Fellin et al.
In this 2004 paper, Fellin et al. investigate the nature and mechanism of slow inward currents (SICs) evoked in CA1 neurons of the Wistar rat hippocampus by stimulation of Schaffer Collaterals (SC). In contrast to the old dogma that glial cells serve only nutritive and insulating functions to neurons, the authors find that SICs are in fact evoked by glutamate released from neighboring astrocytes via a calcium dependent mechanism. To arrive at this conclusion, they first characterize the SICs. SICs: occur spontaneously and upon high-frequency SC stimulation or application of DHPG, a group I mGluR agonist; are mediated by NMDARs; arise from glutamate of nonneuronal, or nonsynaptic, origin (Fellin et al. deduce this by observing that TeNT and TTX do not abolish SICs). Next, by calcium imaging, they find that DHPG elevates intracellular calcium concentration (Cai) in both astrocytes and in neurons, but only the latter is sensitive to D-AP5, an NMDAR antagonist. When Cai of single astrocytes was increased by photolysis of caged calcium, they observed that the SIC was recorded in the neuron at the same time that the calcium wavefront reached the dendrite of the astrocyte. This provided evidence for a causal link between astrocyte Cai increase and the NMDA-mediated SICs in neurons. The authors then used ifenprodil, an NR2B antagonist, and observed that the SIC amplitude was drastically and reversibly reduced. This meant that the NR1/NR2B complex, found mostly in extrasynaptic regions, mediates SICs. Interestingly, these SICs were found to occur in synchrony among neighboring neurons that were shown to be not electrically coupled. The Cai elevations in neurons were also found to be in synchrony and were preceded by elevations in astrocytes. To establish physiological significance, Fellin et al. repeat the experiment in 1mM magnesium. Although they still observe SICs, both in synchrony and in solitude, I am not convinced that these are physiologically significant because the frequency and amplitude of SICs are drastically reduced compared to 0mM magnesium. Furthermore, earlier in the paper (in Figure 3), they used 1mM magnesium to show that it significantly depresses the SIC amplitude and offered this as evidence that SICs are NMDAR mediated.
Diamond et al.
In this 1998 paper, Diamond et al. investigate whether the locus of long-term potentiation (LTP) expression in the CA1 region of the hippocampus is pre- (increased transmitter release) or postsynaptic (heightened postsynaptic sensitivity). The authors used the relatively novel method of recording simultaneously the synpatically activated glutamate transporter currents (STCs) in astrocytes located in the stratum radiatum and field excitatory postsynaptic potentials (fEPSPs). By using kynurenate (KYN), a glutamate receptor antagonist, they were able to separate the STCs from fEPSPs. First, Diamond et al. show that three manipulations of release probability (Pr) ---paired-pulse facilitation, lowered Ca:Mg ratio, and posttetanic potentiation ---and one manipulation of number of release sites (n) ---raising stimulus strength to activate more synapses ---resulted in the increase or decrease of STC amplitude, making it a reliable reporter of the amount of glutamate release. Next, they use high frequency stimulation (HFS) to induce LTP in CA1 of Sprague-Dawley rats, and observed that while fEPSP was increased 1.7-fold, STC was unchanged. This meant that neither Pr nor n is affected by LTP induction in CA1. I thought that this experiment was a relatively simple and elegant way to determine the locus of LTP expression. The control experiments with aracidonic acid, sodium nitroprusside (SNP), and forskolin application CA3 make this paper very complete. The fact that glial cells have the ability to sense or integrate neuronal activity is also fascinating.
Glutamate Release Monitored with Astrocyte Transporter Currents during LTP.
Diamond et al. continued last week’s discussion regarding the mechanisms of LTP in CA1 neurons, specifically looking at whether it is associated with increased glutamate release from Schaffer collaterals. They used a new recording method where glutamate transporter currents (STCs) were recorded in astrocytes to gauge glutamate release, as astrocytes remove extracellular glutamate through electrogenic transport. To manipulate the probability of release, they used three techniques: (1) paired pulse facilitation (2) lowering Ca2+, and (3) burst of high frequency stimulation. The authors then induced LTP finding that the amplitude of the fEPSP increased accordingly, but noticed no change in STC. They proposed that increases in the STC were canceled by astrocytic down-regulation of glutamate transport through arachidonic acid (AA) or nitric oxide (NO), but disregarded their hypothesis upon finding that AA had only a short-term effect on STC inhibition and NO had no effect. Finally, the authors examine LTP in CA3, as this area is associated with increased glutamate release. They were able to detect STCs through forskolin, as it potentiates mossy fiber inputs to CA3 neurons but found that it was blocked by DCG-IV. Overall, I thought this paper was quite interesting in the way that it used a novel technique to examine LTP, when everyone else was focused on measuring the postsynaptic response.
Neuronal Synchrony Mediated by Astrocytic Glutamate through Activation of Extrasynaptic NMDA Receptors.
