Gap junction function in the nervous system

[Pythagorean]

In part branching off a discussion here, but also a discussion I've always been interested in having. I also want to take some time to answer the question in that thread:

What sort of phenomena or behaviour do axo-axonic gap junctions produce?

axo-axonic gap junctions in the hippocampus are thought to play a role in generating ultra-fast oscillations that have a unique "spikelet" shape that participates in some higher level functions [1]:

Recent studies both in vivo and in vitro have revealed rhythmic, synchronous population activity, such as gamma frequency (30–80 Hz) oscillations 34., 87., 98. and 121. and ultrafast sharp-wave “ripples” (>80–200 Hz) 31. and 124., in mammalian brains. These oscillations are thought to play an important role in a variety of cognitive processes, including memory formation, sensory perception, and other higher functions 44. and 87.. Evidence for electrical signaling between homogeneous populations of neurons involved in generating these rhythms is strong. Coupling was initially demonstrated between principal (pyramidal) cells 61., 62. and 101. but also, more recently, between interneurons 5., 36., 39., 40., 50., 64., 65., 99., 100. and 117.. A number of reviews have detailed the properties of gap junctions 8., 9., 28. and 29., their role in interneurons 37. and 102., and network activity, especially during development 25. and 79., and epilepsy 20., 32., 51., 72., 107. and 112.. The aim of this review is to discuss the importance of electrical signaling in synchronizing network activity. In particular, we will focus on two types of rhythmic activity observed in the hippocampus both in vivo and in vitro—gamma frequency oscillations and ultrafast oscillations.

Axo-axonic gap junctions are also capable of antidromic action potentials [2] which can give alternate routes of excitation and inhibition (a gap junction, of course, acts somewhat like an inhibitor with respect to the neuron at the higher potential).

and address this comment:

Ephaptic coupling including electrical gap junctions are prevalent in the brain but are more of an anomaly than a selected design feature, as far as we can tell. There doesn't seem to be any conserved pattern of any type of ephaptic coupling across brain taxa that would suggest it has any specific role in sensory-motor information processing, at least.

Firstly, I'd just like to point at that it's true that ephaptic coupling, but electrical coupling through gap junctions is not considered ephaptic coupling [3]. Ephaptic coupling is coupling through environment, such as local electric fields and local ion exchange with extracellular space, not direct gap junction coupling.

The well known gap junctions role is synchronization [4], but they also pass molecular signaling molecules, leading one author to refer to them as the "rosetta stones" of biology (because they can integrate electrophysiological and metabolic communication) [5]. Electrical synapses are also found extensively coupling GABAergic interneurons in the cortex, and are thought to participate in coincidence detection across inhibitory signals [6]. They also appear to functionally segregate portions of network [7]. Experiments in C.elegan show that interfering with different kinds of gap junctions can lead to numerous different functional deficiencies (from constipation to chemotaxis to death) [8]. And of course, the nervous system is dominated by gap junctions in early development [9].

The chief blockers and openers of gap junctions (because gap junctions can be open, rectifying, or closed) are carbenoxolone and trimethylamine.

[1] http://www.sciencedirect.com/science/article/pii/S0361923003002302
[2] http://www.sciencedirect.com/science/article/pii/S0306452298007556
[3] http://www.nature.com/neuro/journal/v14/n2/abs/nn.2727.html
[4] http://www.sciencedirect.com/science/article/pii/S0166223699014976
[5] http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3058197/
[6] http://www.nature.com/nrn/journal/v2/n6/abs/nrn0601_425a.html
[7] http://www.sciencedirect.com/science/article/pii/S0166223605000998
[8] http://www.ncbi.nlm.nih.gov/pubmed/24575048
[9] http://www.jneurosci.org/content/3/4/773.short

 

[DiracPool]

In part branching off a discussion here, but also a discussion I've always been interested in having. I also want to take some time to answer the question in that thread: What sort of phenomena or behaviour do axo-axonic gap junctions produce?

