On the whole, we humans are excellent at remembering our past experiences. This ability — known as episodic memory — mainly depends on the hippocampus, and many studies have addressed the issue of synaptic changes in this part of the brain that accompany memory-associated processes. Brain derived neutrophilic factor (BDNF) is a growth factor that has been implicated in memory-related synaptic transmission — it facilitates long-term potentiation and synaptic vesicle docking. However despite a substantial body of data from animal models, including transgenic mice, the link between BDNF and human memory has only now been established. The evidence comes from Egan and colleagues who use cohort tests of human subjects to show that a particular SNP at the BDNF locus affects human memory and hippocampal function. Using rat neuron tissue culture, the authors also show that these effects are brought about by the abnormal secretion and subcellular localization of BDNF.

Prompted by the association of BDNF with memory and learning in animal models, Egan et al. set out to investigate whether the only frequently found amino-acid-altering polymorphism in the human BDNF locus — a valine to methionine substitution at position 66 — would have similar effects. Various memory tests on Val/Val, Val/Met and Met/Met individuals showed that Met/Met individuals performed worse than other genotype groups in episodic memory tests but that other types of memory were unaffected. As might have been expected, only the Val/Val individuals had normal hippocampal physiology during memory tests. Although the mechanism that underlies this finding remains unknown, Egan et al. clearly demonstrated a qualitative difference in hippocampal response between the Val and Met BDNF alleles. Using magnetic resonance scanning, they showed that the Val66Met SNP is associated with reduced neuronal integrity and abnormal physiological activity in the hippocampus. The authors, interested in these underlying mechanisms, turned to in vitro experiments. A look inside the cultured hippocampal neurons told them that while Val-containing BDNF localized to cell bodies and distal processes in a punctate manner, the Met version was diffuse and found only in cell bodies. This, together with the result of secretion assays, prompted the authors to suggest that the Val to Met substitution significantly reduces activity-dependent BDNF secretion and prevents the correct synaptic localization of MetBDNF.

Having clearly shown the association between BDNF, memory and hippocampal function, as well as revealing some of the underlying cellular events, the authors end on a speculative note — although present in humans, the relatively deleterious Met allele is absent from lower primates, so, might it confer some as yet unknown compensatory advantage?