Use-Dependent Inhibition of Synaptic Transmission by the Secretion of Intravesicularly Accumulated Antipsychotic Drugs

July 01, 2012

Antipsychotic drugs are effective for the treatment of schizophrenia. However, the functional consequences and subcellular sites of their accumulation in nervous tissue have remained elusive. Here, we investigated the role of the weak-base antipsychotics haloperidol, chlorpromazine, clozapine, and risperidone in synaptic vesicle recycling. Using multiple live-cell microscopic approaches and electron microscopy of rat hippocampal neurons as well as in vivo microdialysis experiments in chronically treated rats, we demonstrate the accumulation of the antipsychotic drugs in synaptic vesicles and their release upon neuronal activity, leading to a significant increase in extracellular drug concentrations. The secreted drugs exerted an autoinhibitory effect on vesicular exocytosis, which was promoted by the inhibition of voltage-gated sodium channels and depended on the stimulation intensity. Taken together, these results indicate that accumulated antipsychotic drugs recycle with synaptic vesicles and have a use-dependent, autoinhibitory effect on synaptic transmission.

  • Ethan Perlstein

    I’ll break the ice by discussing the first half of the
    paper, which describes a method to quantify the subcellular accumulation of
    four antipsychotic drugs (APDs) in rat primary hippocampal neurons based on the APDs’ ability to displace a fluorescent dye called LysoTracker Red, or LTR,
    which labels acidic compartments, including synaptic vesicles. The APDs studied
    were the phenothiazine chlorpromazine; the butyrophenone haloperidol; the atypicals risperidone and clozapine. Although
    structurally different, all four APDs are cationic amphipaths, though in the
    paper they are referred to as weak bases, the terminology of Rayport &
    Sulzer (1995).

    Don’t get me wrong, I’m a big fan of the weak base
    hypothesis. However, I’m leery of painting the four APDs in this paper
    with too broad a brush; and of focusing too much on their cationicity at the
    expense of lipophilicity. For example, chlorpromazine (CPZ) is 100-times more
    lipophilic, and its amine 10-times more basic, than the other three APDs. Does
    clomipramine, and tricyclic antidepressant structurally similar to
    chlorpromazine, also displace LTR in the rat hippocampal neurons? Elsewhere in the paper the
    authors measure extracellular APD concentrations in specific brain regions
    following a chronic dosing in whole animals. What about the concentration of
    APDs in the cellular membranes of those brain regions in chronically treated
    rats, which you’d expect would be higher for CPZ, the most lipophilic APD in the set.

    Stepping back, the use of LTR as a marker of APD accumulation
    is reasonable but imperfect: keep in mind that LTR is a very weak base (pKa =
    7.5) and about 10-times less lipophilic than the least lipophilic APD. Also, I would
    have liked to have seen a control showing that synaptic vesicle number and/or
    structural integrity was unaffected by APDs.

    I understand the dilemma here: chemical derivatization of
    APDs into fluorescent analogs would likely alter their binding and
    physicochemical properties, making it difficult to generalize back to the
    parent compounds. Radiolabeling is the least invasive alteration possible but
    detection is non-trivial, especially in live cells. One way out of the dilemma is
    to leverage the synthetic approach of false fluorescent neurotransmitters
    (FFNs) described by Sulzer and Sames several years ago. Such reagents would be
    superior to LTR as a proxy for APD accumulation in synaptic vesicles, and pH-sensitive FFNs are now available. Replication of this study will require a more direct assessment of APD accumulation
    before we can conclude precisely where and how much these drugs accumulates in human
    brain tissue over time.

    • dave_bridges

      What about the specific effects of the drugs on trafficking. For example, chlorpromazine is is used to inhibit endocytosis, for example , so effects on synaptic transmission will be complicated by that factor. Also if these things are not getting endocytosed, how are they being internalized?

      • Ethan Perlstein

        I agree. In the paper chlorpromazine appears to behave like the other antipsychotic drugs, but it’s much more lipophilic and basic, and so presumably its pharmacology is more complex. Richard Anderson’s group first showed that chlorpromazine affects clathrin adaptor localization (, so there’s still much more than meets the eye.

        We showed in my lab’s recent PLoS ONE paper that sertraline/Zoloft passive diffuses into cells. Maintaining a V-ATPase-driven pH gradient increases the uptake potential of any cationic amphipath; no endocytic internalization is required.

  • dave_bridges

    do you have a link to the paper or am i just blind?