Back in the day, roughly around the turn of the 20th century, all communication between neurons was thought to be electro-electrical, or sort of a "continuous" stream through neurons, essentially equivalent to "gap junction-ish" dynamicism. It was largely through the pioneering work of investigators like Ramón y Cajal, Golgi, and Brodmann, who developed microscopic histological techniques and selective neuron staining methods that the chemical synapse was identified.

Through many more studies over the decades, it was found that the best way to model brain function was as a "contiguous" network of individual neurons communicating through chemical synapses rather than a continuous network communicating through the equivalence of what could be termed gap junction connectivity.

Referring back to the previous thread you referred to, what is the evidence that brain and behavior is regulated through chemical synapses and not gap junction dynamics? Well, we can start with the entire psychopharmeucetical industry and the literature it embodies. Pharmico-kinetics works primarily through synaptic receptor inhibition, agonism, reuptake inhibition, second messenger system potentiation via neuromodulation, etc. As far as motor expression is concerned, twitch fiber synchronization it is regulated by monoamine dopamine in the frontal cortex and basal ganglia, and by acetylecholine in the motor effectors, each of which work through the chemically modifiable synapse.

And that is the key, modifiable. From what I know of gap junctions or ephaptic coupling mechanisms in general, there is no mechanism for the generation of a modifiable synapse that could be seen to affect learning or behavior in any sort of significant or salient plastic manner.

With the 10's of thousands of articles on the subject going back over the decades, of course your going to find a few articles raising a question as to whether other, not so frequently considered, elements of brain function could be making an important contribution. These outlier models crop up in neurobiology just as they do in theoretical physics. And, just as in theoretical physics, they typically stick around for years without shifting the existing zeitgeist, and that's typically because they are just wrong.

IMHO, this is what we are seeing here in these threads with the gap junction argument, and especially with madness's introduction of the participation of astrocytes, which are glial cells, in the participation of brain dynamics. These ideas, again, typically crop up when progress in a certain field stagnates and researchers look for alternative explanations. The glial cell contribution to brain function has been around for well over a decade now and I'm sure one could pull out a few articles arguing for it. And the thing is, yes, they do participate in brain function because they support the neurons that are actually doing the real work. Do they do real work themselves? Maybe in a sense. But the important point is that that contribution is negligible relative to the massive and significant role in brain dynamics that is accomplished through actual neurons and their electrochemical synapses.

Again, as I said in the previous thread, the only behaviorally significant effect to brain dynamics that gap junctions seem to confer are a contribution to local synchronization effects in neuronal tissue who's relevance is still controversial.

 

[Pythagorean]

what is the evidence that brain and behavior is regulated through chemical synapses and not gap junction dynamics?

I don't agree with this implied mutual exclusiveness. Certainly, chemical synapses dominate brain dynamics (in adults) but the point here is that gap junctions still play important functional roles and are not as trivial and insignificant as ephaptic coupling. You don't have to convince me of the evidence for chemical synapses playing a role.

From what I know of gap junctions or ephaptic coupling mechanisms in general, there is no mechanism for the generation of a modifiable synapse that could be seen to affect learning or behavior in any sort of significant or salient plastic manner.

The hemichannels of gap junctions are made of innexins, connexins, and pannexins (depending on species) and the individual hemichannels can be modulated to open, close, or rectify (only allow current through one direction) . Further, individual cells can regulate the amount of gap junctions present to change the coupling strength. [1][2]


[1] http://www.ncbi.nlm.nih.gov/pmc/articles/PMC307687/
[2] http://www.sciencedirect.com/science/article/pii/S0005273612001848

Do they do real work themselves?

This approaches the no true scottsman fallacy. You can start dividing functions into "real" and "not real" and then anytime a function is demonstrated, you can just say "but that's not real work". It's not a very productive approach to discussion, imo.

