsive treatment and also environmental enrichment, increase BDNF expression (Krishnan & Nestler, 2008; Martinowich et al.2007). Thus, BDNF could be a valuable biomarker for a depressive state and the recovery from it. Unfortunately, CSF levels of BDNF are already at the limit of detection in healthy individuals, despite sensitive methods of analysis (Burbach et al.2004). Of course it is also not possible to easily determine BDNF brain tissue levels in the living patient. Therefore, there is great interest in peripheral measures of humoral BDNF, particularly in blood (Brunoni et al.2008; Sen et al.2008). The normally very high concentration of BDNF in blood could be derived not only from the brain, but also from peripheral sources, since most body tissues express this growth factor (Lommatzsch et al.2005a; Nockher & Renz, 2005). Given that cerebral BDNF crosses the blood–brain barrier (Pan et al.1998), it is reasonable to assume that serum BDNF concentrations are associated with BDNF levels in the brain. This assumption is substantiated by animal experiments showing that BDNF serum levels are correlated with BDNF expression in cortical brain regions (Karege et al.2002b), and the recently demonstrated positive association between N-acetylaspartate – a well established marker of neuronal integrity (Moffett et al.2007) – and serum BDNF in healthy human subjects (Lang et al.2007). Within blood, BDNF is mainly stored but not synthesized in platelets/thrombocytes (Fujimura et al.2002; Pliego-Rivero et al.1997). Therefore, there is an ∼200-fold difference between BDNF levels determined in serum vs. plasma, probably due to release of BDNF from platelets into serum during coagulation (Lommatzsch et al.2005b). However, it should be borne in mind that platelets store BDNF from different sources. Notably, cells from the central nervous system and endothelial cells seem to contribute to blood levels of BDNF, and platelets are thought to be a storage compartment (Lang et al.2004b; Lommatzsch et al.2005a, b). Thus, platelets may be actively taking part in the regulation of homeostasis by storing BDNF for later release in periods of increased demand (Fujimura et al.2002). In this sense platelet BDNF might reflect a ‘buffer system’ (Ziegenhorn et al.2007). One situation where an increased BDNF demand cannot be met might be depressive disorder. Indeed, there is increasing evidence linking lowered or reduced BDNF serum levels to depression (Karege et al.2002a, 2005; Lang et al.2004a, 2005; Shimizu et al.2003). Most clinical studies in cohorts of depressed patients were performed with serum. In these studies BDNF levels were lowest in non-treated depressed patients while antidepressant treatment increased BDNF levels (Brunoni et al.2008; Gervasoni et al.2005; Sen et al.2008). However, most of these studies were hampered by the fact that the humoral/serum BDNF levels reported for the healthy control groups were quite different (Karege et al.2005). To give a striking example: the same work from Chiba University (Japan), published in two different papers in 2003, reported mean BDNF-control levels with a two-fold difference (Nakazato et al.2003; Shimizu et al.2003). Even from a more methodological point of view, storage of serum (but not whole blood) at −20°C was associated with a significant decrease in BDNF concentration (Trajkovska et al.2007). Women had significantly higher whole-blood BDNF concentrations than men, and showed a right-skewed BDNF concentration distribution (Trajkovska et al.2007), whereas no significant gender differences on humoral BDNF levels were reported by others for either plasma and serum (Lommatzsch et al.2005b; Ziegenhorn et al.2007). Moreover, BDNF serum levels are not normally distributed, indicating that previously published studies with small sample sizes using parametric testing may be misleading (Ziegenhorn et al.2007). Finally, almost all studies on humoral BDNF are done by expensive, commercially available BDNF-ELISA kits run in duplicate on each sample, thereby ignoring the impact of unspecific binding and/or spike recovery on the obtained results (by contrast, cf. Ziegenhorn et al.2007). Cattaneo and colleagues present a clinical study evaluating BDNF both in serum and leucocytes of depressed patients prior to and during treatment with escitalopram (Cattaneo et al.2009). Interestingly, serum BDNF protein levels as well as leukocyte BDNF mRNA levels were reduced in drug-free patients compared to healthy controls, and recovered gradually during a 12 wk course of treatment. The increase in BDNF mRNA levels also correlated with clinical improvement. These authors speculated that peripherally synthesized BDNF may have a direct effect on brain function and behaviour. This interesting hypothesis should be followed up in animal experiments. The simplest experiment would be peripheral administration of BDNF and a study of the cellular and behavioural effects. A more specific and mechanistic approach would include a leukocyte specific knockout of the BDNF gene and analyses to discover whether this causes a change in depression-like behaviour and/or alterations of the response to antidepressant treatments. Furthermore, the latter approach could clarify to what extent leukocytes contribute to BDNF plasma and serum levels. Even if BDNF down-regulation in leukocytes from depressed patients and its up-regulation subsequent to antidepressant therapy were not causally related to the physiological mechanisms related to depression and its therapy, BDNF expression in leukocytes could represent a valuable biomarker for the diagnosis or clinical course of depressive disorders. As outlined above, BDNF serum levels probably represent an integrative composite from different sources, and the temporal dynamics of changes in BDNF serum levels may not closely correlate with those of brain tissue levels (Sartorius et al.2009). Due to the short half-life of mRNA, leukocyte BDNF mRNA levels could more accurately reflect the central mechanisms of BDNF physiology. At present, however, this is mere speculation which needs to be validated by concurrent investigations of BDNF mRNA expression profiles in experimental animal models of depression and antidepressive therapy. Moreover, studying BDNF serum levels would complete the picture. Although depression-related behaviours in BDNF conditional knockout mice may provide further evidence for a role of BDNF in depression (Monteggia et al.2007), reduced levels of serum BDNF have also been observed in other neuropsychiatric disorders, e.g. schizophrenia, panic disorder, eating disorders, neurodegenerative diseases such as Alzheimer's and Huntington's diseases and others (Ehrlich et al.2009; Lang et al.2004b; Nakazato et al.2003; Post, 2007; Schulte-Herbruggen et al.2007; Strohle et al.2009). This may not be a surprise given that various kinds of stress are able to reduce BDNF levels at least in experimental animals (Krishnan et al.2008; Martinowich et al.2007; Post, 2007). In contrast, decreased expression or availability of endogenous BDNF may be correlated with an increased vulnerability for the development of (or at least some symptoms of) different neuropsychiatric diseases as indicated by different transgenic mouse models of depression (Monteggia et al.2007; Ridder et al.2005) or Alzheimer's disease (Hellweg et al.2006), or even within serum in healthy volunteers with depression-related personality traits (Lang et al.2004a). This view will limit the use of humoral/peripheral BDNF levels as a possible disease-specific diagnostic marker. However, the increasing evidence that BDNF serum levels increase differentially following antidepressant treatment (but see Hellweg et al.2008; Jang et al.2009) may suggest this neurotrophic factor as a potential biomarker to predict an individual response to a specific antidepressant treatment. Further studies are required to determine whether leukocyte BDNF expression levels are a more specific marker than serum/plasma levels for diagnostic purposes or for predicting the clinical course of depression. This work was supported by grants from the Deutsche Forschungsgemeinschaft (GA427/9-1 to P.G. and HE 1392/3-1 to R.H.). None.