Neuromodulation (Tim O'Leary) Flashcards
Why are invertebrate CPGs useful?
- Stereotyped neurons means we can study the exact same neuron in different animals, and it will have the same properties and connectivity.
- CPGs are robust biologically (will function in vitro)
- In inverterates, we can identify neurons anatomically, genetically and functionally.
Crustacean STG - anatomy, connectivity
- Sits on top of the stomach in an artery from the pericardial organ - so can sense hormones and neuromodulatory bloodborne substances
- contains pyloric circuitry (AB-PD pacemaker, to LP, to PY. Constitutively active
- also contains gastric mill circuitry (controls teeth in stomach, so only need to be active when food’s in there
Every connection is inhibitory, either ACh or Glu.
Crustacean STG - activity
Can measure muscle activity in live lobster, by putting EMG wires in. Characteristic triphasic pyloric rhythm. PD muscle contracts, then LP, then PY, like a peristaltic pump
STG in vitro allows intracellular recording, triphasic rhythm remains. This is ‘fictive’ behaviour.
Other e.g.s of fictive behaviour
pre-Botzinger complex in mammalian brainstem - controls breathing rhythm. If you put it in a dish, rhythm remains. If you make it hypoxic, you get fictive gasps.
Thoracic ganglion in flying insects esp locust - can record in a tethered animal, you get reciprocal bursts that correspond to wingbeats. If you blow on the animal you activate the rhythm. Can de-afferent (so you remove any sensory input) and rhythm remains, but is slower, ‘wingbeat frequency less stable’.
Evidence of flexibility
Gastric rhythm only switches on when CoG neurons signal stomach stretch. This is short-term flexibility.
Buschges et al 1991 - removing hindwing tegulae decreases wingbeat frequency. 80% of animals recover normal motor function within 2 weeks. Removing forewing tegulae at this point causes a similar decrease in WBF as before. So there was functional reorganisation. This is long-term flexibility.
Cell-intrinsic properties and firing patterns that result
1) pacemaker activity - due to hyperpolarisation activating fast inward current, and deactivating a slow regenerative inward calcium current from T-type channels. Causes bursting. The frequency of bursting is dependent on exact balance of conductances.
2) Bistability - a small current pulse will not affect the cell, whereas a large current pulse will switch it into a persistent depolarised state, where it keeps firing rapidly until it ‘gives up’.
3) escape from inhibition - because hyperpolarisation deactivates various currents
4) rebound bursting - a cell is tonically active, doesn’t fire whilst it’s inhibited, then once the inhibition is removed it will burst before returning to its tonic activity
5) delayed rebound
How is a stable half-centre produced?
Sharp et al 1996 - took two follower neurons from two different crab STGs, put them in dishes. Connected using artificial inhibitory synapses, via a dynamic clamp, that allowed them to alter synaptic threshold.
Neither cell was a pacemaker, so this was a network rhythm, reliant on how quickly each cell could escape from inhibition.
Stable half-centre rhythm required that the synaptic threshold was within the voltage envelope of the oscillations themselves
Depolarising synaptic threshold initially increases period of oscillations, then decreases it.
Neuromodulators altering cell-intrinsic properties. Convergence and Divergence
Convergence - neuromodulators can converge on the same current, via different receptors
Marder, 1987 - pilocarpine and 5-HT both activate pacemaker
Swensen and Marder 2000 - multiple neuromodulators act on the same current in the isolated LP neuron
Divergence - neuromodulators will act on various targets and have various effects
DA will activate calcium current in LP neuron, but inhibit calcium current in the PD neuron
Marder and Weimann 1992 - removing descending input to STG slows or stops the pyloric rhythm. Applying various neuromodulators will start it up again, but each substance produces a different form of the classic triphasic rhythm. So different modulators can substitute for each other if the existence of a rhythm is more important than its exact form, especially if the NMJ acts as a temporal filter.
Neuromodulators altering network properties
In control saline, LP action potentials have no effect on PD neuron. After bath application of RPCH, LP APs cause IPSPs in PD.
Dickinson et al 1990 - RPCH strengthens cardiac sac IVN neuron synapses onto pyloric circuit neurons. This couples the two circuits together and produces a rhythm.
How do we prevent overmodulation?
1) Voltage dependency - I current is activated by pilocarpine, among others. but it is also voltage-dependent. So whilst it can alter membrane potential, it can’t push it outside an appropriate range. E.g. artificial proctolin current (by modelling the I properties) speeds up rhythm in LP cell but does not depolarise it, whereas constant current injection does.