Fellin et al. examined the role of astrocytes in relation to the observed slow inward currents (SICs) found in CA1 neurons upon stimulation of Schaffer collaterals. The authors first found that these SICs were mediated exclusively by NMDA receptors upon application of the NMDAR antagonist D-AP5, which abolished the SIC. They then applied DHPG, a metabotropic glutamate agonist, and found that it triggered SICs but argued that the glutamate was nonsynaptic in origin, as application of tetanus neurotoxin, which blocks synaptic release of neurotransmitters, had no effect on the SIC. They then used calcium imaging experiments to discover that DHPG causes an increase in intracellular calcium concentration in astrocytes and through photolysis were able to elicit increased calcium levels in a single astrocyte to measure the SIC of a single adjacent pyramidal neuron. In looking at the mechanism for glutamate release, they used the glutamate transporter inhibitor TBOA and found that it does not abolish DHPG-triggered SICs, suggesting that reverse operation of glutamate transporters does not release glutamate. In looking at the mechanism for the SIC response, they used the NR2B antagonist ifenprodil, which eliminated the SIC amplitude, suggesting that extrasynaptic NMDARs mediate the SICs. Finally, they looked at the considerable synchrony of the SICs using paired recordings, finding that this synchonization was not due to electric coupling through gap junctions but was NMDAR mediated. The results from their experiment overturn the dogma that glia are mere support cells for neurons.
Diamond et al. 1998
Diamond et al. used simultaneous whole-cell voltage clamp recordings of fEPSP and synaptically activated glutamate transporter currents (STC) from astrocytes as measures of neuronal glutamate release in order to determine if the probability of presynaptic vesicle release is altered following LTP induction. The authors had to isolate the transporter-mediated current from the extracellular field potential and potassium currents that also comprised the complex astrocytic electrical response. They did this by administrating KYN, a glutamate receptor antagonist. The remaining current was diminished by glutamate uptake blockers, indicating that it was mediated by glutamate transporters. Diamond et al. then used three different manipulations to determine if the STC measure was sensitive enough to discern subtle changes in neurotransmitter release. The results of these manipulations demonstrated that STC amplitude was both sensitive to changes in pr and n, more so than the fEPSP. The authors then selected two independent pathways in which they recorded STCs before and after LTP induction. There was no difference in the STC measures, indicating no change in the amount of glutamate release. In order to determine if arachidonic acid (AA) or nitric oxide (NO) could account for these results by inducing a downregulation of glutamate transport into the astrocyte, the authors carried out a series of tests in which they applied AA and NO to the cell solution. While AA caused a reduction in STC, NO had no effect. Still, they concluded it was unlikely AA plays such a compensatory role as its effects are rapidly reversible. Lastly, the authors analyzed LTP in mossy fiber projections to CA3 neurons and noted forskolin was necessary to make the transporter current discernible. They conclude by discussing the theory of glutamate "spillover" as a means for LTP expression and giving reasons against increased glutamate release following LTP induction.
Fellin et al. 2004
Fellin et al. begin by noting how stimulation of Schaffer collaterals evoked slow inward currents (SICs) in CA1 neurons that were characterized by a slow rise/decay time. SICs were also found to occur spontaneously and after administration of DHPG, a group 1 mGluR agonist. These SICs were blocked upon application of D-AP5, indicating their dependence on NMDAR activation. As SICs were still found following administration of TTX and tetanus neurotoxin, the authors were able to conclude the glutamate responsible for SIC induction was nonneuronal in origin. Given how SC stimulation is known to increase intracellular Ca2+ concentrations in astrocytes and ultimately trigger glutamate release, Fellin et al. then examined if DHPG would induce SICs by a similar mechanism. Using imaging techniques, they found DHPG administration successfully elevated Ca2+ levels in astrocytes. Then, using photolysis, they were able to locally increase astrocytic Ca2+ levels and record the SICs evoked in corresponding pyramidal neurons. Again, D-AP5 decreased the SIC response, proving that this current was indeed a result of astrocytic glutamate release mediated by NMDARs. Blockade of glutamate transport inhibitors by TBOA did not affect SICs, indicating glutamate was not released by reverse operation of astrocytic glutamate transporters. Use of ifenprodil, an NR2B antagonist, diminished SIC amplitude, showing that SICs are mediated by extrasynaptic NR1/NR2B NMDA receptor complexes. Using paired cell recordings, the authors were able to note synchronous activity in subsets of pyramidal neurons after stimulating glutamate release from astrocytes. This coordination of synchronous activity was also observed under normal cell conditions in the presence of Mg2+ and the absence of picrotoxin, a GABA receptor antagonist. The authors conclude by assessing the physiological relevance of this mechanism for inducing synchronous activity, suggesting it is triggered during high levels of synaptic activity and that it may possibly play a role in information processing.
Diamond et al.
The authors of this paper attempt to determine if LTP in CA1 neurons is associated with increased glutamate release from Schaffer collaterals, through the use of astrocyte recording techniques. This method is viable as astrocytes take up glutamate that is spilled into the synaptic cleft via electrogenic transport, thus a current (a STC) can be recorded. They began by ensuring that astrocyte currents were sensitive to changes in glutamate by comparison of STCs and field EPSPs, along with manipulation of release probability (p) and the number of synapses recruited (n). STCs show a biphasic response as result of the fEPSP current, which the authors simplified by application of KYN, a glutamate receptor antagonist. The remaining current is blocked by THA and DHK, proving it is mediated by glutamate transporters. Paired pulse facilitation, post tetanic potentiation, low Ca2+, and varied stimulus strengths all affected both STCs and fEPSPs, showing that astrocyte currents are in fact sensitive to changes in glutamate levels. Thus, the authors were able to induce LTP and note that, while fEPSPs were potentiated, STCs remained unchanged, suggesting that neither p nor n are changed by LTP induction. They noted it was possible that an increase in STC was canceled out by downregulation of glutamate transport by either arachidonic acid or NO, but this was deemed unlikely as AA has only a short-term effect and NO donors had no effect. The authors attempted to examine LTP in CA3, which is associated with increased glutamate release, but found STCs difficult to detect unless release was enhanced even further by forskolin.
