A neurohypophysial end game: spreading excitation with sildenafil

Neurotransmitter release is a function of the pattern and frequency of propagated action potentials arriving at the axon terminal, and central nervous system (CNS) synapses can exhibit a variety of forms of synaptic plasticity in response to this activity. Seminal studies at the frog neuromuscular junction and giant squid synapse revealed that the clustering of action potentials transiently facilitates transmitter release, primarily through enhanced Ca²⁺ signalling. This is evident at many mammalian synapses as well, but understanding the full complement of mechanisms that successfully code arriving action potentials has been hindered due to the small size of most CNS terminals (often ∼1 μm), which makes direct recordings difficult even with the advanced forms of optical guidance that characterize modern neurophysiology. A notable exception is the calyx of Held in the auditory system, a large terminal that can be recorded simultaneously with its postsynaptic partner.

An additional, more specialized case exists in the mammalian neurohypophysis, where the axonal swellings of hypothalamic oxytocin- and vasopressin-producing neurons can reach ∼15 μm. These allow direct terminal recordings with sharp electrodes from isolated neural lobes, as well as whole cell and single channel recordings from acutely dissociated terminals or from the visualized axonal swellings in neurohypophysial slices. We now have a detailed knowledge of terminal membrane ion channels in the neural lobe and their association with peptide secretion, the latter of which can be assessed with capacitance recordings to supplement the vast amount of data correlating the neuronal firing patterns of these neurons with peptide release in vivo and in vitro. Oxytocin and vasopressin neurons can both adopt bursting patterns, albeit very different in form, and this pattern is clearly the most efficient means of releasing a large amount of hormone.

Facilitation of peptide release, i.e. an increase in the amount released per spike, occurs during spike clustering and is closely tied to spike broadening, which in these neurons results largely from the frequency-dependent inactivation of a repolarizing, I_A-like K⁺ current. Broadened action potentials allow increased Ca²⁺ entry per spike from high voltage gated channels, and subsequently, increased stimulus-secretion coupling (Bourque, 1991; Jackson et al. 1991). But this facilitation is transient, and with sustained activity secretory fatigue sets in, the underlying mechanisms of which are incompletely known and undoubtedly multifaceted. Typically, fatigue is offset by a phasic bursting pattern, with recovery during silent intervals. Spike failures during prolonged repetitive activity contribute to fatigue at the neurosecretory terminal and by contrast, factors increasing this security could facilitate release by spreading excitation to a larger number of terminals. These failures are likely to occur at large varicosities with low safety factors (Zhang & Jackson, 1995).

In this issue of The Journal of Physiology, Zhang et al. (2007) bring an impressive and expertly employed array of tools to focus on a novel means by which blood-borne factors may affect neurohypophysial stimulus–secretion coupling. The paper stems from an earlier study from the same lab (Klyachko et al. 2001) demonstrating that activation of a BK-type Ca²⁺-activated potassium channel, critical for action potential repolarization in these terminals, is enhanced through the following sequence of events during spike trains: (1) spikes generate Ca²⁺ influx, initiating BK activation and increasing neuronal nitric oxide (NO) production; (2) NO activates soluble guanylate cyclase and cGMP production; (3) cGMP activates protein kinase G (PKG); and (4) PKG phosphorylates BK channels to shift their voltage dependence leftward, so that channels open earlier (at less depolarized potentials) during the spike depolarization. The enhanced BK activity increases spike afterhyperpolarizations, which help remove Na⁺ channel inactivation, increasing the likelihood that subsequent depolarizations will generate a spike. Thus, in parallel with its effect on exocytosis, Ca²⁺ influx during spike trains activates a chain of events that transiently secures spike fidelity.

BK channel enhancement is limited by the catabolism of cGMP by phosphodiesterase (PDE) activity, and therefore PDE inhibition can prolong facilitation by forestalling spike failure. The twist in the present study is that the specific PDE involved in the neural lobe is PDE5; potent inhibitors of PDE5 are the active components of the drugs widely used to address erectile dysfunction, Viagra® (sildenafil) and Levitra® (vardenafil). The study shows that PDE5 inhibition markedly prolongs BK activity enhancement, and correspondingly, decreases spike failures and increases their spatial propagation in the neural lobe. The authors then show that this enhancement leads to a threefold increase in oxytocin release to 25 Hz spike trains in isolated neural lobes. These effects are all dependent on NO production.

Thus, the actions of drugs commonly used to treat male sexual dysfunction may not be restricted to vascular targets such as the corpus cavernosa. Oxytocin is released in massive amounts during orgasm in humans and in other mammals, where it is thought to influence sperm transport, and this release would be enhanced in subjects taking PDE5 inhibitors. In addition, these inhibitors cross the blood–brain barrier and could affect CNS terminals that might utilize NO–cGMP signalling pathways, including a central oxytocin system that is known to be involved in maternal, social and sexual behaviour (Argiolas & Melis, 2004). Thus the present results, in combination with a mounting appreciation of the effects of PDE5 inhibition on a variety of physiological functions (Uthayathas et al. 2007), suggest neuronal targets deserve increasing attention in this regard.