An ambitious attempt to link cognitive tests, functional brain imaging, and cell biological assays implicates a polymorphism in brain-derived neurotrophic factor (BDNF) as the source for highly specific memory deficits, alterations in hippocampal activity, and protein processing defects. The study, led by researchers at the National Institutes of Health, appeared in the January 24 Cell.

Beyond its well-documented role in supporting neuronal growth and survival (see related news item), BDNF plays a significant role in hippocampal long-term potentiation (LTP), and in learning and memory (see Lu and Gottschalk, 2000; Poo, 2001). Given the presence of a frequent polymorphism in the human BDNF gene (a valine for methionine substitution), and some evidence of hippocampal dysfunction in schizophrenia patients and their siblings, Daniel Weinberger of the National Institute of Mental Health and Bai Lu of the National Institute of Child Health and Human Development, both in Bethesda, Maryland, led a team to test whether BDNF genotype might be a risk factor for hippocampal dysfunction in these groups. The researchers also explored whether BDNF polymorphisms might affect the cellular processing of the peptide.

The cohort (n = 641) for the memory experiments included schizophrenia patients, their unaffected siblings, and normal controls, all drawn from a large sibling study of schizophrenia. The intent of the experiment was to determine whether the BDNF genotype was related to memory dysfunction in schizophrenia, but the researchers found no such correlation. Instead, they found that possession of the met/met genotype was associated with a significant, and similar, episodic memory deficit relative to val/val or val/met in all three patient subgroups. Interestingly, there was no such difference in tests of word recall, semantic memory, or working memory/executive function-memory domains that may depend less upon the hippocampus.

In two imaging experiments, the researchers were able to correlate this suggestion of hippocampal dysfunction with hippocampal activity differences. They used fMRI to monitor brain function in subjects performing the N-back test, a memory trial that involves primarily neocortex while tending to "disengage" the hippocampus. In two independent, small cohorts of normal controls (13 and 17 subjects, respectively), the researchers found that val/val subjects had the expected deactivation of the hippocampi, whereas val/met subjects had an inappropriate overactivation of both hippocampi. (There were not enough met/met subjects to make a valid assessment of this subgroup.)

In the second imaging study, the researchers used MRI spectroscopy to measure n-acetyl-aspartate (NAA), which has been proposed as an indirect marker of neuronal integrity and synaptic abundance (Maier et al., 1995). Again studying all three groups (schizophrenia patients, siblings, and controls), they found that the NAA signal was significantly reduced in the left hippocampus of val/met subjects relative to val/val subjects. In the right hippocampus, the researchers noted only trends in this same direction. Again, the number of met/met subjects was too small to allow direct comparison to the other groups, but a multiple regression analysis yielded a linear reduction in NAA levels in the left hippocampus with an increasing number of met alleles, suggesting an allele dose effect.

The third level of experiments looked at the effects of BDNF genotype on cultured rat hippocampal neurons. Neurons were transfected with either val- or met-BDNF, and the polymorphisms did not appear to affect the production of the protein, or its ability to function as a neurite-inducing growth factor. However, fluorescence and double-labeling techniques indicated that, whereas val-BDNF occurred both in the cell body and on dendrites, met-BDNF was confined to the cell body. Fluorescence analysis suggested that met-BDNF levels were lower than val-BDNF levels, and that the met-BDNF accumulated near the nucleus, whereas the val-BDNF showed a punctate distribution in the cell body and dendrites.

In response to a depolarizing challenge, val-BDNF-transfected, but no met-BDNF-transfected neurons markedly increased their secretion of the protein into the culture medium. In contrast to this activity-dependent secretion, there was no difference in the amounts of regularly (constitutively) secreted protein between the two groups of cells.

Hypothesizing that the met polymorphism may lead to sorting of the protein into the wrong secretory pathways, the researchers examined its subcellular locations. Markers for various intracellular organelles indicated that only val-BDNF is successfully sorted from the Golgi apparatus to secretory vesicles. This would seem to explain why val-BDNF, but not met-BDNF, is found near synapses and is secreted in response to the depolarization challenge.