Fellin et al.
In this paper the authors explored slow, inward, excitatory currents in CA1 neurons (SICs) that are mediated by glutamate released from astrocytes onto extrasynaptic NMDARs. They noted that the type of NMDAR associated with slow kinetics tends to be found in extrasynaptic membranes. The authors invoked SICs with high frequency stimulation of Schaffer collaterals, and noted their appearance in a subpopulation of CA1 neurons. The SICs could be evoked repetitively and occurred at low frequency. Additionally, they could be blocked by D-AP5, an NMDAR antagonist. SICs are also invoked by DHPG, an mGluR agonist. These SICs persisted even after addition of tetanus toxin, suggesting an non-neuronal source of glutamate. DHPG, like SC stimulation, was shown to increase Ca2+ concentration in astrocytes, via Ca2+ imaging. An associated Ca2+ increase in CA1 neurons was noted, which was blocked by D-AP5. The authors then increased Ca2+ in single astrocytes with caged Ca2+ and photolysis, and recorded temporally corresponding SICs in nearby CA1 neurons, which could be blocked with either D-AP5 or Mg2+. Use of TBOA proved that SICs were not mediated by glutamate transporters. The suspected type of NMDAR was then confirmed to be vital by the use ifenprodil, a selective NMDAR antagonist. The authors also found that SICs and Ca2+ increases could be temporally correlated among multiple neurons that are not electrically coupled. The SICs noted did not always have the same amplitude, and could be synchronized across considerable distances. Experiments under physiological conditions proved SICs could occur in the presence of Mg2+ though at reduced frequency and amplitude. Synchronization could still be found.
November 9, 2010
Diamond et al.
The authors of this paper claim that expression of LTP does not involve increases in the amount of neurotransmitter released from the presynaptic cell. They studied the Schaffer collateral/CA1 synapse and its interactions with astrocytes whose processes reside in close proximity with these synapses. Astrocytes contain many glutamate transporters and are believed to be responsible for removing glutamate escaping from the synaptic cleft. This transport of glutamate generates a current, known as a synaptically activated glutamate transporter current (STC) which the authors measure to gauge the probability of glutamate release from Schaffer collaterals. They found that the amplitude of the field EPSPs and the STC both increased in response to high frequency stimulation, indicating that the STC amplitude is correlated to the amount of transmitter released. They then induced LTP in the CA1 cell and found that although the amplitude of the fEPSP increased accordingly, the amplitude of the STC remained the same. The authors interpreted this result to mean that expression of LTP did not involve an increase in the amount of transmitter released, since only fEPSP was altered. The alternative explanation was that the astrocyte downregulates transport in response to an increase in the STC. They disproved this by adding arachidonic acid which depressed the amplitude of the STC, but only transiently, so a long-lasting release of AA would be an unlikely mechanism for downregulation. To prove that the STC is sensitive to increases in transmitter release as a result of LTP, they studied LTP induction at the mossy fiber/CA3 synapse, where LTP expression is known to be presynaptic. They saw that forskolin, which potentiates mossy fiber input, enhanced the STC response. This enhancement was blocked by addition of DCG-IV which inhibits release from mossy fibers. Their study is interesting since it measures synaptic activity indirectly by looking at astrocyte activity. While their data is convincing, it seems that using a different cell type and molecular mechanism introduces many different unknown variables that could affect the results and their subsequent interpretations.
Fellin et al.
The authors of this paper were addressing the cause for slow inward currents (SICs) that they measured in CA1 pyramidal neurons. They recorded from pyramidal cells in the presence of TTX to block action potentials. They determined that SICs were mediated by NMDA receptors and contained no AMPA component. They then used DHPG, a group I mGluR agonist, in order to stimulate the pyramidal cells and find a source for the excitation. Using calcium imaging, they found that astrocytes in the hippocampus had elevated levels of intracellular calcium whenever DHPG was bath applied in the presence of TTX. They also applied TeNT which blocks synaptic transmission, to prove that SICs were not due to synaptic glutamate release. Previous studies have shown that elevations in [Ca]i can elicit release of glutamate from astrocytes, and the authors hypothesize that this non-synaptic release of glutamate is the cause for the NMDA-mediated SICs in the pyramidal neurons. To prove this, they stimulated the astrocyte with DHPG and recorded SICs in pyramidal neurons that were abolished in the presence of AP5. Next, they used caged calcium in order to determine whether stimulation of individual astrocytes can elicit SICs in adjacent pyramidal cells. UV photolysis caused detectable calcium wavefronts in astrocyte dendrites that were in close proximity to pyramidal neurons, which elicited SICs in these adjacent cells. Using ifendprodil, an NR2B antagonist, they determine that the glutamate released by the astrocyte acts preferentially at extrasynaptic NMDA receptors. The most important result they arrived at was that the glutamate released from these astrocytes caused neurons within its vicinity to synchronize their responses. The involvement of astrocytes in establishing neuronal synchrony reveals a novel mechanism used by the brain to coordinate the activity of a sub-population of neurons. They go on to claim that this synchrony can also occur under physiological conditions. A question I had hoped they would address was the mechanism through which DHPG elicits increased in [Ca]i in astrocytes.
