Plastic axonal trees

You can’t teach an old dog new tricks…or can you?

We’ve all heard about how the brain slows down as we age.  We’re constantly losing brain cells.  Neurons become “static” and cannot make new connections.  Is this true?  Are we really doomed to a lifetime of deteriorating mental function?

A paper by Oberlaender et al came out this week in the journal Neuron that disputes this common view of the adult brain.  They studied plasticity in the adult rat somatosensory cortex.  Plasticity refers to the ability of neurons to change shape, connections and activity levels in response to environmental changes.  The somatosensory cortex is the area of the brain that processes touch sensation.  In the rat (and mouse), the somatosensory cortex is organized into barrels, where each cylindrical chunk of brain responds only to signals sent from an individual whisker.  The sensory nerves in a whisker travel to the brainstem first, then they make a stop in the thalamus (a deep region of the brain), and finally the thalamic neurons synapse with cortical neurons in layer 4 (remember that the brain cortex is arranged into 6 layers). 

The somatosensory circuitry, from a sensory neuron (pink) 
to the brainstem (medulla) to the thalamus and then somatosensory cortex.

It has been shown before in the adult rat somatosensory cortex that the neurons can undergo structural and functional changes in response to changes in activity.  This kind of plasticity has only been observed before in the cortex; other areas of the brain were thought to be static once the brain reaches the adult stage.  Oberlaender et al. show, however, that the neurons from the thalamus can also undergo structural changes in the adult brain.

Changes in axonal morphology

To induce neuronal plasticity, the researchers trimmed a single whisker on a group of rats.  This is a painless procedure, so they don’t have to take into account responses due to an injury.  This trimmed whisker will no longer be sensing the environment, so its sensory neurons will be silent.  Three days later they filled the thalamic neurons associated with that whisker with a dye, so they could image the shape of the neurons. 

We need to take a brief pause here to discuss neuronal anatomy: Neurons have a round cell body, dendrites which receive signals from other neurons and axons which send signals to other cells.  Axons can travel great distances and make synapses with many different cells.  The general consensus in the field is that as we learn something new, more synapses and connections are made between neurons.  The image below is a thalamocortical rat neuron (a thalamic cell that makes synapses with the cortex).  You can really see that the axon and dendrites have lots of branches and each of those branches may have multiple synapses to other neurons.  This is why neurons are often compared to trees with their branching limbs.

Thalamocortical neuron.  From Destexhe et al., 1998, J. Neuro.

The authors compared the morphology of thalamic neurons from control rats to those that had their whisker trimmed.  The neurons corresponding to the  trimmed whiskers had considerably shorter and less branched axons.  Remember, these neurons in the thalamus are no longer receiving signals from the whisker, and in just three days they started to retract.  This often happens in the cortex, where an unused area of the brain will just shrink up or get taken over by other neurons.  This is the first time this has been shown for a non-cortical region of the brain. 

Functional compensation

The shortened axons are obviously making fewer synapses with cortical neurons, so these neurons should be less active.  However, when they recorded electrical activity, there was no difference in L4 cortical cells in the trimmed mice compared to controls.  The authors investigated this more thoroughly and looked at synchrony between cells.  Neurons that are active at the same time will often add up their signals at the next connection, so this is another way of looking at activity in the brain.  The trimmed mice had more synchronous cortical cells than the controls.  That makes no sense, right?  They have fewer synapses, so how could they be more synchronized?  Apparently there must be some form of compensation for the decrease in axon length and synaptic connections.  In other words, the remaining synapses become stronger to maintain a normal level of electrical activity.  We call this process homeostasis, which happens at many different levels in our bodies (temperature regulation, blood sugar, etc).

To summarize, trimming the whiskers results in less signaling to the whisker area of the thalamus.  As a result, the thalamic neurons become shorter and less branched.  They make fewer synapses onto the cortex, but it doesn’t matter because these synapses increase their strength to maintain a physiological activity level. 