The accumulated results from this collaboration lead the authors to suggest that these problems with intracellular trafficking and activity-dependent secretion of the met-BDNF could play a role both in the altered hippocampal function seen in imaging scans and the deficits on episodic memory tests. Further, "it is reasonable to speculate that the gene will impact the manifestation of diseases where function of the hippocampus and memory are impaired by the disease. Thus, one can imagine that a condition such as Alzheimer's disease, which destroys the hippocampus, may produce more dramatic effects or have a worse or more rapid course in individuals who have the met-BDNF genes in comparison to individuals with the val form of the gene. Similar phenomena may also occur with normal aging and in depression," said Weinberger in a Cell press release.—Hakon Heimer


  1. We were impressed by the research presented in Egan et al., which clinches a decisive role for BDNF in cognitive processes in humans. In animal models there has been an increasingly strong basis supporting an important role of BDNF in cognitive function, and the Egan study has translated this effect to humans. Essentially, the message is that when BDNF is not functioning properly, cognition suffers in the long term. Importantly, the study demonstrates that these cognitive effects occur in young/middle-aged, cognitively intact individuals. It is especially interesting that their cell culture results indicate that the effect is not in the synthesis of BDNF, but rather in the processing and release. The polymorphism is in the proBDNF (thus, on the preprocessed protein rather than in the active protein), and is affecting intracellular trafficking and also activity-dependent release.

  2. In the present studies, Daniel Weinberger and colleagues have reported that valine (val) substitution to methionine (met) in BDNF was associated with a failure to localize in secretory granules and synapses. The studies using transfected neurons appeared to translate to some degree into the human condition, where the val/met polymorphism was associated with poorer episodic memory.

    This is a difficult series of studies and the authors have generated some interesting data. Their imaging data were complicated by the numbers of subjects in the met/met group, but nonetheless, the authors have shown that there was no correlation between BDNF genotype and memory dysfunction in schizophrenia. In addition, the studies indicated that possession of the met/met genotype was associated with a significant episodic memory deficit relative to val/val or val/met in all three patient subgroups.

    These data strengthen rodent data indicating that BDNF plays a key role in memory and learning. While the molecular mechanisms of memory formation are complex and the role of long-term potentiaton widely debated, it seems possible that the initial rapid events (calcium entry, early LTP) then lead to a secondary phase of protein-synthesis dependent (late LTP) and later phenomena. The recruitment of cAMP and CREB signalling pathways could clearly result in increased secretion of BDNF, and there is an accumulating body of evidence that this neurotrophic factor plays a key role in synaptic processes that underlie memory formation.

    The lack of effect in schizophrenia is of interest. A recent postmortem study by Chen and coworkers (Chen et al., 2001) reported increased BDNF levels in the hippocampus of depressed patients treated with antidepressants (chronic antidepressant treatment is well known to increase cAMP, CREB and BDNF in rodents), but no such increase was observed in schizophrenia patients. Taken together, these data may suggest that BDNF plays a key role in learning, memory and depression, but different additional factors may come into play in schizophrenia.

    It is clear from preclinical studies that tetanic stimulation and memory tasks increase BDNF gene expression, and gene knockout studies and infusion of anti-sense to BDNF impairs spatial memory tasks. Therefore, the data seem to show BDNF is pivotal in cognitive processes. What is less clear is how this links to activity-dependent processes and how BDNF is packaged and secreted at the correct time and in the correct synaptic locations. The data from Weinberger and colleagues suggest that BDNF with the met polymorphism may be processed differently, have a low synaptic level, and may be released in lower levels in response to depolarization. These data suggest that BDNF does play a key role in memory formation in humans.

    It is possible that, in the aging brain or in disease situations (AD, depression), the met polymorphism would be associated with a more severe or rapid onset of symptoms. The data also suggest that viral vector delivery of neurotrophins or small-molecule approaches that increase synaptic BDNF levels may have potential for treating symptoms and perhaps the neuronal abnormalities in these disease states.


    . Increased hippocampal BDNF immunoreactivity in subjects treated with antidepressant medication. Biol Psychiatry. 2001 Aug 15;50(4):260-5. PubMed.