Diamond et al.
In this article, the researchers examine whether transporter currents in astrocytes are sensitive to changes in glutamate release from Schaffer collateral/commissural terminals. They hypothesized that they would be. The researchers recorded whole-cell recordings of STCs from astrocytes and field EPSPS. To manipulate p, the probability of release, they used paired-pulse facilitation, lowered the Ca-Mg ratio, and used high frequency stimulation. They also applied glutamate rapidly to outside-out patches and applies nitroprusside. They found that when excitatory responses were isolated, the astrocytes responded to stratum radiatum stimulation with a biphasic response. They found that when p was increased, paired-pulse facilitation decreased. They concluded that LTP doesn't alter the amount of glutamate released during synaptic stimulation and that astrocyte STC amplitude can act as a measure for changes in p.
Fellin et al.
In this paper, the authors characterize how NMDA receptors and astrocytes mediate neuronal synchrony. They hypothesize that extrasynaptic NMDARs act as targets for glutamate from astrocytes, and that astrocytes can evoke synchronized neuronal responses. To test these ideas, the researchers perform HFS on Schaffer collaterals, perform patch clamp recordings with applications of various toxins, and use calcium imaging and uncaging. They find that stimulation of Schaffer collaterals results in SICs mediated by NMDARs, that these SICs are due to non-synaptic glutamate, that the glutamate's source is nearby astrocytes, that the NMDARs in question are extrasynaptic, and that this mechanism induces synchronized responses.
Glutamate Release Monitored with Astrocyte Transporter Currents During LTP
In last week’s papers, we studied whether LTP in the hippocampus is a presynaptic or postsynaptic phenomenon. In this paper, the authors measure glutamate release by recording synaptically activated transporter currents in hippocampal astrocytes. In the brain, almost all glutamate is transported into glia before it is “repackaged” in the neuron, and so an increase in glutamate transport would imply an increase in released glutamate. The authors show that LTP in the CA1 does not involved an increase in glutamate release during synaptic stimulation. They first use voltage-clamps, current-clamps, and whole-cell recordings of astrocytes to illustrate the evoked synaptic activity in the hippocampus. They find a biphasic current due to the extracellular field potential and the transporter mediated current. They also show reliable and predicted changes in STC during paired-pulse facilitation (increases glutamate release) and decreasing Ca++ levels (decreases glutamate release). These findings indicate that the amplitude of the STC is a reliable measure of the changes in the amount of glutamate release. They then show that LTP induction does not result in an increase in pr or n, but are wary of the possibility that an increase in the STC is offset by a concomitant downregulation of transport mediated by arachidonic acid or nitric oxide. They find that AA effectively reduces a glutamate elicited current, but that it is reversible and transient, and that NO has no effect on the current. However, it is unlikely that physiological AA has any effect on the STC, as it would require that the induction of LTP result in a substantial, long-lasting increase in extracellular AA. While the beginning figures in this paper focus on CA1 neurons, the authors also address CA3 pyramidal neurons. In these, the mechanism for LTP expression is generally agreed upon to involved increased glutamate release. Thus, mossy fiber-evoked STCs from astrocytes in the stratum lucidum of CA3 may be increased in size after LTP induction. Stimulation of stratum lucidum 100 μm from the astrocyte, while it elicited a long-lasting potassium component, failed to elicit an STC. For this reason, the authors used forskolin, which induces glutamate release, to show STC potentiation in CA3 astrocytes. Overall, the authors propose that LTP is expressed by the addition of functional AMPA receptors to previously “silent” synapses expressing only functional NMDA receptors, as was proposed last week.
Neuronal Synchrony Mediated by Astrocytic Glutamate through Activation of Extrasynaptic NMDA Receptors
According to the neuron doctrine, the only functional units of the nervous system are the neurons. As such, only neurons should be able to release neurotransmitter and cause excitation. The authors of this paper, however, provide support for a distinct mechanism of neuronal excitation and synchrony in which astrocytes play a key role. They show that astrocytes can release glutamate as a result of Ca++ activation, and that this glutamate acts preferentially on extrasynaptic NMDA receptors in a synchronous manner. This glutamate causes slow inward currents that can occur spontaneously. The authors use DHPG, a group I mGluR agonist, as well as TTX and TeNT to show that SICs arise from nonsynaptic origins and activate NMDARs. Using Calcium imaging, they show that SICs correspond to increases in astrocyte Ca++i, and that the DHPG response in astrocytes is not blocked by D-APV. The causal link between astrocyte Ca++i increases and NMDAR-mediates responses in neurons was measured by Ca ++ photolysis. A photolysis-induced Ca++ wavefront arrived at the pyramidal cell contemporaneously as the SIC and repetitive stimulation of the astrocyte resulted in neuronal responses with similar latency, demonstrating that astrocytes are responsible for SIC initiation. Perhaps because many drugs (such as amphetamine) operate through a reversal operation of transporters, the authors tested the possibility that glutamate transport was being reversed in astrocytes. They found that this was not the case, and that glutamate acted on extrasynaptic NMDAR NR1/NR2B complexes. The results here are somewhat confusing. The authors use ifenprodil, a NR2B antagonist, to show that SIC amplitude is severely reduced, but that NMDA-mediated EPSCs evoked by SC stimulation are only slightly affected. Why should this be the case? Should not both of these be significantly reduced? The authors do, however, show that Ca++i is coupled among astrocytes and neurons, and that synchronous SICs are often paired in neurons and due to astrocyte Ca++ elevations. Experiments have shown that neuronal glutamate release can cause Ca++ oscillations in astrocytes, but the reverse has not been tested until now. The level of astrocyte-neuron coupling exhibited in this paper has also not been shown to be so high. Overall, the authors conclusively reveal that glutamate released from astrocytes acts on extrasynaptic NMDA receptors to promote synchronized activity. This action may have a role in the formation of functional modules in the brain. However, since the release of glutamate is involved, it can also have a role in epileptic and excitotoxic events during malfunction.
