Turner Lectures (7-8) Flashcards
Hippocampus and neocortex in mammalian episodic memory
Entorhinal cortex receives sensory information (via primary sensory cortex etc). Perforant pathway feeds this info into the hippocampus via the dentate gyrus (CA3 then CA1). This is fed back to the EC via the subiculum and then distributed across the neocortex.
Hippocampus: fast but temporary memories, new learning may overwrite old. Neocortex: slow but permanent, new learning is integrated with old.
First observation of LTP
Jerje Lomo in anaesthetized norwegian rabbit hippocampus (in particular the perforant path input to dentate gyrus) IN VIVO
First stimulated path <0.1Hz, then brief HFS/tetanus for 3-4 secs (100Hz), then return to LFS.
Resulted in steeper slope of the extracellular field EPSP, plus increased population spike (=increased number of cells firing).
This was later shown to last for several hours and also shown in awake rabbit.
Main glutamatergic (excitatory) pathways in hippocampus
Hippocampus consists of Cornu Ammonis (CA) region (strip of pyramidal neurons) and the dentate gyrus (DG), (consists of granule cells).
Trisynaptic circuit:
1) Perforant path fibre (perforant fibres connecting to dentate granule cells)
2) Mossy fibres (dentate granule cells connected to CA3 pyramidal cells)
3) Schaffer collateral (CA3 pyramidal cells connecting to CA1 pyramidal cells)
LTP at CA3–>CA1
PTP
Hippocampal circuitry can be maintained in slices 300-500um thick.
Put stimulus on CA3 (Schaffer collateral), record from CA1 (extra or intracellular).
As before, LFS followed by HFS (this time for 1sec) and then return to LFS.
Intracellular: EPSP and EPSC (excitatory postsynaptic current) both increase.
Extracellular: field EPSP increases.
Following HFS, you get post-tetanic potentiation (PTP) = SHORT TERM (~10mins). This decays to a sustained level. With 1 HFS stimulus, this sustained level = “early LTP” (~1hr) (decays back to baseline at 2 hrs). With 4 HFS (1 every 5 mins), get early and late LTP (upto 24hrs),
Theta burst stimulation
TBS more closely mimics natural rhythms of activity in the brain (theta waves)
TBS is another stimulation protocol: instead of 100Hz for 1sec, 100Hz
use a four-pulse burst at 100 Hz and repeat it ten times at five bursts per second (200 millisecond gaps).
What does CA3–>CA1 LTP induction depend on? (x3)
How are they blocked?
1) NMDA-R activation (blocked by APV/AP5)
2) Postsynaptic increase in Ca2+ (blocked with Ca2+ chelator, EGTA)
3) Postsynaptic depolarization (blocked by injection of -ve current)
Experiment demonstrating this: start with hyperpolarised cells. Briefly depolarise to 0Hz, paired with a LFS (2Hz for 20s).
EPSC after this pairing is increased, even though HFS wasn’t used (don’t necessarily need AP?)
Input interactions during LTP induction
1) WEAK HFS input from single pathway –> insufficient postsynaptic depol. –> no LTP induced. Below threshold - need certain degree of postsynaptic depol. to induce LTP.
2) Stronger stim. of same single pathway –> more fibres activated, “threshold for induction” is reached –> LTP (minimum no. of inputs required within a pathway)
3) LTP is input specific - it will only be induced at inputs that are active during HFS (not adjacent inactive ones)
4) However, combining weak inputs from different locations can generate sufficient depol. even if they couldn’t on their own (=CO-OPERATIVITY)
5) Weak HFS input from one location paired with strong HFS from another can also generate LTP at both places (=ASSOCIATIVITY) (can occur in weaker inputs than (4))
LTP at other synapses
Almost all synapses within the extended hippocampal region are Hebbian EXCEPT the mossy fibre pathway (dentate gyrus –> CA3). Postsynaptic depolarisation not neccessary so non-Hebbian.
Protein kinases in CA3–>CA1 LTP
EARLY LTP
EARLY LTP - Dependent upon PKC and CaMKII (blocked by peptide PKC19-31 and peptide CaMKII273-302 respectively)
LTP associated with phosphorylation of GluR1 subunit of the AMPA-R. Both kinases can target the Ser831 site –> increased conductance of AMPA –> increased current flow and therefore larger EPSC amplitude and EPSP.
Phosphorylation at Ser845 doesn’t have this effect BUT is involved in receptor trafficking (–> R insertion into plasma membrane). Ser831 phosphorylation –> translocation into the postsynaptic density. This leads to increase in MACROscopic conductance (rather than just MICROscopic conductance increase associated with individual AMPA-Rs).
Presynaptic & postsynaptic locus for LTP?
Changes in AMPA-Rs = postsynaptic locus for LTP.
Is debate about a presynaptic locus, which increases the probability of vesicle release.
Would involved intercellular signalling molecules e.g. diffusible messengers (NO, arachidonic acid…) and membrane-bound extracellular proteins (integrins, ephrin receptors…)
Anatomical changes in LTP
May also be rapid ultrastructural changes - spine splitting.
Apply TBS to CA3–>CA1.
E.M. serial sections have been reconstructed to show two spines instead of 1 post-LTP induction. Increased number of spines connecting same axon and dendrite.
30 mins after HFS, postsynaptic density splits and spinule pushes into presynaptic terminal. Postsynaptic terminal splits into two (60mins) and presynaptic gets remodelled.
Unlikely to be mediated via genome, probably local mechanism involving cytoskeleton.
(LATE) LTP blocked by…
H89 (seletive PKA inhibitor)
Anisomycin (translational inhibitor)
Actinomycin D (transcriptional inhibitor)
Late stages therefore similar to Aplysia:
Increase in postsynaptic Ca2+ –> activation of PKA –> phosphorylation of CREB –> binding to CRE –> transcription of IEGs –> transcription of LRGs –> NEW PROTEIN SYNTHESIS
CA3 –> CA1 LTD
If synaptic weights can only be increased, eventually all synapses would become fully potentiated. Need a means of reversing LTP –> LTD/depotentiation
Same experimental setup as before (record from CA1, stimu. CA3 Schaffer).
First monitor baseline at <0.1Hz. Then 1Hz for 15 minutes (900 stimuli) (LFS). Then return to baseline.
Get sustained depression of responses to about 80% of control level.
CA3–>CA1 LTD depends on… (differences?)
Same as LTP: NMDA-R activation (blocked by AP5), increase in postsynaptic Ca2+, postsynaptic depol.
How can this produce two different effects?
Testing different frequency stimulations (900 stimuli between 1-50Hz) shows: >10Hz –> LTP, <10Hz –> LTD.
Plasticity change is determined by the degree to which presynaptic activity affects postsynaptic neurons.
During LFS, pre and postsynaptic activation is correlated but this leads to decrease in synaptic weight and therefore is ANTI-HEBBIAN (synapse can be both - depends on the conditions)
Mechanisms of LTD
Protein kinases important for LTP so maybe opposite effect (dephosphorylation) required for LTD. Prevent dephosphorylation –> prevent LTD.
E.g. okadaic acid (protein phosphatase 1 & 2A inhibitor) and FK506 (calcineurin/protein phosphatase 2B inhibitor).
Protein kinase inhibitors DON’T block LTD.
Dephosphorylation seems to be targetted to Ser845 of GluR1 subunit of AMPA-R. Seems to decrease open time probability, which disengages AMPA-Rs from their binding partners and initiates retrieval of the R from membrane by endocytosis (reverse of priming insertion).