  3. Several lines of evidence show that BDNF is implicated in hippocampal long-term potentiation, learning, and memory in nonhuman species, but until now the involvement of this factor in human memory and hippocampal function has not been examined directly.

    In this respect, Egan and colleagues provide evidence that a valine/methionine substitution polymorphism at codon 66 (V66M) in the 5’ pro-region of the human BDNF gene affects intracellular distribution, packaging, and release of BDNF protein in vivo. Furthermore, they elucidated the effect of different genotypes, in human, on verbal episodic memory, hippocampal physiological activation, neuronal integrity and synaptic abundance. The authors did not find evidence that V66M polymorphic system was associated with schizophrenia, but suggest a possible role in other neurological and psychiatric disorders. Last year, several papers reported a genetic association of V66M polymorphism with different neurological diseases. In particular, the val allele has been associated with susceptibility to Alzheimer’s disease (AD) (Ventriglia et al., 2002) and, in two family-based independent studies, with an increased risk to develop bipolar disorder (Neves-Pereira et al., 2002; Sklar et al., 2002). Similarly, the met allele was associated with Parkinson’s disease by (Momose et al.., 2002) and recently the V66M has been associated with susceptibility to anorexia nervosa (Ribasés et al., 2002).

    In light of these results, further studies are needed to clarify the effective role of the two BDNF alleles in the pathogenesis and symptomatology of each disease.


    . Association between the BDNF 196 A/G polymorphism and sporadic Alzheimer's disease. Mol Psychiatry. 2002;7(2):136-7. PubMed.

    . The brain-derived neurotrophic factor gene confers susceptibility to bipolar disorder: evidence from a family-based association study. Am J Hum Genet. 2002 Sep;71(3):651-5. PubMed.

    . Family-based association study of 76 candidate genes in bipolar disorder: BDNF is a potential risk locus. Brain-derived neutrophic factor. Mol Psychiatry. 2002;7(6):579-93. PubMed.

    . Association studies of multiple candidate genes for Parkinson's disease using single nucleotide polymorphisms. Ann Neurol. 2002 Jan;51(1):133-6. PubMed.

    . Met66 in the brain-derived neurotrophic factor (BDNF) precursor is associated with anorexia nervosa restrictive type. Mol Psychiatry. 2003 Aug;8(8):745-51. PubMed.

  4. The exciting paper by Egan et al. is a far-reaching collection of experiments ranging from cell culture to human behavior that convincingly demonstrate the differing properties of proBDNF molecules carrying val66 or met66 polymorphic substitutions and their role in human episodic memory. It is very clear from the transfection experiments in hippocampal neurons that valBDNF localizes to dendrites, whereas metBDNF localizes to cell bodies. The green fluorescence is of a different character (punctate for valBDNF-GFP, diffuse for metBDNF-GFP), reflecting localization in different subcellular compartments, and only valBDNF, not metBDNF, is released at the synapse by regulated secretion. The authors have made an important discovery, namely, that the polymorphism in the pro region of BDNF is important for intracellular trafficking and activity-dependent release of BDNF. However, the number and size of the proteins, and more importantly their processing to the mature form, are not affected by the polymorphism.

    The authors carefully studied the biological activity of the mostly mature BDNF resulting from the processing and secretion of either valBDNF or metBDNF by COS cells (Fig. 7A). Not surprisingly, since the polymorphism is in the pro region, the mature forms exhibit no differences in TrkB activation or neurite outgrowth activity. What is important to determine next, however, is whether there are any differences in biological activity between the valproBDNF or metproBDNF precursors themselves. Both proBDNF and mature BDNF are present in the human CNS (Fahnestock et al., 2002; Michalski & Fahnestock, Mol. Brain Res., in press), both are secreted by transfected cells, including neurons (Egan et al., Figure 6E; Mowla et al., 2001), and both are capable of activating TrkB (Mowla et al., 2001). Thus, secreted proBDNF may exert some biological effects in the brain separate from its role as a precursor. Constitutively released proBDNF (val or met), which may regulate neuronal growth and survival for example, has yet to be investigated in the absence of mature BDNF.