The important point here is that adult neurons can undergo structural plasticity (shorter axons) and functional plasticity (strengthening of synapses) as a result of experience (or lack thereof).  These changes happened really quickly – only three rat days.  The authors conclude that “thalamocortical input to cortex remains plastic in adulthood, raising the possibility that the axons of other subcortical structures might also remain in flux throughout life.”  There’s hope for us after all!

This blog title was somewhat inspired by the Radiohead song "Fake Plastic Trees".  Oh the 90's.

There's more than one way to get out of a nucleus

This week’s paper totally blew my mind.  It’s so amazing to me that we know so much about how cells work and yet this paper describes a totally new process that has never been observed before.  How’s that possible?  What else remains to be discovered?

In this week’s issue of Cell, Speese et al. describe a new method for molecules to exit the nucleus.  The nucleus is where the genome is stored and where genes are expressed into RNA (see Fundamentals section for more details).  The nucleus is surrounded by two nearly impenetrable membranes, an inner and outer membrane.  All biology students are taught that there is only one way into and out of the nucleus: the nuclear pore complex (NPC).  It’s an elaborate and selective protein pathway that spans the inner and outer membranes of the nucleus (I’ve written about the nuclear pore complex before in the context of long-lived proteins).  If the molecule has the right kind of chaperone, then it can pass through the NPC to get into the cytoplasm. 

The lab that did this research studies a signaling pathway at the Drosophila neuromuscular junction (that’s what I study too!)  The neuromuscular junction is a synapse (remember synapses?) between a motor neuron and a muscle.  The pathway that they study is called the Wnt pathway and they’ve shown that it helps regulate the development of the neuromuscular junction.  The neuron releases a signal that binds a receptor on the muscle.  The active receptor, called Frizzled, is taken into the cell where it gets cut into two pieces.  One piece enters the nucleus and they have observed that it accumulates into big aggregates (we’ll call them “foci”).  But what happens next?  They suspected that the foci may leave the nucleus, but how could such a big glob of proteins get through the relatively narrow nuclear pore complex?  Ay, there’s the rub.

Nuclear Lamins

The first thing the authors investigated was whether nuclear lamins might be associated with the nuclear foci.  Nuclear lamins are the resident proteins in the nucleus that help spatially organize the chromosomes and provide scaffolding for the nucleus and nuclear pore complexes.  Lamins are important medically because there are a number of “laminopathies”, human diseases associated with mutated lamins.  One particular lamin, called LamC, does appear in the Frizzled aggregates in the muscle nucleus.  In fact, if you get rid of LamC, then the Frizzled receptor no longer forms foci in the nucleus.  The association with LamC, though, doesn’t explain how the Frizzled receptors function in the nucleus.

Perhaps Frizzled binds to DNA; that’s the major molecule in the nucleus after all.  Nope.  The authors found that there is no DNA associated with the Frizzled clumps.  Well, perhaps it binds mRNA then.  Remember that DNA is copied into small pieces of mRNA which then travel out of the nucleus and direct protein synthesis (see Fundamental #1).  Sure enough, they find that there are some mRNA molecules bound in the Frizzled/LamC aggregates.  It is not that uncommon to find proteins bound to mRNA in the nucleus and usually they all transport out of the nucleus through the pores.  mRNA can only be translated into protein in the cytoplasm, so they have to leave the nucleus to be functional.  The problem is that these Frizzled globs are way too big to go through the pores.  The aggregates must disassemble before leaving the nucleus, right?  Right?  Not necessarily.

Nuclear budding

The researchers did lots of imaging of the Frizzled foci and found something strange.  They sometimes would see an aggregate surrounded by nuclear membrane on both sides.  In other words, the aggregates appeared to be in the space in between the inner and outer membranes.  What?!  They also imaged the foci over time and could see them staying intact as they leave the nucleus.  In the end, what they believe is happening is that the particles get surrounded by the inner membrane and then pinch off into the intermembrane space.  At that point they are surrounded all the way around by a sphere of membrane.  That membrane fuses with the outer membrane, spitting out the particles into the cytoplasm.  See the figure below for a visual representation of this process.  The role of LamC in this process is probably to help promote rearrangement of the membrane scaffold to allow nuclear budding. 