 

[Pythagorean]

As for astrocytes, my understanding is that it's known accepted that they release D-Serine (a necessary co-agonist for activation of NMDA receptors) [1] and can also release the neurotransmitter glutamate at the tripartite synapse [2]. They've also been shown to regulate Fabp7 and PSD through transcription [3].

[1] http://www.ncbi.nlm.nih.gov/pubmed/23485803
[2] http://www.ncbi.nlm.nih.gov/pubmed/16221850
[3] http://www.ncbi.nlm.nih.gov/pubmed/18286188

But perhaps we should start an astrocyte thread as well :)

 

[DiracPool]

This approaches the no true scottsman fallacy. You can start dividing functions into "real" and "not real" and then anytime a function is demonstrated, you can just say "but that's not real work". It's not a very productive approach to discussion, imo.

I didn't say that they don't do any real work. I said maybe they do in some tangential sense. My theme was about "preponderance of the evidence" for what does the "lions" share of real work, of you re-read it.

Do they do real work themselves? Maybe in a sense. But the important point is that that contribution is negligible relative to the massive and significant role in brain dynamics that is accomplished through actual neurons and their electrochemical synapses.

 

[DiracPool]

But perhaps we should start an astrocyte thread as well :)

See, this is exactly the point I was trying to make. Now I'm in a position, as in the previous thread, where I'd have to prove a negative to win the argument, which, of course I can't do. That argument being that gap junctions or glial cell activity have never participated significantly in any sort of brain process.

So, I'll just stick to my above argument, for the record.

Although, I'd have to say that I didn't know C. Elegans got constipated?! Who would have guessed?

Experiments in C.elegan show that interfering with different kinds of gap junctions can lead to numerous different functional deficiencies (from constipation to chemotaxis to death)

I'd say that enough of a reason right there to keep your gap junctions running effectively.

 

[Pythagorean]

I think what is interesting about the contribution of things like astrocytes* and gap junctions is that their contribution to processing (not support roles) is an open question and there's lots of evidence suggesting it's worth investigating. I guess chemical synapses are boring to me. I do find the diversity and classifications of GABAergic interneurons in the cortex very interesting.

I also think an interesting aspect of biology, in general, is how support roles aren't cleanly separable from processing roles. For example (in addition to electrical synapses and astrocytes), ATP functions both as an energy source (support) and a signaling molecule (processing).

*astrocytes are the only glial cells I know of that have significant evidence for their contributions to processing.

 

[madness]

As for astrocytes, my understanding is that it's known accepted that they release D-Serine (a necessary co-agonist for activation of NMDA receptors) [1] and can also release the neurotransmitter glutamate at the tripartite synapse [2]. They've also been shown to regulate Fabp7 and PSD through transcription [3].

[1] http://www.ncbi.nlm.nih.gov/pubmed/23485803
[2] http://www.ncbi.nlm.nih.gov/pubmed/16221850
[3] http://www.ncbi.nlm.nih.gov/pubmed/18286188

But perhaps we should start an astrocyte thread as well :)

Have you seen this? http://www.pnas.org/content/early/2014/07/23/1410893111.abstract

Seems like pretty strong evidence for the role of astrocytes in circuit dynamics and cognition.

 

[Pythagorean]

Have you seen this? http://www.pnas.org/content/early/2014/07/23/1410893111.abstract

Seems like pretty strong evidence for the role of astrocytes in circuit dynamics and cognition.

I didn't read the paper, just the abstract. I agree somewhat with Dirac Pool about something like this:

By creating a transgenic mouse in which vesicular release from astrocytes can be reversibly blocked, we found that astrocytes are necessary for novel object recognition behavior and to maintain functional gamma oscillations both in vitro and in awake-behaving animals.

Now, they may show how the role is a processing role and not a support role, but from just the sentence above, it could very well be a support role, and without the support of the astrocytes, the neuron responsible for the activity can't function.