2) Convergence - multiple neuromodulators act on the same current, so will occlude each other, causing a ceiling effect.
Marder and Meyrand 1989 - endogenous modulators will keep voltage within an appropriate range while speeding up rhythm and increasing amplitude. Exogenous modulators like nicotine will depolarise the cell, even though it’s acting on the same current. So endogenous modulators are tuned to the properties of the circuit and the cell and the current.
The gastric mill rhythm, and its coexistence with the pyloric rhythm
Gastric mill rhythm is only needed when there’s food in the stomach, which is detected by stretch receptors and transmitted in CoG neurons.
When CoG neurons are inative, the gastric circuitry exhibits a weaker following of the typical pyloric rhythm (which is triphasic).
When the CoG neurons are stimulated, gastric cells switch from simple follower cells into bistability, and show strong frequent firing that causes persistent contraction of the gastric mill teeth, then are inactive for a while. This occurs in a half-centre pattern, i.e. the gastric circuitry overall has a biphasic rhythm. The pyloric rhythm is maintained when this occurs.
Some neurons seem to show both rhythms, we don’t know why. Sometimes the teeth show pyloric-regulated ‘jittering’. We think this may be to help chew particularly tough food, since the pyloric rhythm is faster.
Weimann and Marder 1993 - bath application of endogenous polypeptides released from CoG neurons will turn on the biphasic gastric rhythm. This relies on network properties, but is set up by cell-instrinsic switch from follower cell to bistability.
What’s the basis for the variability?
Weimann, Meyrand and Marder 1991 - Took 49 crustacean STGs and measured from every identified cell from every one. Found some that only showed a pyloric rhythm (AB, PD, LP, PY), one that only showed a gastric rhythm (DG), and many that were gastric in some animals and pyloric in others. AM was 50/50. So the same neuron can have different outputs.
Golowalsh et al 1999 - voltage clamp experiment. found that the main conductances varied 3-5 fold between the same cell type in different preps. This means we can’t simply average out (Golowalsh et al 2002 took threemodels with the same output but differing conductances, produced a model with conductances that were the average of all those, and found a totally different output).
So maybe variability comes from differing conductances
Evidence for differing conductances
Using a computational model, we find that pacemaker activity is abolished with just a 20% change in conductances. Yet you can make changes this big as long as conductances covary.
Prinz et al 2004 - used a database approach, generating over 20 million models with different cellular and synaptic parameters. found ‘equivalent models’ which had different parameters but same output. 2.4% of models produced ‘pyloric rhythm’. 20% produced ‘pyloric-like’, which basically meant triphasic.
Schulz et al 2006 - used PCR to measure RNA levels of different channels in same neuron in different preps. Found that the different channels do have varying expression levels, and that RNA copy number correlated well with different electrophysiologically measured conductances, and that they covary. Additionally, this correlation was different in different cell types.
Evidence against differing conductances
Nowotsky et al 2007 - a critique of database approach. They took actual lobster pyloric circuits, put them in vitro, and applied 4-aminopyridine, a potassium channel blocker. Found that each prep responded in the same way, suggesting they all had the same channel density. Then took three ‘equivalent’ models from Prinz et al, and computationally reduced the I<a> conductance to 75%, then 50%. Found that the models reacted differently. Suggested that they were not, then, truly equivalent.
- counter-argument: animals respond differently to drugs all the time. This could be due to differing densities of their target channels. -
Also, an accurate model would need tens of parameters, each varying in tiny increments (because the pacemaker rhythm is so sensitive to tiny changes), and thus the database doesn’t cover enough.
Also, a single neuromodulator applied to a circuit in its standard experimental state will produce the same effect every time.
Also, the idea of hyperpolarising currents balancing out depolarising currents (and therefore they can covary with no effect on output) would require them to have the same activation and inactivation curves and timescale.</a>
Is there a limit to this variability? I.e. how do we keep circuit behaviour robust?
Altering the point in the phase where a synapse fires will alter the effect it has on time period of the next burst. This depends on the strength of the synapse, but the impact of synapse strength on this effect saturates (due to hyperpolarisation-activated inward currents). So to keep variability robust, you need a strong (aka big) synapse. This is expensive, since large currents need to pass. SO there’s a trade-off.
This means some variability in parameters (synaptic strength) won’t affect the activity. So variability is compatible with reliability.