Neuronal Synchrony Mediated by Astrocytic Glutamate through Activation of Extrasynaptic NMDA Receptors
This study examined the finding that D-AP5-sensitive slow inward currents (SICs) were generated in CA1 pyramidal neurons in response to DHPG, an mGluR agonist, even when synaptic transmission was blocked by tetanus neurotoxin (TeNT) and action potentials were blocked by TTX. If synaptic transmission was blocked, how did an mGluR response lead to an NMDAR response? The researchers pursued the possibility of extrasynaptic transmission mediated by astrocytes. They looked at the response of astrocytes and found that DHPG caused elevated intracellular calcium levels. They then used photolysis to evoke increased calcium levels in a single astrocyte and recorded the responses of a single pyramidal neuron. They found that the calcium wavefront in the astrocyte reached the dendrite of the pyramidal neuron at the same time that the pyramidal neuron responded with an SIC. They then investigated the mechanism for release, blocking the glutamate transporter in the astrocyte using TBOA. This did not block SICs, suggesting that a reversal of the glutamate transporter is not the mechanism for release. They then examined the mechanism of the response, finding that NR2B blockade with ifendopril nearly eradicated the response,, suggesting that extrasynaptic NMDAR are responsible for mediating the SICs. They next examined the synchronization of the neuronal response using paired recordings in the pyramidal neurons while stimulating astrocyte glutamate release using DHPG, calcium, or SC stimulation. SC stimulation was stopped and TTX was added to prevent synaptic signaling, but the glutamate that had already been released was sufficient to activate the astrocyte-mediated SIC response. All synchronous responses were abolished by D-AP5, suggesting that the mechanism was NMDAR mediated, not due to gap junctions. They repeated the recordings at more physiological conditions and still found that synchronous SICs were present. They were not sure how this synchrony arises, but they hypothesized that either one astrocyte releases glutamate and the extrasynaptic NMDARs on the dendrites of different neurons are all close enough to be activated simultaneously in response, or that there is simultaneous glutamate release from two different sites, either on the same or different astrocyte. I think that the latter option seems less feasible, because what would cause that synchronous release? If it was a random coincidence, then why were the synchronous responses so common? Anyway, I thought that overall the paper was extremely thorough and it overturned the dogma that glia do not participate in neuronal signaling.
Glutamate Release Monitored with Astrocyte Transporter Currents during LTP
This paper discussed a new method for determining whether LTP is induced by presynaptic or postsynaptic mechanisms in the CA1 region of the hippocampus. They recorded glutamate currents through the glutamate transporters on astrocytes during uptake of transmitter in synaptic transmission. They verified that this technique was sensitive to changes in release probability by measuring currents due to paired-pulse facilitation, lowered calcium to magnesium ratio, and a burst of high frequency stimulation. They found that the STC was actually more sensitive to these conditions than the field postsynaptic potentials. They then tested STC due to LTP by recording a baseline using KYN, which blocked fEPSPs, then washing it out and stimulating with three bursts, then adding KYN again after 15 minutes. They found that there was no change in STC. They then examined the potential downregulation of transporter due to AA or NO. AA had too short an action, and NO did not have an effect. Thus, they concluded that the lack of change in the STC was not due to less transporter action and increased transmitter, but rather a lack of change in transmitter. I thought that this paper was very creative. Last week, none of us came up with this idea for measuring presynaptic release versus postsynaptic response, but I think that it was effective.
Diamond: Glutamate Release Monitored with Astrocyte Transporter Currents during LTP:
The debate is still on as to whether LTP causes increased transmitter release, or increased post-synaptic sensitivity- so Diamond et. Al, measure glutamate currents in transports on astrocytes (which soak up excess glutamate released from the pre-synaptic side) in order to better understand how LTP affects the amount of glutamate released. The researchers record STCs (synaptically activated glutamate transporter currents) in the astrocytes of rate straitum. Overall, they find that Pr (probability of release) and N (number of release sites) will change STC, however, induction of LTP does not; therefore, LTP does not alter the amount of glutamate released on the pre-synaptic side. Three mechanisms to test whether a change in Pr changes STC were used: (1) paired pulse facilitation (2) lowering Ca2+ (3) burst of high frequency stimulation. Overall, changes to the Pr had the expected consequential effects. Therefore, if LTP causes increased Pr, the STCs should increase- but this paper shows that LTP does not change STCs. To be careful that this increase in STC is not just off-set by a downregulation of transporter, they use AA and NO to see if this compensatory mechanism is in place; however, neither messenger had an effect on the STCs. Overall, I thought this paper was very careful to consider other options besides that confirming their hypothesis and to test them, before ultimately concluding that LTP does not increase Pr because STCs are not increased.
