Rapid synaptogenesis following long term potentiation in adult hippocampus
Long term potentiation (LTP) is considered to be like the cellular mechanism for long term memory and is (justifiably) widely researched, with 9889 pub med hits. Using an innovative technique, Bourne and Harris have demonstrated how LTP correlates with the structural plasticity of dendrites. Briefly, this was their procedure:
First, they stimulated the axons of rat hippocampus slices in CA3 with concentric bipolar stimulating electrodes. These electrodes produce action potentials in all of all axons beneath them but can induce LTP selectively, allowing the researchers to compare control dendrites to those recieving LTP to isolate the effects of LTP. The electrodes emitted theta burst stimulation, which entails 8 "trains" delivered 30 seconds apart of 10 bursts per 2 seconds of 4 pulses at 100 Hz.* This frequency is apparently more resemblant of natural firing patterns in the hippocampus than the alternatives and allows for the back propogation of action potentials, which is important for structural plasticity.
Second, they determined which slices had the proper physiological characteristics of LTP, which are: 1) a dose-dependent increase of output current as the input stimulating electrode intensity is increased, 2) a lack of LTP at the control site, and 3) an induction of LTP at the correct site, as indicated by an increase in the slope of the field excitatory post synaptic potential. They split their tissues into three groups characterized by the amount of time in the experimenter after the induction of LTP: 5 mins, 30 mins, and 120 mins. Hippocampal slices matching all of the three characterstics were "fixed" by immersion in mixed aldehydes within seconds of the end of these time frames.
Third, they vibra-sliced the tissues across their width at 70 micrometer thickness and determined which tissues were preserved to high quality, as indicated by the clear identification of the proper intracellular components (i.e. microtubules of dendrites, cristae of mitochondria). Tissues that passed this anatomical criteria were sectioned, at ~ 38 to 60 nm. The researchers visualized the CA1 region of these tissues with transmission electron microscopy (TEM).
The time-dependent differences in anatomy following LTP that they found were fascinating. Dendrites from tissues 5 / 30 mins after LTP generally had more of the transitional structures associated with synaptogenesis including asymmetric shaft synapses, stubby spines, and nonsynaptic filopodia. But these changes were gone in the dendrites of the 120 mins samples. Instead, the dendrites of the samples 120 mins after LTP had a significant increase in postsynaptic density, i.e. in one subset, an increase in average area from ~ 17 +/- 1.5 square micrometers in the non-LTP sample to ~ 22 +/- 1.5 square micrometers in the LTP sample. This is evidence for the theory that the first initial changes in synaptic strength following LTP are underlied by transient changes, such as synapse unsilencing and AMPA trafficking, but later the changes are mediated by an increase in synaptic area.
Another time-dependent measure that the researchers were able to look at was the location and magnitude of polyribosome upregulation following LTP. Ribosomes are ~ 18-25 nm in diameter and are connected in polyribosomes by a grey fuzz which can be detected by TEM. At 5 mins after LTP, in the head and neck of dendrites, polyribosomes were upregulated from ~ 0.09 +/- 0.02 polyribosomes per micrometer in the control group to 0.20 +/- 0.04 polyribosomes per micrometer in the LTP group. But at 120 mins after LTP, in the head and neck of dendrites, polyribosomes were downregulated from ~ 0.34 +/- 0.07 polyribosomes per micrometer in the control group to 0.10 +/- 0.02 polyribosomes per micrometer in the LTP group. The net control upregulation can be explained by supporting the ongoing formation of small dendritic spines. And although the LTP group had a overall downregulation by 120 mins, this makes sense in terms of localizing the polyribosomes to dendritic spines that have enlarged postsynaptic densities. So, ribosomes can be transported around CA1 dendrites fast.
Overall, the model of synaptic plasticity emerging from their data set and analysis is that dendrite segments can act as functional units. If an inhibitory or excitatory spine is lost, the remaining ones should increase in size to compensate by ~ 120 mins. Moreover, it's possible that polyribosomes or other intracellular organelles in the dendrite could be responsible for the regulation of dendritic homeostasis independently of the cell's nucleus. This study also shows the utility of the time-dependent ssTEM technique, one that will surely begin to recieve more use in the next decade as machine learning algorithms continue to make the data analysis easier.
* I might as well admit that I don't really get this part of their experiment that well.
Reference
Bourne JN, Harris KM. 2010 Coordination of size and number of excitatory and inhibitory synapses results in a balanced structural plasticity along mature hippocampal CA1 dendrites during LTP. Hippocampus Early View. PubMed. Doi: DOI 10.1002/hipo.20768.