  • Teja Groemer

    Chlorpromazine didnt have any effects on CME. We were aware of the evidence on this but we didnt observe this in our stimulation paradigms which didnt serve bulk endocytosis (50 AP 10 Hz).

    • Ethan Perlstein

      Just to be clear, Teja, we’re talking about Figure 4? I agree that the treatment regime you used may not have been sufficient in terms of time or dose to have an effect of clathrin-mediated endocytosis. What happens when you expose to APDs for longer than 15 minutes, say 1 hour? Or do a dose around 50µM, at least for CPZ, which is much more lipophilic than the atypicals.

      • Teja Groemer

        Well, we didnt try it. I personally think that 50uM is really a high dose. Once, we had titrated acutely cpz concentrations with pHluorin and in these concentration range the was a baseline-increase indicating intravesicular pH breakdown. This happens if over 10uM are applied.

        • dave_bridges

          So do you think the effects on CME are nonspecific at the higher doses?

          • Ethan Perlstein

            I would think yes.

        • Ethan Perlstein

          That makes sense. I also think 50µM CPZ is a high dose. And I agree that 10µM CPZ is probably the highest you want to go before the effects on membranes becomes net-destructive. In yeast the IC50 of CPZ is around 15µM, so that’s also consistent with your findings.

  • Ethan Perlstein

    After reading the paper carefully several times, I have a much better appreciation for why I think it’s a must-read for card carrying Pharmacologists. Although it’s behind a paywall, at least academics with access to university library subscriptions should give it a read.

    I was particularly intrigued by the electrophysiology results and the model of “self dosing.” Digging into the weeds a bit, the authors argue that haloperidol preferentially binds to the inactivated form of the sodium channel. But I’m not sure how they envision binding at the molecular level, and why is the binding site only available on the inactivated conformation of the channel? Could inactivated channels have a different set of protein-lipid interactions that are sensitive to drug accumulation, or is the binding site exclusively made of protein? In general, what is the extent of specific membrane partitioning by APDs, not just their extracellular concentrations in specific brain regions? Do the authors propose that APDs reside in the interfacial membrane region?

    All that gets to my broader question about membrane-based blind spots in this still excellent study. For example, the authors used haloperidol in most of the electrophysiology experiments, and I’m curious to know what the results would have looked like with chlorpromazine, which is 100 times more lipophilic. If chlorpromazine and haloperidol are truly interchangeable with regard to the electrophysiology results, than it suggests that the effect does not depend on membrane partitioning. But I would still argue that given enough time and a minimum dose or a high enough dose acutely, membrane effects would start to manifest as a function of the drug’s intrinsic lipophilicity.

    I have many more questions, but now it’s up to some other brave souls to jump in and keep the discussion going.


  • Andrew Morton

    OK, I’m finally taking the plunge. What’s the worst that can happen, I look stupid? Have you seen my profile picture?! @Ethan, first I just want to commend your efforts at catalysing discussion around papers in your sphere of interest. Your site is amazing and it must surely become some sort of gold standard for lab websites from now on! I’m sorry for not posting sooner. I’m chipping in not necessarily because I have anything radically insightful to say about the work, but more as a gesture of encouragement for efforts like this at “post-publication peer review.”

    As a quick disclaimer (for that, read “plug”), I co-authored the Preview piece in Neuron that accompanied Teja’s paper, with my PhD supervisor Mike Cousin ( Rather than go through that it seems better just to share a couple of more speculative points I had, but which we didn’t include.

    The activity-dependent scaling in efficacy of APDs seen in the study could effectively serve as a different mechanism of targeting drugs to where they are needed in the brain independent of receptor affinity, targeting circuits based on activity rather than possession of particular receptors… It would be interesting to incorporate drug effects with this mechanism into circuit/network models of pathophysiological activity patterns in schizophrenia.

    The pHluorin and calcium imaging experiments were performed in the presence of micromolar extracellular concentrations of APDs, based on the predictions from mathematical modelling that such concentrations would emerge in the synaptic cleft downstream of vesicular accumulation and release after exocytosis. Perhaps the killer experiment would be to mimic the hypothetical clinical situation by dosing neuronal cultures for longer periods of time with nanomolar (i.e. circulating plasma) concentrations of the APDs, chronically in the culture medium for a period of days/weeks. The authors’ model would be really strengthened if vesicular accumulation of APDs in this experimental design was sufficient to induce the same autoinhibition of presynaptic function that was described in the presence of micromolar extracellular concentrations of APDs.

    Cheers for now,