    The discovery reported in this paper—that the regulated release of BDNF at synapses is compromised by the met polymorphism and contributes to BDNF’s effects on learning and memory—is exciting and significant. It is of note that there was no correlation between the met polymorphism and schizophrenia. A genetic association has been reported, however, between the val66 allele of the BDNF gene and bipolar disorder (Sklar et al., 2002; Neves-Pereira et al., 2002), and between other polymorphisms of the BDNF gene and Alzheimer’s disease (Kunugi et al., 2001; Ventriglia et al., 2002). This paper will surely encourage further examination of the role of BDNF in human cognitive function.


    . Neurotrophic factors and Alzheimer's disease: are we focusing on the wrong molecule?. J Neural Transm Suppl. 2002;(62):241-52. PubMed.

    . Biosynthesis and post-translational processing of the precursor to brain-derived neurotrophic factor. J Biol Chem. 2001 Apr 20;276(16):12660-6. PubMed.

    . Family-based association study of 76 candidate genes in bipolar disorder: BDNF is a potential risk locus. Brain-derived neutrophic factor. Mol Psychiatry. 2002;7(6):579-93. PubMed.

    . The brain-derived neurotrophic factor gene confers susceptibility to bipolar disorder: evidence from a family-based association study. Am J Hum Genet. 2002 Sep;71(3):651-5. PubMed.

    . A novel polymorphism of the brain-derived neurotrophic factor (BDNF) gene associated with late-onset Alzheimer's disease. Mol Psychiatry. 2001 Jan;6(1):83-6. PubMed.

    . Association between the BDNF 196 A/G polymorphism and sporadic Alzheimer's disease. Mol Psychiatry. 2002;7(2):136-7. PubMed.

  5. The study by Egan and colleagues deserves much attention because it attempts to track the role of BDNF in hippocampal function from human (!) memory traces down to the underlying molecular mechanisms. The observed polymorphism in position 66 of the pro domain, in fact, shows very interesting consequences for hippocampus-related memory tasks and hippocampal metabolism. Besides these very valuable advancements, some additional issues brought up by the authors would be interesting to focus on in future studies. Among these are:

    1. During the N-back memory task (which, as the authors mention, primarily relies on neocortical function), this study observes an unsual hyperactivation (i.e., hyperoxygenation) in the hippocampal region, thus indicating increased hippocampal neuronal activity in the met mutants. In contrast, the N-acetyl-aspartate measurements indicate a decreased metabolism of the val/met heterozygotes in the hippocampus. These experiments seem to indicate both increased and decreased BDNF-dependent hippocampal function, depending on the experimental protocol used. It will be interesting to determine which effect on hippocampal neuronal function is the more dominant one.

    2. Localization of overexpressed rat BDNF-GFP (which relates to the human val-BDNF individuals) in rat neurons has been shown to vary significantly between different cells of the same preparation (own unpublished results), and to show dissimilar subcellular localization in closely related neuronal preparations (compare e.g., Kohara et al., 2001 and Hartmann et al., 2001). Given this relatively large scatter of BDNF localization in seemingly similar neurons, it will be interesting to find out whether the met mutant is consistently mistargeted in all neuronal populations, and whether this effect is species-independent.

    3. Last, but not least, the authors seem to have expressed human BDNF in rat neurons. Although human BDNF is known to activate rodent BDNF receptors (i.e., TrkB and p75), perhaps we can’t just assume that human BDNF is targeted similarly to rat BDNF in these rodent neurons. This issue is of particular importance, given that even a single amino acid substitution in this study seems to change completely the cellular targeting of BDNF.

    In this respect, the paper by Egan et al. opens new avenues to linking BDNF biochemistry and BDNF-dependent neuronal function. However, additional studies (e.g., with rat BDNF mutated in the respective valine position) are clearly needed to settle the issue that targeting is ruled by the pro domain of these secreted molecules.


    . Synaptic secretion of BDNF after high-frequency stimulation of glutamatergic synapses. EMBO J. 2001 Nov 1;20(21):5887-97. PubMed.

