Hyperpolarization-activated current (I(h)) in vestibular calyx terminals: characterization and role in shaping...

Calyx afferent terminals engulf the basolateral region of type I vestibular hair cells, and synaptic transmission across the vestibular type I hair cell/calyx is not well understood. Calyces express several ionic conductances, which may shape postsynaptic potentials. These include previously described tetrodotoxin-sensitive inward Na(+) currents, voltage-dependent outward K(+) currents and a K(Ca) current. Here, we characterize an inwardly rectifying conductance in gerbil semicircular canal calyx terminals (postnatal days 3-45), sensitive to voltage and to cyclic nucleotides. Using whole-cell patch clamp, we recorded from isolated calyx terminals still attached to their type I hair cells. A slowly activating, noninactivating current (I(h)) was seen with hyperpolarizing voltage steps negative to the resting potential. External Cs(+) (1-5 mM) and ZD7288 (100 μM) blocked the inward current by 97 and 83 %, respectively, confirming that I(h) was carried by hyperpolarization-activated, cyclic nucleotide gated channels. Mean half-activation voltage of I(h) was -123 mV, which shifted to -114 mV in the presence of cAMP. Activation of I(h) was well described with a third order exponential fit to the current (mean time constant of activation, τ, was 190 ms at -139 mV). Activation speeded up significantly (τ=136 and 127 ms, respectively) when intracellular cAMP and cGMP were present, suggesting that in vivo I(h) could be subject to efferent modulation via cyclic nucleotide-dependent mechanisms. In current clamp, hyperpolarizing current steps produced a time-dependent depolarizing sag followed by either a rebound afterdepolarization or an action potential. Spontaneous excitatory postsynaptic potentials (EPSPs) became larger and wider when I(h) was blocked with ZD7288. In a three-dimensional mathematical model of the calyx terminal based on Hodgkin-Huxley type ionic conductances, removal of I(h) similarly increased the EPSP, whereas cAMP slightly decreased simulated EPSP size and width.