VGB is particularly effective for seizures in TSC patients, but the mechanism of this unique relationship between VGB and TSC is poorly understood. In this study, we show that VGB strongly inhibits seizures in a mouse model of TSC. Consistent with its known mechanisms of action, VGB causes the expected increase in brain GABA levels in the KO mice. Furthermore, as a novel and unexpected finding, VGB inhibits mTOR pathway activity, which could represent an additional mechanism of action that may contribute to the distinctive efficacy of VGB in TSC.
VGB appears to have unique therapeutic properties in TSC for a couple of reasons. First, medical intractability is especially common in epilepsy in TSC, occurring at a much higher rate (~65%) in epilepsy in TSC than in epilepsy overall (~30%) 
. However, in contrast to most other antiseizure medications, VGB is particularly effective for seizures in TSC, especially infantile spasms, a typically devastating type of childhood seizure. VGB has been reported to eliminate spasms in about 95% of cases, which is much higher than all other treatments for spasms due to TSC or any other cause 
. In fact, this special relationship between VGB and TSC is one of the few documented examples of a medication having individualized efficacy for a specific epilepsy syndrome or cause of epilepsy. Furthermore, early treatment with VGB has been reported to improve the long-term outcome of neurological development and epilepsy in TSC patients 
, although additional controlled clinical trials are needed to confirm these findings. In the present study in a mouse model of TSC, VGB was very effective in inhibiting epilepsy, causing almost complete suppression of seizures. The efficacy of VGB was comparatively better than other standard antiseizure medications in Tsc1GFAP
CKO mice, such as phenytoin and phenobarbital, which reduced seizure frequency by 55–68% 
. Thus, this mouse model represents an appropriate system to investigate the mechanisms of action of VGB in TSC.
The standard, accepted mechanism of action of VGB is inhibition of the catabolism of GABA, the major inhibitory neurotransmitter in the brain 
. Irreversible inhibition of GABA transaminase by VGB leads to elevated brain GABA levels 
. The resulting potentiation of inhibitory GABA systems in the brain is a rational explanation for the antiseizure effects of VGB. However, the effects of VGB on GABA physiology is not straightforward, as VGB causes a paradoxical reduction in evoked fast inhibitory postsynaptic potentials of neurons but instead potentiates tonic GABAergic inhibition 
. Furthermore, enhancement of GABA signaling by increased availability of GABA may not explain the unique effectiveness of VGB for epilepsy in TSC, as other GABA-potentiating drugs, such as barbiturates and benzodiazepines, do not show comparable efficacy in TSC. In Tsc1GFAP
CKO mice, a previous study found no abnormalities in GABAergic synaptic transmission 
and the present study demonstrates that baseline GABA levels in neocortex and hippocampus of the KO mice are similar to control mice. While Tsc1GFAP
CKO mice did show a better response to VGB in GABA levels in neocortex than control mice, this was a modest difference, and the effects of VGB on GABA levels in hippocampus were similar in control and in Tsc1GFAP
CKO mice. Thus, other mechanisms of action independent of GABA may need to be considered in order to account for the unique efficacy of VGB for seizures in TSC.
