Ghrelin anticonvulsive properties: is it matter of desacylation?

We have read the article by Portelli et al. (2012) that thoroughly reviews the still scarce literature existing on the role of ghrelin and its receptor (GHSR1a) in epilepsy. As the authors recognize, clinical studies on the relationship between ghrelin and epilepsy are largely contradictory, since some of them show increased ghrelin levels, other no changes or even a decrease in patients affected by epilepsy: so, factors causing such variability need to be carefully addressed in future investigations. On the other hand, Portelli and coworkers look at data coming from animal studies as clearly demonstrating a role for ghrelin as anticonvulsant. Supporting this view, they cite two different studies showing that ghrelin is able to counteract the convulsive effects of pentylenetetrazole (Obay et al., 2007) or kainic acid (Lee et al., 2010), and a third one that reports negative findings both in pilocarpine and kainate models (Biagini et al, 2001). In our opinion, these discrepancies are only apparent, since Lee and coworkers (2010) reported beneficial effects not on seizure prevention, but on seizure severity, whereas we evaluated just seizure induction. More intriguingly, Obay and coworkers (2007) found that ghrelin dose-dependently increased latency to (but did not prevent) myoclonic jerks, tonic generalized extension and generalized clonic seizures. Duration of tonic generalized extension was also decreased by ghrelin, but duration of the initial myoclonic jerk was increased as well as the duration of clonic generalized seizures. Furthermore, a small effect on overall seizure duration was obtained only with the highest ghrelin dose. Alternatively, we have reported that GHSR1a ligands other than ghrelin prevent seizures in pilocarpine-treated rats, and that desacyl ghrelin increases the latency to first generalized seizures in the kainate model.A key observation to understand these discrepancies is that “…In human plasma, acylated ghrelin was found to disappear more quickly than total ghrelin, with elimination half-lives of 9-13 and 27-31 min, respectively (Akamizu et al., 2004)”, as cited by Portelli and coauthors (page 586 of the review). In spite that ghrelin was always administered 30 min before the tested convulsant (Obay et al., 2007; Lee et al., 2010), in no one of the mentioned studies circulating levels of ghrelin and desacyl ghrelin were assessed. In view of the probable conversion of ghrelin to desacyl ghrelin by butyrylcholinesterase (De Vriese et al., 2004) during latency to convulsions, it cannot be excluded that the anticonvulsive effects attributed to ghrelin could be due to desacyl ghrelin. Accordingly, the highly effective GHSR1a agonist JMV-1843 (also known as EP01572 or ARD-07), which has a better kinetics than ghrelin, in our hands failed to counteract the pilocarpine effects. We believe that the present evidence is still too poor to suggest a role for GHSR1a agonists in seizure prevention, at least in animal models. The same questions apply to clinical investigations, in which the balance between ghrelin and desacyl ghrelin is not clearly addressed.