One cell, two bursting mechanisms. In vivo conditions change the in vitroburst in pyramidal cells of the ElectroLateral Lobe (ELL) of electric fish

The intrinsic mechanisms underlying burst generation in vitro where neurons are in relative isolation are generally well understood [1]. However, how these mechanisms are implemented under in vivo conditions where cells receive massive synaptic bombardment is still not clear. Pyramidal cells within the electrosensory lateral line lobe (ELL) of weakly electric fish have a well-defined burst mechanism in vitro (Fig. 1A), which is based on a somato-dendritic interaction [2]. Surprisingly, in vivo recordings from ELL pyramidal cells (Fig. 1B) do not show any of the characteristics associated with bursting found in vitro [3]. The goal of this project is to understand how in vivo conditions can give rise to these differences.

Figure 1

Bursting in the ELL pyramidal cells. (A) Representative trace of the in vitro recorded burst. Somatic spikes backpropagate into the dendrites, generating a dendritic spike which will move back into the soma, generating a depolarizing afterpotential (DAP), which will trigger a new somatic action potential The burst terminates with a characteristic doublet, followed by the burst after hyperpolarization (bAHP) (B) Representative trace of the in vivo recorded burst. Here there is no significant decrease in the interspike interval during a burst, bursts do not terminate with a doublet and the bAHP is absent. Also, the somatic spike shape recorded in vivo differs from the one recorded in vitro, with a pronounced hyperpolarization (AHP) following every spike within a burst.

One of the striking differences between in vivo and in vitro conditions is the absence of glutamatergic input to the cells in vitro which might provide the major source of Ca²⁺ to the cell via NMDA receptors. To test this hypothesis, we injected the calcium chelator, BAPTA, in pyramidal cell in vivo. The resulting removal of intracellular Ca²⁺ changed the cell bursting pattern to one characteristic of in vitro recordings.

To understand these observations, we have used a computational approach to propose a cellular mechanism for burst generation in vivo. In our computational model, which is based on the in vitro ghost-burst model, Ca²⁺ enters the cell through NMDA channels in the dendrites. When Ca²⁺ diffuses into the soma, it affects the Ca-activated potassium current. Gradual increases in Ca²⁺ concentration increases this current and eventually terminates the burst. This current also creates a spike shape characteristic to an in vivo burst, with a strong hyperpolarization after every spike within a burst.