    . Activity-dependent transfer of brain-derived neurotrophic factor to postsynaptic neurons. Science. 2001 Mar 23;291(5512):2419-23. PubMed.

  6. Egan et al. have performed an extraordinary study ranging from population to molecular levels. BDNF val66met polymorphism was shown first to affect episodic memory and hippocampal activation. Met-BDNF was revealed to fail in trafficking into secretory vesicles, resulting in the reduction of its release when synapses were depolarized. Thus, this SNP leads to dysfunction of BDNF in synaptic plasticity, causing memory impairment, although the biological activity of met-BDNF itself is as potent as that of val-BDNF.

    This story seems to be beautifully verified and convincing. BDNF SNPs have attracted much interest in other stories, too. The val66met polymorphism was reported to occur more frequently in patients with Parkinson’s disease (Momose et al., 2002). Other BDNF polymorphisms, such as 270C/T and 196A/G, were suggested to be associated with Alzheimer’s disease (Riemenschneider et al., 2002; Ventriglia et al., 2002). The pathological significance of BDNF SNPs should be pursued further.

    In the study by Egan et al., the met allele was revealed to be associated with poorer episodic memory, but not with altered semantic or working memory. This observation was made only with young subjects. If old met/met homozygotes were tested, the results could be different. As decreases in general cognitive function are seen in aging people, deficits in cognitive ability other than episodic memory may occur, and they might be more manifested in met/met homozygotes.

    BDNF is known to act as a factor for stimulating neuronal regeneration. The major effects of BDNF are ascribed to enhance axonal outgrowth and neuronal survival. Many studies have reported that BDNF promotes recovery from neural lesions (Blesch et al., 2001; Heibert et al., 2002; Ando et al., 2002). According to Egan et al., the release of mBDNF is much lower than that of vBDNF when synapses are depolarized. Therefore, mBDNF may be less efficient at neuronal restoration. Although Egan et al. mentions that mBDNF has the same potency for synaptic plasticity as vBDNF, it would be interesting to examine whether any delay in the recovery from neuronal damage such as ischemic insults could be seen in met/met homozygotes.


    . Association studies of multiple candidate genes for Parkinson's disease using single nucleotide polymorphisms. Ann Neurol. 2002 Jan;51(1):133-6. PubMed.

    . A polymorphism of the brain-derived neurotrophic factor (BDNF) is associated with Alzheimer's disease in patients lacking the Apolipoprotein E epsilon4 allele. Mol Psychiatry. 2002;7(7):782-5. PubMed.

    . Association between the BDNF 196 A/G polymorphism and sporadic Alzheimer's disease. Mol Psychiatry. 2002;7(2):136-7. PubMed.

    . Neurotrophism without neurotropism: BDNF promotes survival but not growth of lesioned corticospinal neurons. J Comp Neurol. 2001 Aug 6;436(4):456-70. PubMed.

    . Brain-derived neurotrophic factor applied to the motor cortex promotes sprouting of corticospinal fibers but not regeneration into a peripheral nerve transplant. J Neurosci Res. 2002 Jul 15;69(2):160-8. PubMed.

    . Animal model of dementia induced by entorhinal synaptic damage and partial restoration of cognitive deficits by BDNF and carnitine. J Neurosci Res. 2002 Nov 1;70(3):519-27. PubMed.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Run For Your Brain: Exercise Boosts Hippocampal Gene Expression, Neurogenesis

Paper Citations

  1. . Modulation of hippocampal synaptic transmission and plasticity by neurotrophins. Prog Brain Res. 2000;128:231-41. PubMed.
  2. . Neurotrophins as synaptic modulators. Nat Rev Neurosci. 2001 Jan;2(1):24-32. PubMed.
  3. . Proton magnetic resonance spectroscopy: an in vivo method of estimating hippocampal neuronal depletion in schizophrenia. Psychol Med. 1995 Nov;25(6):1201-9. PubMed.

Further Reading

No Available Further Reading

Primary Papers

  1. . The BDNF val66met polymorphism affects activity-dependent secretion of BDNF and human memory and hippocampal function. Cell. 2003 Jan 24;112(2):257-69. PubMed.