Fellin: Neuronal Synchrony Mediated by Astrocytic Glutamate through Activation of Extrasynaptic NMDA Receptors:
Fellin et.al, show that astrocytes release glutamate which can coordinate neurons through activation in a synchronous fashion. They also hypothesize that extrasynaptic NMDA receptors might bind preferentially with glutamate that is released from a non-neuronal source (astrocytes). They are interested in measuring the SIC (slow inward current) and using TTX to block neuronal sources, still see a SIC, therefore concluding that these currents represent NMDA activation by glutamate not from a synapse. They also performed Ca2+ imaging experiments with DHPG, to find that SIC in neurons and Ca2+ increases in astrocytes come from the same event. Although they have already concluded that the astrocyte-released glutamate activates NMDArs of non-neuronal origin, they test to see whether this glutamate can activate neuronal NMDArs, which seems to be the case. I thought the most important part of this paper was their decision to test whether these events would still happen under physiological conditions. Even with GABAergic (inhibitory) transmission and Mg++ in solution, synchronous SICs and activation of neurons was observed; therefore, this mechanism seems to have great physiological implications. Overall, this new found mechanism for wide-spread synchrony questions this use of gap-junctions- I would have liked them to block gap-junctions and run these same tests to determine the link between SICs and neuronal activation.
“Glutamate Release Monitored with Astrocyte Transporter Currents during LTP”
In a continuation of the presynaptic vs. postsynaptic expression of LTP controversy, Diamond et al. record glutamate currents in hippocampal astrocytes to measure the evoked glutamate release from Schaffer collaterals in CA1. While most researchers around this time measured the postsynaptic response to support their argument for a pre- or postsynaptic locus of expression, Diamond et al. use a relatively independent technique. They record glutamate transporter currents (STCs) in astrocytes to monitor glutamate release from Schaffer collaterals.
First, they establish the existence of STCs in astrocytes. Then they use three techniques to alter release probability (pr): paired pulse facilitation, reduction of Ca:Mg ratio, and high frequency stimulation. They also look at STC responses to a change in the number of release sites (n) by increasing stimulus strength. Results of such manipulations demonstrate that the STC amplitude depends on the amount of released glutamate, either as a result of changes in Pr or n. When Diamond et al. induce LTP, they note no change in STC. They present an alternative explanation for the lack of change in STC. Specifically, they state that an increase in STC is balanced by down-regulation of transport. Based on previous research, they hypothesize that arachidonic acid (A) or nitric oxide (NO) modulates this compensatory response. However, AA does reversibly inhibit the STCs but the effect is so transient, that AA is not a likely candidate. NO does not affect the STCs at all. Finally, Diamond et al. consider LTP results in CA3, as LTP in this region is due to an increase in glutamate release. Detection of a transporter current is difficult, but application of forskolin, an adenylyl cyclase activator, potentiates mossy fiber input to CA3 neurons. Ultimately, the authors’ conclusions are straightforward. They observe parallel changes in the fEPSPs and STCs while altering pr and n. However, LTP does not affect STC amplitude. Therefore, it seems that the locus of LTP expression is not related to an increase in pr or n. In their discussion, the authors mention complications in comparing STCs and fEPSPs. Although they believe that it is unlikely that astrocytes do not sense the LTP potentiated synapses, they do not present convincing evidence otherwise.
“Glutamate Release Monitored with Astrocyte Transporter Currents during LTP”
Historically, researchers viewed glia as support cells for neurons. Fellin et al. present a novel functional role for glia, as they explore the slow inward currents (SICs) observed in CA1 neurons and their relationship to astrocyte glutamate release. They use such techniques as patch clamp recordings, confocal microscopy, and photolysis to answer their overarching question regarding the link between astrocyte glutamate and extrasynaptic NMDAR.
First, they establish that SICs result from NMDAR activation by non-neuronal glutamate. They test high frequency stimulation of Schaffer collaterals (SCs) and note that the NMDAR antagonist D-AP5 reversibly blocks the SICs. They verify that mGLuR stimulation can trigger SICs via use of metabotropic glutamate agonist DHPG. The glutamate is non-neuronal, as application of tetanus neurotoxin blocks neurotransmitter release but not SICs. Next, using Ca imaging experiments, Fellin et al. find that DHPG causes an increase in Ca concentration in astrocytes, which is associated with an NMDAR response in pyramidal neurons. Employing a caged glutamate technique, they are able to elicit an increase in Ca concentration in single astrocytes and measure the resulting SIC in the adjacent neurons. Glutamate is not released through a reversal of the glutamate transporters, as inhibition of the glutamate transporter does not abolish DHPG triggered SICs. Fellin et al. show that the NMDARs are mostly composed of the NR1/NR2B complex by demonstrating the drastic reduction in SIC amplitude in the presence of NR2B antagonist ifenprodil. Interestingly, there is a great deal of synchrony in terms of SICs in different neurons due to astrocytic glutamate. However, this synchrony is not due to electrical coupling. Such synchrony parallels the synchrony in Ca elevation in the various neuronal domains tested. Finally, the authors verify the physiological relevance of the astrocytic glutamate evoked SICs. Overall, I agree with their interpretations regarding the astrocytic origin of NMDAR mediated responses in CA1 pyramidal neurons, mostly because the authors are extremely thorough in terms of their use of various controls and alternative explanation. Also, it would have been interesting had the authors been able to better clarify the exact mechanism of glutamate release that elicits SICs.
