Walk-Through: Drugs and Optogenetics in the Network Module

If you have not already done the Network Walk-Through Warm-Up, I suggest you do that before proceeding, so that you are familiar with the general processes involved in building neural circuits.

The default circuit shows 2 neurons, with N1 making a type a spiking chemical synapse onto N2. You can check the properties of the synapse through the Synapses: Spiking chemical menu command, but, in short, it is a nicotinic cholinergic type.

Available drugs are listed in the Drugs group of the Experimental Control dialog. Each drug has an apply checkbox to the left of its name.

Here is some general information about drugs in this module:

Note that the Experimental Control drug options can be hidden through the Options: Configure View menu command if they are not relevant for an activity and you want to simplify the configuration.

The drug properties (normally available through the Drugs menu) can be hidden through the Options: Set puzzle menu command. The drugs are still available for use, but their properties cannot be seen or changed.

Synapse Effects

There are 10 editable rows in the dialog, for the 10 possible user-defined drugs.

The left column is headed Drug name, the remaining columns specify what the drug does. To be enabled for the user, the drug must have a name. If it has a name, but no other properties, it will appear in the Drugs list, but it will do nothing.

Note that Drug 1 (αBT: alpha bungarotoxin) affects type a spiking and non-spiking chemical synapses, and its Effect (right-hand column) is to Block. In reality, the snake venom αBT does indeed block nicotinic acetylcholine receptors, so this is appropriate. Drug 4 (ACh: acetyl choline) also affects type a spiking and non-spiking chemical synapses, but its action is to Activate. Since ACh is the natural transmitter for these synapses, this too is appropriate.

The new N2 trace in the results is completely flat because we have completely blocked the synapse.

N2 now spikes continuously because the ACh receptors are continuously activated.

Note: In Neurosim, receptors do not desensitize when an activator drug is applied.

Timed Application

Each drug has an Auto checkbox which, if checked, will mean that the drug is applied and removed at the specified On time and Off time. If the drug is already on when the simulation reaches the specified On time, the drug remains on. Similarly, if the drug is off at the specified Off time, the drug remains off.

Note that an asterisk appears after the drug name in the Drugs list of the Experimental Control dialog. This is a visual indication that the drug status may change during a simulation run without user intervention (if the Timed Application dialog were not showing, the user might not realize that Auto apply was active).

At this point, both the On time and Off time are set to 0 in the dialog, so ACh would be applied and immediately removed at the start of the simulation, and thus have no effect.

The simulation encounters the Off time at 0 s, but ACh was not applied at this time anyway, so nothing happens. The simulation progress until it encounters the On time at 0.04 s. At this point ACh is applied (and the ACh box in the Experimental Control panel is checked). Consequently the neuron depolarizes and spikes. There is no further Off time, so ACh remains applied for the rest of the simulation, and is still applied at the end. The drug box in the Experimental Control panel remains checked.

The simulation starts with ACh applied, but it immediately encounters the 0 Off time, and so the ACh is removed. It is re-applied at the 0.04 s On time.

The Result may not be quite as expected. At the start, the ACh is still applied from the previous run, so the On time of 0.04 s is ignored, and the ACh remains applied until the Off time at 0.08 s.

Now the simulation starts with ACh not applied. The ACh is applied at 0.04 s and removed at 0.08 s, so the Results now shows the effect of applying ACh just during the specified Auto apply times.

Hopefully, this gives you a good idea of how you can auto-apply drugs in the Network module.

Fraction Effect

In reality, drug effects often depend on the concentration of the drug when it is applied - low concentrations usually have weaker effects than high concentrations. This can be simulated in Neurosim as described next.

The circuit is similar to the default, except that N2 has been set to be non-spikingThe neuron was made non-spiking by unchecking the 'Use integrate-and-fire spikes' box in the 'Neuron Properties' dialog, accessed by double clicking the neuron. so that its spikes do not interfere with the drug effects. The synapse type a is still a nicotinic cholinergic type, but it does not show facilitationThe 'Relative facilitation' has been set to 1 in the 'Spiking Synapse' properties dialog, accessed from the 'Synapses: Spiking chemical' menu command.. In the Results view, the bottom trace shows the chemical synaptic conductance in N2. The conductance changes underlying the EPSP are obvious.

The right hand column is labelled Fraction effect. A value of 1 (the default) means that the effect is maximal (either total block, or total activation of all receptors), values less than 1 reduce the effect in proportion.

A new sweep occurs, and the synaptic conductance is completely blocked.

Another sweep occurs, but this time the block is partial. A Fraction effect of 0.7 means that 70% of the receptors are blocked, leaving 30% available to carry synaptic current. Hence the synaptic current is only about one third of that for the unblocked receptors, and the EPSPs are reduced in size but not abolished. The 30% can be regarded as the residual conductance fraction.

Multiple Blocking Drugs

You can specify more than one drug as a blocker of a particular synaptic type (although I would not normally recommend this when teaching).

The new drug should appear both in the Drug list in the Experimental Control dialog, and in row 5 in the Drug Timed Application and Strength dialog (assuming that you kept that open - if not, re-open it now).

A fourth sweep occurs, and now the synaptic conductance is very small, as is the resulting EPSP.

Activation and Occlusion

The N2 synaptic conductance immediately jumps to a high level, and N2 depolarizes.

Note: The post-synaptic conductance increase caused by maximum drug activation (Fraction effect = 1) is set heuristically at 2x the baselineThe baseline conductance and facilitation rate are set in the 'Synapse Properties' dialog, accessed through the appropriate Synapse sub-menu. synaptic conductance, with a further increase if the synapse facilitates. The increase is to allow for the possibility of temporal and spatial summation of the post-synaptic conductance.