Recent evidence suggests that abnormal mTOR pathway activity is critical for epileptogenesis in TSC and that mTORC1 inhibitors, such as rapamycin, may be particularly effective therapies for epilepsy, not just as an anticonvulsant to treat seizures, but also as antiepileptogenic therapy to prevent epilepsy in TSC 
. mTOR is a protein kinase, which normally regulates a number of important functions, such as cell growth, proliferation, metabolism, and protein synthesis, many of which could affect epileptogenesis under pathological conditions. Thus, the possibility that VGB could exert at least some of its effects on epilepsy in TSC via interaction with the mTOR pathway is an intriguing, but previously untested, hypothesis. In the present study, we have provided evidence that VGB can inhibit the mTOR pathway. First, mTORC1 pathway activity, as reflected by downstream S6 phosphorylation, was decreased by VGB administered to Tsc1GFAP
CKO mice in vivo. It is important to note that other signaling pathways, such as ribosomal s6 kinases (RSK) and mitogen-activated protein (MAP)/extracelluar-signal-regulated (ERK) kinases, may also modulate S6 phosphorylation independent of mTOR. However, this mTORC1-independent phosphorylation appears to only involve the Ser235/236 phosphorylation site of S6, not the Ser240/244 site 
. Thus, our finding that VGB decreases Ser240/244 phosphorylation supports the involvement of the mTORC1 pathway. As seizures themselves can cause mTOR activation 
, one simplistic explanation for the apparent mTOR pathway inhibition by VGB in the TSC mouse model is that VGB's inhibition of seizures secondarily decreased mTOR activity. However, this possibility is basically excluded by the similar findings of mTOR pathway inhibition by VGB in normal control mice in vivo
, as well as in cultured astrocytes from both control and Tsc1GFAP
CKO mice. By comparison, the GABA potentiating drug, phenobarbital, inhibits seizures, but has no effect on mTOR activity 
. Furthermore, the inhibition of mTOR pathway activity by VGB in cultures in vitro
also eliminates other confounding factors in the brain in vivo
, suggesting that VGB directly inhibits the mTOR pathway.
The prototypic mTORC1 inhibitor, rapamycin, has potent effects in preventing epilepsy, prolonging survival, and blocking associated pathological and molecular changes that promote epileptogenesis in Tsc1GFAP
CKO mice, as well as in other mouse models of TSC 
. In the present study, in addition to seizures, VGB also had some inhibitory effects on glial proliferation, at least in hippocampus. As VGB inhibited P-S6 in both hippocampus and cortex, a similar effect of VGB on astrocyte number likely also occurred in neocortex, but the sample size may have been underpowered to reach statistical significance. Such effects of VGB on glial proliferation can likely be attributed to mTOR pathway inhibition. However, the effects of VGB were not as strong or extensive as rapamycin, as VGB only had modest effects on survival and glial proliferation and no effect on neuronal organization. Since VGB was effective in inhibiting seizures, this suggests that the pathological abnormalities in these mice are not secondary to ongoing seizures. As rapamycin is a very potent mTORC1 inhibitor 
, the difference in the effectiveness of rapamycin and VGB may be due to the milder suppression of mTOR activity by VGB in vivo, which maximized at approximately 60% of the levels of vehicle-treated Tsc1GFAP
CKO mice (). Given these differences between VGB and rapamycin, it is possible that VGB does not directly inhibit the mTOR kinase itself like rapamycin, but may also involve other components of the mTOR pathway. Future studies are needed to determine the specific molecular mechanisms by which VGB regulates the mTOR pathway.
Other limitations of this study include issues intrinsic to the mouse model. While there are now a variety of animal models of TSC, involving inactivation of Tsc1
at different developmental time points and in different subsets of brain cells, there is no perfect model that recapitulates all neurodevelopmental features of TSC. Tsc1GFAP
CKO mice involve primarily inactivation of Tsc1
in glial cells, although a subset of neurons is also affected. The mechanism of action of VGB in TSC may depend on the cell type(s) affected, but this issue is not addressed with this one model of TSC. In addition, in patients with TSC, VGB is most effective against infantile spasms. Neither Tsc1GFAP
CKO mice nor any other animal model of TSC have been documented to have spasm-like seizures. Interestingly, however, rapamycin has been shown to selectively suppress spasms in a non-TSC rat model of infantile spasms 
. Finally, the present study has not determined the relative contribution of GABA potentiation and mTOR pathway inhibition in decreasing seizures. Future studies need to define in more detail the specific cell types, seizure types, and specific mechanisms involved in VGB's effect in TSC. Despite these current limitations, the present study is significant in identifying a potential novel mechanism of action of an antiseizure medication involving the mTOR pathway. This interaction of VGB with the mTOR pathway may account for the unique efficacy of this drug for a common genetic epilepsy.