“Glutamate Release Monitored with Astrocyte Transporter Currents during LTP”--- Diamond et al
This paper was focused on determining the mechanism by which LTP in CA1 region of the hippocampus occurs. Synaptically activated glutamate transporter currents (STCs) were used to monitor postsynaptic glutamate release. In order to see if TCs are sensitive to changes in glutamate release, Diamond et al simultaneously recorded fEPSPs and STCs using a whole-cell voltage clamp. The results from the voltage clamp were then converted to potential units and the authors found that with picrotoxin (blocks the gamma-aminobutyric acid-activated chloride ionophore) you could elicit a complex, biphasic response. The field component (the first part of the biphasic response) and most of the slower K+ current could be eliminated with bath applications of kynurenate. Everything could be blocked with THA and DHK, which showed the TCs were glutamate-mediated. Then Diamond et al altered p to test STC sensitivity. The authors found that with elevated p facilitation of the STC was decreased, and that STC was sensitive to changes in n, and then using these results the authors concluded that the STC was more sensitive to changes in p and n than the fEPSP. Next the authors tested whether it was an increase in p or n that followed LTP by eliciting STCs and fEPSPs with two different pathways. Since the authors found that LTP didn’t significantly alter n or p (and confirm this with AA and NO), the authors concluded that LTP therefore does not alter the amount of glutamate released during synaptic stimulation. Their findings counter the idea of LTP being used to ‘unsilence’ immature, silent neurons, and instead the authors suggest that these results could be explained by a spill-over theory.
“Neuronal Synchrony Mediated by Astrocytic Glutamate through Activation of Extrasynaptic NMDA Receptors”--- Fellin et al
The authors find that NMDA receptor response occurs in synchrony in multiple CA1 neurons. Fellin et al used intense stimulation of SCs and found that in the absence of extracellular Mg2+ there was a delayed trigger and a spontaneous, slow inward current (SIC), which occurred at low frequencies. The fast AMPA component was always absent from the SIC. Since SIC are due to activation of NMDAR, the authors next questioned whether stimulation with DHPG can trigger SICs. Using a patch-clamp recording and DHPG, the authors found that SICs were indistinguishable from SC-stimulated SICs. The authors also found desensitization of AMPAR is the reason for the absence of an AMPA component by using D-AP5 and TTX. Fellin et al also looked at SICs in varies postnatal weeks and found no change. Next in order to study astrocyte response to DHPG, the authors used Ca2+ imaging experiments, and concluded that SICs and transient [Ca2+]i are echoing the same event. Then Fellin et al, confirmed the existence of a causal link between [Ca2+]i increases and NMDAR-mediated responses in neurons using photolysis (loading astrocytes but not neurons with caged Cat2+) and concluded that this current is mediated by NMDAR activation. The authors also checked to make sure that the reverse operation of glutamate transporters did not mediate SICs with TBOA. Next, Fellin et al used ifenprodil to show that astrocytic glutamate prefer the NR1/NR2B complex. The authors also found high synchrony between SICs using paired-recordings, and used additional confocal imaging to confirm that glutamate release could act on multiple neurons. Next, the authors tested whether activated astrocytes can signal back to neurons and trigger NMDAR-dependent synchronized responses in neurons. Using a Mg2+-free and picrotoxin solution, and found delayed responses and neuron-neuron synchrony. Then Fellin et al tested whether these responses could be triggered by astrocytic glutamate in the absence of picrotoxin (to mimic physiological conditions) and found no significant change. Fellin et al therefore, concluded that activation of [Ca2+]i elevations in CA1 pyramidal neurons elicits repetitive, highly synchronized NMDAR-mediated responses.
Fellin et al.
The question the authors of this paper are asking is if the NR1/NR2B complex NMDA receptors, located mostly extrasynaptically, are the targets of astrocyte glutamate release. They also look into the mechanism and purpose of this glia-neuronal interaction. They hypothesize that increases in astrocyte calcium and the resulting NMDAR response can be elicited by stimulation of Schaffer collaterals and that they have a role in synchronizing neuronal firing, perhaps during high frequency stimulation. They find that the activation of increases in calcium in astrocytes causes the release of glutamate, which elicits NR1/NR2B complex-mediated responses in CA1 pyramidal neurons, and they provide evidence that astrocytic glutamate release can cause synchronous activity in several neurons. This helps overturn the dogma that glia are passive and serve only supportive and nutritive functions, as it shows a direct involvement of astrocytes in synaptic transmission. To find these results, they first characterize the slow inward currents (SICs), they use high frequency stimulation of Schaffer collaterals (SC) in the absence of magnesium. They find that these are mediated by NMDARs (they are blocked by AP5), they can be triggered by mGluR activation (by agonist DHPG), are very consistent in the absence of AMPAR activation (in the presence of NBQX), are unmasked because of AMPAR desensitization (cyclothiazide masks them), and finally that they are non-neuronal (they are not blocked by TTX). They establish that they are elicited by astrocytes using caged glutamate and finding that the SICs correspond very well with the increase of calcium in the glia. They found that the SICs were mediated by the NR1/NR2B complex by abolishing them with ifenprodil and also showed that they were unaffected in the presence of a glutamate transporter inhibitor. They then show the synchrony of increases in calcium in astrocytes and nearby neurons using calcium imaging, and they show that stimulus trains in SCs evoke calcium elevations in both CA1 neurons and astrocytes. I agree with their results for the most part, in part because they did many control experiments, showing that the synchronous activity was not due to gap junctions, for instance, and that the caged glutamate experiment was not due to the uncaging process alone and did not work in neurons. I think their explanation of the mechanism of the synchronous activity is incomplete, but not unconvincing.