The paper that sticks out that I have read on astrocytes modulating glutamate release at the tripartite synapse carefully went through each process in the chain (I believe they found it effected nr2b receptors on the postsynaptic cell, leading to LTP in response to ATP signaling on the astrocytes y2p1 receptors). For each process, they presented their evidence.

I believe this is the paper, but not on a privileged account currently:
http://www.nature.com/neuro/journal/v10/n3/abs/nn1849.html

 

[madness]

Now, they may show how the role is a processing role and not a support role, but from just the sentence above, it could very well be a support role, and without the support of the astrocytes, the neuron responsible for the activity can't function.

Sounds like a false dichotomy to me. Can you provide a suitable definition of support versus processing roles? Couldn't similar arguments be made about interneurons etc.?

 

[Pythagorean]

There's no dichotomy implied. In fact, earlier in the thread:

support roles aren't cleanly separable from processing roles

The point is not that it's one or the other. The point is that evidence for processing needs to be presented; for example, if you remove all the blood from an organism and present it with a math test (which it then fails, having no blood) then conclude that blood is important to mathematics, you'd be technically correct, but it wouldn't be a very meaningful statement. Blood doesn't participate in the processing itself, it merely makes sure other cells have the resources they need to do the processing.

 

[madness]

There's no dichotomy implied. In fact, earlier in the thread:



The point is not that it's one or the other. The point is that evidence for processing needs to be presented; for example, if you remove all the blood from an organism and present it with a math test (which it then fails, having no blood) then conclude that blood is important to mathematics, you'd be technically correct, but it wouldn't be a very meaningful statement. Blood doesn't participate in the processing itself, it merely makes sure other cells have the resources they need to do the processing.

Right, but again you need to provide a definition of support vs processing, otherwise it's not possible to provide evidence.

 

[Pythagorean]

Right, but again you need to provide a definition of support vs processing, otherwise it's not possible to provide evidence.

It's actually quite common terminology in the literature. It often boils down to the debate of whether the role is "just supporting" or not. That is the nature of DiracPool's criticism. It's not a new concept.

Briefly, processing serves to relay information, whereas support is the provision of energy, structural integrity, or components for building proteins. Of course, there is no dichotomy... for instance, actin provides structural support for the cell, but actions on actin (breaking of the structures) can also trigger actin constructing and deconstructing processes (a signaling role).

 

[madness]

Briefly, processing serves to relay information, whereas support is the provision of energy, structural integrity, or components for building proteins.

In that case, I'm not sure why you would consider vesicular release of neurotransmitter by astrocytes to be a support rather than processing role.

 

[DiracPool]

Right, but again you need to provide a definition of support vs processing, otherwise it's not possible to provide evidence.

The most straightforward definition of support versus processing is the criterion of "does the cell support an action potential?" If it does, it's typically designated a neuron, if it doesn't it's typically designated neuroglia. Glia, meaning glue, which holds or glues the processing neurons together and supports them through supplying of nutrients, insulating the axons, processing metabolic waste, etc.

Does that mean that there's never been an instance when an otherwise designated glial cell has participated in some sort of information processing? Probably not.

http://en.wikipedia.org/wiki/Glial_cell

For over a century, it was believed that the neuroglia did not play any role in neurotransmission. However 21st century neuroscience has recognized that glial cells do have some effects on certain physiological processes like breathing,[2][3] and in assisting the neurons to form synaptic connections between each other.[4]

However, when conducting perceptual or behavioral trials in test animals, you're going to get a lot more relevant information on network dynamics by recording spike trains and local field potentials from actual neurons rather than targeting the glial cells.

 

[madness]

The most straightforward definition of support versus processing is the criterion of "does the cell support an action potential?" If it does, it's typically designated a neuron, if it doesn't it's typically designated neuroglia.