It turns out that this same sort of process happens during infection by a herpes virus.  The viral genome goes into the host nucleus and directs formation of new viruses.  The complete virus is much too big to exit the nucleus through the pores, so it uses this method of nuclear budding to get out of the nucleus.  Everyone assumed that this process was specific for viruses, but now we see that it’s actually a normal cellular activity.  In fact, the virus probably hijacks the apparatus for normal nuclear budding to get itself out of the nucleus.  Viruses always know more than the scientists.

Synaptic translation

Now we’ve solved the problem of how the Frizzled/LamC/mRNA complexes can exit the nucleus, but what happens next?  How does this continue the Wnt signaling pathway?  The authors followed the path of one mRNA that binds Frizzled in the nucleus.  They found that the mRNA localizes near the synapse (in the muscle side).  It’s been shown before that local protein synthesis at the synapse can direct synapse development and plasticity.  When LamC is impaired, this mRNA no longer goes to the synapse, which really goes to show how important lamins are in cellular signaling.

Review time:

1) Signal from neuron to muscle activates the Frizzled receptor

2) A piece of the receptor gets cut off and enters the nucleus (via the NPCs)

3) The Frizzled piece binds up with LamC and particular mRNAs

4) LamC helps rearrange the nuclear membrane, so budding can occur

5) The large Frizzled particles exit by budding straight through the membranes

6) Frizzled and the mRNA travel to the synapse

7) The mRNA is translated into a particular protein that directs synapse development

This helps solve a problem about the laminopathies (diseases caused by mutation of a lamin).  Often these diseases cause muscular dystrophy.  Well, maybe what happens at the Drosophila neuromuscular junction could happen in humans too.  If a lamin is mutated, it may prevent membrane budding, so RNA-protein complexes cannot exit the nucleus.  If they get stuck in the nucleus, the mRNA cannot be translated at the synapse, so no new proteins will be made and the neuron-muscle synapse will be impaired.  That would explain defects in the muscle system caused by these diseases. 

Electrically coupled cells make a connection

Inside our brains are billions of neurons, which communicate to each other via chemical messengers called neurotransmitters.  Neurotransmitters are released from one cell and bind to receptors on a neighbor cell, creating a chemical synapse between the two neurons.  Precise connections between neurons create microcircuits made up of multiple cells, which are responsible for processing sensory information or controlling motor output.  Our brains contain an outer cortex that has 6 layers of cells (you can think of it as 6 layers of an onion peel with layer 6 being the closest to the center).  Between the layers are functional columns made up of neurons that all serve a similar function and connect to each other in a microcircuit.  In other words, cells in layer 1 connect to layer 2 and layer 3 cells and so on, until there is a circuit of inter-connected neurons spanning all 6 layers. 

Diagram of the 6 cortical layers. From thebrain.mcgill.ca, an excellent neuroscience primer

What’s interesting is that cells right next to each other in one layer may not form chemical synapses with each other at all, opting instead to synapse with a neuron in another layer.  What’s so special about the neurons in other layers?  How do neurons decide with which cells to make a chemical synapse? 

Another interesting thing about these microcircuits that make up the columns is that the interconnected neurons were “born” from the same stem cell.  During the development of an organism, there are many stem cells (also known as progenitor cells) that divide to make mature cells.  When a progenitor cell divides, it makes one copy of itself, as well as one cell that will mature to become an adult cell, which is no longer capable of dividing.  What’s interesting about this process in the brain is that one progenitor cell will end up making many “sister” neurons and these are the neurons that make synapses with each other, forming the microcircuits.  We can now refine our initial questions to ask: How do sister neurons find each other in the brain, through multiple layers, and then make chemical synapses with each other?  How do they know they were made from the same progenitor cell?