When an activator drug is applied with a Fraction effect of 1, the conductance increase that would normally be caused by transmitter released from the pre-synaptic neuron is occluded - i.e. disappears. This is because all the post-synaptic receptors are assumed to be activated by the drug, so further release of transmitter from the pre-synaptic neuron has no effect.

Now the instant conductance increase caused by ACh application is smaller, and the pre-synaptic spikes cause a small further increase in conductance. This is because the receptors are not fully occluded with the weaker activator effect. This is meant to simulate a lower concentration of ACh.

Both the ACh and residual PSP conductance increases are reduced. This is because the blocking effect of αBT takes precedence over the activating effect of ACh.

Voltage-Dependent Channel Effects

Drugs that block voltage-dependent channels operate in a very similar way to those that block synaptic channels - they simply reduce the calculated channel conductance by the specified Fraction effect factor.

Activator drugs act differently. The key difference is that they are obligate factors. A voltage-dependent channel can be specified as needing activation, and if this is done, that channel will have 0 conductance unless an activating drug is applied. This allows a drug to act as a neuromodulator that "enables" a voltage-dependent channel, but it does not affect the voltage-dependency itself.

Three identical brief depolarizing stimuli are applied to N1, at one-second intervals. Each stimulus elicits a burst of spikes that long outlasts the stimulus that initiated it. This is because each stimulus generates a plateau potential due to a persistent sodium current, which is shown in the bottom trace. (If you need more information on the plateau potential mechanism, look at this tutorial.)

Note that this neuron has 3 voltage-dependent channels. The first two (K delayed rectifier and Na fast) are modified Hodgkin-Huxley channels, and produce standard spikes. The third channel, Na persistent, is a persistent sodium channel (i.e. it inactivates very slowly), and it is this that generates the plateau potential.

In the Results view, the spike bursts have disappeared, and each stimulus now just elicits a single spike (although you would need to zoom in the timebase to see these clearly). This is because we have specified that the persistent sodium channel requires activation for it to generate any current. Luckily, we have a drug (neuromodulator) that does just that: octopamine.

Octopamine is defined in the top row as activating Na-p channels, while in the second row the drug PRAX-562 is defined as a Na-p blocker. Note that PRAX-562 will not block the fast sodium channel because its short name is Na-f, but TTX will block both types of sodium channel because both short names contain 'Na'.

Now the first and last stimuli elicit single spikes but the middle stimulus elicits a plateau potential and consequently a burst of spikes. The burst occurs because octopamine is auto-applied just before the middle stimulus, and auto-removed about one second later. The application timing is indicated by the purple horizontalThis is visible because the Monitor box is checked in the Experimental Control dialog, and purple because that is the selected colour. line. The drug application enables the persistent sodium channel, which causes an immediate small depolarization because it is partially activated at rest. However, as soon as the stimulus occurs and generates a normal HH spike, the channel becomes strongly activated, and generates a long-lasting plateau potential, which leads to multiple HH spikes.

This pre-blocks the persistent sodium channel, and the plateau potential is abolished. The 3 stimuli now just elicit single spikes each.

TTX blocks both types of sodium channel, and so neither spikes nor plateau potentials are generated. The membrane potential changes generated by the stimuli are just passive depolarizations.

Optogenetics

A drug that activates an ion channel can simulate light in an optogenetic experiment (in which case, of course, the term "drug" is inappropriate).

The neuron is an endogenous burster that contains 3 voltage-dependent channels. Na (V) and K (V) produce standard HH-type spikes (but with modified kinetics), while the third channel, Ca (inactivating), is a low-threshold, slowly inactivating calcium channel. It is this channel that drives the oscillations responsible for bursting (this is the same neuron used in the tutorial on endogenous bursting).

In the Results view, the top trace shows the membrane potential, and the bottom trace shows the conductance of the calcium channel.

Assume that a genetic engineer has induced the neuron to express two types of channelrhodopsinsChannelrhodopsins are light-activated ion channels that originally derived from unicellular green algae, but through genetic engineering are now available in a variety of forms. Their expression can be induced in particular cell lines in many model organisms used in neuroscience research, such as the fruit fly or mouse. in its membrane. The first is opened by blue light and is anion (i.e. chloride) selective, and therefore inhibits the neuron when activated. The second is a non-selective cation channel (like a nicotinic acetylcholine receptor) that is opened by red light and which excites the neuron when activated. The aim is to modify the existing model neuron to incorporate these new channels. This can be done by adding new "voltage-dependent" channels, but without making them voltage-dependent. If you want to jump ahead and see the finished model, it is available as the file Optogenetics finish.

This adds a new channel (4), but by default it is an HH sodium channel ('Another fast Na channel'), which is not what we want.

At this point we have an anion channelrhodopsin embedded in the membrane.

This implants a new channel (5) which is identical to the one we just implanted.

We have finished implanting the two channelrhodopsins into the neuron. Now we need a way to activate them.

Now we have two "drugsThe fact that light is not a drug is covered by the "etc." in the group name." that can activate the channelrhodopsins. It would be nice to monitor the channelrhodopsin conductance directly.

You are now ready to run an experiment. The idea is to manually "switch on" first the blue light, and then the red light, while observing what happens to the neuron's activity.

If you want to simulate the effect of light at reduced (sub-optimal) intensity:

This has the same effect as would reducing the maximum conductance in the Voltage-Dependent Channel Properties dialog for the channelrhodopsin that the light activates.

You have now finished the walk through.

The final model is available as the file Optogenetics finish. The lights are switched on and off automaticallyAutomatic light switching is specified in the 'Timed Application and Strength' dialog, which is accessed through the 'Drugs: Timed application, strength' menu command., first the blue light, then the red light, so you can just Start the simulation and sit back and watch.

NEXT

Copyright © W. J. Heitler (2019)