Well if you define processing roles to be generating action potentials, then it's trivially true that only neurons do processing. However, glia "support" action potentials in other cells (by vesicular release of neurotransmitter). And, as shown in the recent study I linked to, this process has a marked effect on the gamma frequency LFP and on novel object recognition. I find this to be convincing evidence of a processing role, but clearly it will depend on your definitions.

 

[atyy]

Well if you define processing roles to be generating action potentials, then it's trivially true that only neurons do processing. However, glia "support" action potentials in other cells (by vesicular release of neurotransmitter). And, as shown in the recent study I linked to, this process has a marked effect on the gamma frequency LFP and on novel object recognition. I find this to be convincing evidence of a processing role, but clearly it will depend on your definitions.

It shows that astrocytes can be used to control neural activity. But just going by the abstract, doesn't it depend on tetanus neurotoxin being artifically expressed in astrocytes?

 

[madness]

It shows that astrocytes can be used to control neural activity. But just going by the abstract, doesn't it depend on tetanus neurotoxin being artifically expressed in astrocytes?

The tetanus neurotoxin was selectively expressed in order to disrupt vesicular release. The resulting decrease in gamma power and novel object recognition demonstrates that this vesicular release is somehow implicated in those processes. This is similar to lesion studies, where you destroy part of the brain and see what effect it has, but more targeted.

 

[DiracPool]

but clearly it will depend on your definitions.

That's a fair point. It also depends on what you're interested in. In college my initial interest in neurobiology was in receptor chemistry. However, after a while it shifted to electophysiology when I realized that the important processing that the brain does is in the mesoscopic dynamic range revealed in LFP and small surface cortical arrays. These recordings average out the participation of anywhere between 10's of thousands to 10's of millions of individual neurons and several orders of magnitude more synapses than that.

Thus, whether or not gap junctions or glial cells play a role in information processing and what the extent of that is I am not really an expert on and, frankly, am not so concerned about. At the end of the day what is important is the network dynamics. These dynamics are self-organized, and every aspect of the neuropil plays an important role, gap junctions, glia, neurons, blood vessels and the like. They all work together to generate and sustain the important convergent dynamics and oscillations in the cortex.

 

[atyy]

The tetanus neurotoxin was selectively expressed in order to disrupt vesicular release. The resulting decrease in gamma power and novel object recognition demonstrates that this vesicular release is somehow implicated in those processes. This is similar to lesion studies, where you destroy part of the brain and see what effect it has, but more targeted.

Thanks! Is the vesicular release continuous?

Incidentally, that reminds me of another one I saw http://www.ncbi.nlm.nih.gov/pubmed/23012414, because the paper you mentioned uses carbachol, which IIRC affects acetylcholine.

 

[Pythagorean]

In that case, I'm not sure why you would consider vesicular release of neurotransmitter by astrocytes to be a support rather than processing role.

Because it doesn't demonstrate (in the abstract) where the neurotransmitter acts as a ligand. That was one of the stronger aspects of the paper I posted: it follows the analyzes each leg of the signaling chain. Anyway, I guess a stronger point is that one paper isn't enough. There's a whole body of literature developing on astrocytes (and gap junctions) today.

when conducting perceptual or behavioral trials in test animals, you're going to get a lot more relevant information on network dynamics by recording spike trains and local field potentials from actual neurons rather than targeting the glial cells.

That may be true, but it's essentially a strawman. The point is to expand the information sources, not throw one source (neurons and chemical synapses) aside for another (astrocytes and electrical synapses).