The answer is: sister neurons are electrically coupled with each other early in development, which promotes the formation of chemical synapses.  This process was described recently by Yu et al. in the online version of Nature.  Before we look at the experiments, we need to understand what it means to be “electrically coupled”.

Electrical vs chemical synapses

I mentioned already how neurons are functionally connected to each other through chemical synapses.  Packages of neurotransmitters, a type of chemical, are released from one cell to tell a neighbor cell to become activated (or inactivated).  This is sort of an indirect way for cells to communicate since it requires a chemical messenger as a mediator. 

There are also more direct connections between neurons called electrical synapses.  In this case, a tube of proteins, called a gap junction, acts as a tunnel between two neurons.  When one neuron gets activated, positive ions rush into the cell.  These ions can go through the gap junction and enter the second neuron, also activating it.  In this way, the two neurons are electrically coupled, because an activating current can easily spread from one cell to another in an instant.  In the adult nervous system, most of our synapses are chemical synapses, but one prominent example of electrical synapses occurs in the heart.  A small group of muscle cells, called the pacemaker, set the heart rate.  When they become activated, they quickly spread the electric signal to the rest of the cells in the heart because they are all directly connected to each other through gap junctions.  The message sent through an electrical synapse is less likely to fail than a chemical synapse, so that’s one reason why it’s important for our heart cells to be electrically coupled.

From University of Tokyo, Life Science web textbook

 

Electrically coupled sister neurons

The researchers set out to show that sister neurons, borne of the same progenitor cell, form electrical synapses with each other shortly after they are made.  They found a way to label sister neurons with a fluorescent protein.  They then stuck recording electrodes into two labeled sister neurons and two other nearby neurons that came from other progenitors.  All four neurons were overlapping, so they could easily form synapses with each other if they wanted.  They then injected current into one cell and found that the current traveled to the sister neuron, but not into the other nearby neurons.  What does this tell us?  There are electrical synapses between sister neurons but not between other cells.  The authors go on to show that the electrical synapses are only present during the first few days of life.  By the 6th day after birth, the sister neurons are no longer electrically coupled.

What’s the benefit of being electrically coupled early on in development?  It allows the neurons to be activated in synchrony.  Another important aspect of neuronal signaling is that there is a minimum threshold necessary for a cell to become activated.  A small amount of ion flow (i.e. current) will dissipate and the neuron will not be active or send signals to other cells.  Once the current reaches a certain threshold, though, the cell fires what we call an “action potential”.  Basically the neuron will get super positive inside really quickly and it will release chemical signals to other neurons.  This is what I mean when I say a neuron is “active”.  So if two cells are electrically coupled, a subthreshold current in one cell, plus another subthreshold current in the second cell will add up to push the neurons into the active state.  Thus both neurons will fire an action potential at the same time.

Okay, so now we have one last question: sister neurons are electrically coupled, so they fire action potentials together in synchrony, so what? 

“Neurons that fire together wire together”

It’s been shown that neurons that are active at the same time are more likely to form chemical synapses.  This finding has been used to explain how neurons can form new synapses during the process of learning (see: Hebbian plasticity).  If we apply this theory to development, then it would follow that the electrically coupled sister neurons are active together, which leads to the formation of chemical synapses between them.

To prove this hypothesis, the authors created mice that had defective gap junctions and therefore no electrical synapses (only in one region of the brain).  They found that in these mice, sister neurons no longer formed chemical synapses with each other.  The electrical coupling between sister neurons is necessary for these neurons to form chemical synapses later on in development.

Let’s review

            1) Progenitor cell divides many times to make sister neurons

            2) Early in development, sister neurons form electrical synapses

            3) Sister neurons are active in synchrony (they “fire” together)

            4) Synchronous activity leads to chemical synapse formation

            5) The electrical synapses go away later in development, leaving behind a microcircuit 
                of sister neurons connected via chemical synapses

Yay for electrical synapses… no longer a footnote in Neuroscience textbooks!