Various experimental approaches have been used in the past to demonstrate positive effects of select antidepressant agents on the IP arm of PI signaling cascades [15, 27, 30, 66]. The present data demonstrate for the first time that antidepressant agents could increase by several-fold the production of CDP-diacylglycerol, a crucial signaling intermediate that is both a derivative of diacylglycerol and a precursor for the biosynthesis of PIs. This effect appeared to be common across agents from diverse chemical and pharmacological classes, seeing it was obtained with the tricyclics imipramine and desipramine, the SSRIs fluoxetine and paroxetine, the atypicals maprotiline and nomifensine, and the MAO inhibitors phenelzine and hydralazine. Thus, the findings could point to a mechanism (enhanced phospholipid biosynthesis) and mediator (CDP-diacylglycerol) for the biochemical and possibly clinical effects that may be common across diverse classes of antidepressants.
Earlier studies observed that several antidepressants enhanced [3H]IP accumulation and [3H]PI labeling in rat cortical slices [27, 29]. The mechanism of this response was confounding, seeing other studies that directly assayed phospholipase C activity suggested that the drugs could stimulate or inhibit PLC-mediated PI hydrolysis [15, 24]. In the present study, the antidepressants were equally effective in enhancing CDP-DG in the presence or absence of LiCl, whereas the presence of Li+ was necessary to demonstrate the effects of the drugs on IP accumulation. This implies that the compounds do not inhibit IP breakdown (otherwise they would have substituted for Li+), and that their effects on IP accumulation is probably secondary and passive to the enhanced production of upstream CDP-DG and PI substrates.
Numerous agents acting at diverse receptor systems can enhance PI metabolism, but few such pure receptor agonists are known to exhibit an antidepressive effect in humans or animals . How then might an effect of antidepressant agents on CDP-DG and PI synthesis be associated with the antidepressive efficacy of the compounds? An attempt to address this question led to comparisons of the ratio data between the antidepressants as a group and agonists at alpha-adrenergic, 5HT2 serotonergic, and dopaminergic receptors (which are implicated in depression) as well as the muscarinic cholinergic receptor that is not known to contribute to the actions of antidepressant agents. While all these receptors are coupled to PI hydrolysis, only some 5HT2 agonists and SKF38393 have been shown to elicit antidepressive effects in rodent models [67–69]. The ratio of CDP-DG to IP components for phenylephrine and carbachol decreased, for α-methylserotonin remained unchanged, and for SKF38393 increased, with increasing concentrations of agonist. For the antidepressant agents, the ratios were not only significantly elevated, but actually increased in a concentration-dependent fashion. This was true of all classes of antidepressants examined, including the effective members among the MAO inhibitors. The similarity between the effects of SKF38393 and the antidepressants may underlie the behavioral antidepressant efficacy of the compound as previously demonstrated in the rodent model . Hence, to the extent that CDP-DG production might be relevant to depression or the mechanism of antidepressant drug action, an antidepressive agent should not merely increase PI hydrolysis, but the CDP-DG produced must exceed and probably precede the production of PI messengers.
It was noteworthy that even the SSRIs produced significant increases in each CDP-DG ratio, whereas the direct 5HT2 agonist α-methylserotonin did not. If the biological actions of the SSRIs were limited to the actions of the drugs to enhance synaptic serotonin levels, then one would expect direct serotonin receptor stimulation to elicit similar effects. This disparity should suggest that facilitation of synaptic serotonin levels may not be the sole or sufficient mechanism of action of the SSRIs. Rather, antidepressant-enhanced synaptic serotonin may work in concert with antidepressant-facilitated neurolipid biosynthesis to achieve the type and level of downstream signaling responses that may contribute to the antidepressive effect.
Although CDP-DG production induced by the antidepressants may be derived from phosphoinositide breakdown, it is not impossible that the CDP-DG pool may be generated from additional endogenous sources. In an initial attempt to address this, we observed that antidepressant-mediated accumulation of [3H]CDP-DG was completely blocked by the non-specific phosphoinositide inhibitor, neomycin, in prefrontal or hippocampal tissues. Conversely, the selective PLC inhibitor U73122 could only partially decrease CDP-DG production, while it completely blocked the release of IPs. These results suggest that, while the integrity of the phosphoinositide pool is essential to the full effect of antidepressant agents on CDP-DG (based on the neomycin data), the PLC-accessible pool of phosphoinositides may not be the only source of antidepressant-mediated CDP-DG production (based on the U73122 data). It is known that neural (and other) cells maintain multiple pools of phosphoinositides not all of which may be accessible to PLC-mediated cycling. The possibility that the antidepressants could mobilize these additional reserves of PI substrates, particularly following acute or chronic metabolic depletion of the substrates, should be an interesting subject for future investigations.
A critical question that was also attempted relates to the extent to which the CDP-DG response may be specific to antidepressant agents versus other psychotropic drugs. After testing a wide range of compounds, we observed that neither the antipsychotics chlorpromazine and haloperidol, nor several other psychotropic agents were capable of inducing the degree of CDP-DG effects observed with the antidepressant agents. While this suggests that the CDP-DG effect, particularly the dose-related effect on CDP-DG/IP ratio, could reflect a characteristic property of antidepressant medications, we were equally surprised by the disparity in efficacy among the MAO inhibitors. The ineffective agents included the nonselective MAO A/B inhibitor clorgyline, and the selective MAO-B inhibitors pargyline and selegiline. At least one of these, clorgyline, is used clinically for the treatment of depression. Conversely, other MAO inhibitors, including phenelzine, hydralazine and tranylcypromine were significantly effective in inducing CDP-DG. The chemical or biological basis for this disparity among the MAOIs is still unclear. Indeed, considering the relatively marked effects of phenelzine and hydralazine, it is possible that inhibition of MAO-mediated monoamine breakdown may not be the predominant mechanism by which these compounds modulate CDP-DG signaling.
It remains to be determined how the present in vitro observations may relate to in vivo drug concentrations or behavioral effects. While all tested antidepressants generally induced significant CDP-DG or PI effects at concentrations of 1–10 μM (and as low as 0.1–0.3 μM for phenelzine and hydralazine), the in vivo concentrations or doses needed to induce comparable effects have not been determined. Nevertheless, a recent report suggests that antidepressants may induce in vivo CDP-DG or phosphoinositide effects at doses commonly used to elicit antidepressant-like behaviors in animal models . Additional studies should help to clarify these questions.