Also, here's a good overview of astrocyte contributions:

A simplified scheme of the different signaling pathways between synaptic terminals and astrocytes for the case of excitatory synapses in the hippocampus (see text for a detailed description). Action potentials arriving at the presynaptic terminal trigger release of glutamate, which can spill over from the synaptic cleft. Perisynaptic astrocytes take up glutamate using their plasma membrane transporters (EAATs) while glutamate, by acting on astrocytic metabotropic receptors (mGluRs), triggers Ca2+ signaling in the astrocyte. This signaling pathway includes production of IP3 and causes an increase of cytosolic Ca2+ due to efflux of this ion from the endoplasmic reticulum (ER). At some synapses, such as in the dentate gyrus, the same Ca2+ signaling pathway could also be mediated by astrocytic purinergic P2Y1 receptors, likely activated by synaptically-released ATP (see text for details). Astrocytic Ca2+ excitability can in turn lead to exocytotic release of several neuroactive substances (or “gliotransmitters”) such as glutamate (Glu), D-serine (D-ser) or ATP which can target specific receptors on pre- and post-synaptic terminals and differentially modulate synaptic transmission. Glutamate acting on presynaptic GluRs could enhance synaptic release, whereas ATP and its derivate adenosine (Adn) could depress it (red path) through presynaptic purinergic receptors (PRs). On the postsynaptic spines [depicted here by a standard RC circuit (Ermentrout and Terman, 2010)], the ensuing effect of gliotransmitters could substantially modify postsynaptic currents by enhancing activation of NMDA receptors (D-serine) or by altering expressions of AMPA receptors therein. Astrocytes could also release TNFα by Ca2+-dependent activation of the matrix metalloprotease TNFα-converting enzyme (TACE), while extracellular TNFα could in turn regulate glutamate release from the astrocyte as well as postsynaptic AMPAR expression. Moreover astrocytic Ca2+ could also propagate across different regions of the same cell or to other neighboring astrocytes by intracellular IP3 diffusion through gap junction channels (GJCs) or via extracellular ATP-dependent pathways, extending gliotransmission to some distal sites away from the considered synapse. For clarity both endocannabinoid-mediated Ca2+ signaling (Navarrete and Araque, 2008), retrograde activation of presynaptic glutamate receptors (Navarrete and Araque, 2010), regulation of postsynaptic NMDARs by presynaptic adenosine receptors (Deng et al., 2011), and the possibility for astrocyte-derived adenosine to enhance synaptic release (Panatier et al., 2011) are not included in this scheme. (B) Equivalent modeling scheme for astrocyte-synapse interactions. The astrocyte (ASTRO) constitutes a third active element of the tripartite synapse in addition to the presynaptic (PRE) and postsynaptic (POST) terminals. In its presence, the flow of input (IN) signals to the output (OUT) is no more unidirectional but presynaptically released neurotransmitter can affect astrocyte function through the interaction pathway A. In turn, the astrocyte can regulate both synaptic terminals via pathways B and C. In addition, the astrocyte could receive additional inputs from or send output to remote synapses in a heterosynaptic fashion (I/O).

[1] http://journal.frontiersin.org/Journal/10.3389/fncom.2012.00098/full

 

[madness]

Because it doesn't demonstrate (in the abstract) where the neurotransmitter acts as a ligand.

Are there any documented cases in which neurotransmitters perform a support rather than processing role?

Anyway, I guess a stronger point is that one paper isn't enough. There's a whole body of literature developing on astrocytes (and gap junctions) today.

Of course. The paper I linked to was published only 2 weeks ago. I'm sure there will be further research attempting to clarify the mechanisms soon.

 

[Pythagorean]

Are there any documented cases in which neurotransmitters perform a support rather than processing role?

Neurotransmitters like glutamate and GABA are regularly recycled through metabolic processes[1]. Astrocytes take up excess glutamate from neural activity... maybe they just recycle it into other products. Maybe they are just sending them back to the neuron to get redocked in the neuron's vesicular reservoir.

It could also serve a compensating role (regulation that serves to keep effective excitability stable when there's local environmental changes). Again, I haven't read the paper, so I don't know if/how they addressed such issues, and the result is still positive in the context of all the other literature.

[1] http://www.nature.com/jcbfm/journal/v27/n12/images/9600490f1.gif