Overall interpretation
A plethora of diverse data has been published over the
past several years in an attempt to explain through a
variety of mechanisms the efficacy of valproate in its
different clinical indications. These data support the
involvement of valproate in several pathways in which
it had not been previously implicated. Yet, reviewing
this information, it is remarkable that the number of
cellular targets that have been proven to be directly
affected by valproate is very small, and this number
becomes even smaller when considering the relevance
of some of these direct targets to epilepsy, bipolar
disorder or migraine, at least as we understand their
pathophysiology today.
There is little doubt that valproate directly inter-
feres with GABA metabolism to increase GABA
brain levels, and that this effect likely plays a
significant role in the immediate control valproate
exerts over epileptic seizures [5] and possibly also
mediates some early therapeutic effects of val-
proate in bipolar disorder [8] and migraine [11] . The
direct inhibition of GSK-3 by valproate, however, is
at best controversial [70] , and the inhibition of
microsomal long-chain fatty acyl-CoA synthase [42]
still awaits further reproduction and substantiation
of its relevance to human disease. Likewise, what
underlies valproate inhibition of brain myo-inosi-
tol-1-phosphate synthase activity, and whether this
is at all relevant to bipolar disorder therapy, still
needs elucidation [44] .
In contrast to the latter cellular targets, valproate-
induced changes in the expression of multiple genes,
mediated at least partially through the direct inhib-
ition of HDAC [18, 121, 122, 142] , have been
repeatedly demonstrated and may very well be
relevant to the therapeutic effects of this drug through
interference with intracellular signalling, e.g. the
inactivation of GSK-3 [73] , and neurotrophic and
neuroprotective effects, e.g. through the promotion of
BDNF expression [27] . Considering the large number
of genes whose expression is altered by valproate [108,
109] , it is reasonable to hypothesise that gene
expression changes plays not an insignificant role in
the long term effects of the drug. Future research
should try and assess for every newly discovered
valproate-induced cellular effect whether it is depend-
ent or independent of HDAC inhibition or AP-1 DNA
binding promotion by this drug. Nonetheless, many
genomic effects of valproate cannot be explained with
our current knowledge of its influences on AP-1 DNA
binding and HDAC inhibition, and additional mech-
anisms through which valproate, or its metabolites,
can affect gene expression should be sought.
In conclusion, influence of valproate at the genomic
level may provide insights into therapeutic effects
relevant to all three indications of epilepsy, migraine
and bipolar mood disorder [12, 124, 138] . The ��unified
field theory�� for the mechanism of action of valproate
in neuropsychiatric disorders possibly comprises acute
effects mediated essentially through the enhancement
of GABAergic transmission followed by a variety of
longer-term effects primarily resulting from gene
expression changes.