Diamond et al.
In this paper, the authors use a new method to address the question of whether the induction of LTP is post- or pre-synaptic. They find that, in the case of CA1 neurons, the induction is not due to an increase in release probability (Pr) or the number of presynaptic vesicles (n). More important is that they measure the glutamate release from Schaffer collateral terminals by recording the glutamate transporter currents (STCs) of nearby astrocytes after LTP induction. In theory, this technique would allow the research to measure the release separately from the post-synaptic response (which was measured by field potentials in this case). They first show the measure is sensitive to changes in release probability and vesicle number by comparing the field potentials and STCs after using high frequency stimulation, paired-pulse facilitation, lowered calcium to magnesium ration, and raising the stimulus strength. They then induce LTP and show that the STC does not change, which could be interpreted as a lack of change in either release probability or vesicle number. They offer the possibility of interference by arachidonic acid or nitric oxide, but they show that neither is likely (nitropusside has no effect and the effect of arachidonic wears off very quickly). They then compare these results to the results in CA3, which is generally agreed to have an LTP that uses presynaptic mechanism. However, they do not detect much of a different in the STC until they use forskolin, which potentiates the synapse much more than LTP would. I do not think the results are very convincing that this is a very good measure of presynaptic release. The experiment with CA3 neurons makes the measure seem imperfect, since they had to use non-physiological potentiation to see changes in the STC. They also mentions several different interpretations and problems with the use of the measurement, one of which is that the astrocytes sense synaptic activity different from the field electrode, so the comparison in flawed. They argue that in this case the astrocyte was not saturated, but it seems that it would be a saturable measure. It is an intriguing measurement, but not a totally convincing one.
Malgaroli & Tsien
The authors of this paper observed the appearance of LTP in the CA3-CA1 synapse in the hippocampus, and associated it with an increase in the frequency of miniature postsynaptic potentials. They invoked LTP by the local application of an Mg2+-free glutamate solution and noted an increase in the frequency of spontaneous events. Mini frequency could also be potentiated with LTP induced by the promotion of exocytosis from the presynaptic neuron via hypertonic challenge. Potentiation was not seen if the experiment was performed in a Ca2+-free medium or with the postsynaptic cell hyperpolarized. These facts implicated NMDARs in the process, and potentiation was accordingly blocked by the NMDA blocker MK-801. No significant changes in mini amplitude or time course were noted, although the authors proved amplitude changes were detectable. They also attempted to prove that potentiation did not simply result from the recruitment of inactive receptor clusters by noting that the response to enhanced vesicle release (from hypertonic challenge) did not scale up to the same extent as mini frequency potentiation. Removal of Ca2+ after the experiment did not abolish potentiation. The authors hypothesize that LTP is induced postsynaptically, but expressed presynaptically.
Isaac et al.
This paper investigated the existence of “silent synapses,” which contain NMDARs but no AMPARs. As a result, they will not display EPSCs at negative membrane potentials, but only when depolarized. These ESPCs are also blocked by the NMDAR blocker D-APV. The authors located such synapses by providing stimulation at a low level, where no EPSCs were noted at resting membrane potential. They then depolarized the postsynaptic cell and continued stimulating at the same rate, leading to the sudden appearance of EPSCs via NMDARs. They also noted an increased NMDAR component of EPSCs by recruiting NMDARs with small increases in stimulus intensity and noting the changes in EPSC size both early and late during the EPSCs. A greater change was seen late, in accordance with the slower time course of NMDA currents. The authors then invoked LTP and noted the appearance of new AMPA EPSCs following the pairing protocol. This experiment was compared to controls in which pairing was blocked by keeping the postsynaptic cell hyperpolarized or in which D-APV was used to block LTP. Both methods abolished the new AMPA-mediated EPSCs.
Glutamate-induced long-term potentiation of the frequency of miniature synaptic currents in cultured hippocampal neurons
Malgaroli and Tsien were interested in exploring the mechanism of increased synaptic strength at hippocampal CA3-CA1 synapses from LTP initiation. Specifically, they wanted to know whether presynaptic transmitter release is enhanced during LTP, using whole-cell recordings to measure miniature synaptic events (minis) and postsynaptic responsiveness. First, they focally stimulated presynaptic neurons and then induced synaptic potentiation by exogenous or synaptically released glutamate to the postsynaptic cell, finding that this strongly increased both the amplitude of the evoked synaptic current and the frequency of the spontaneous excitatory postsynaptic potential (EPSC). Next, the authors examined the involvement of the postsynaptic cell by varying its membrane potential during glutamate application, finding that hyperpolarization abolished mini frequency potentiation. They supported this finding using MK801, a selective blocker of the NMDA receptor pore, as it had the same abolishing effect. Next, the authors compared mini EPSC amplitude during a control period and a period after induction of mini frequency potentiation, discovering that while there was a twofold increase in the number of events with potentiation, there was no change in the time course or mean amplitude of the mini. The authors then further tested for changes in postsynaptic responsiveness by application of exogenous ligands such as glutamate and 2-(aminomethyl)phenylacetic acid (AMPA) to a postsynaptic area in the region where the enhancement was induced. While there was no change in postsynaptic currents, they did find a clear mini frequency potentiation. The authors went on to argue that these findings were related to the hypothesis that LTP results from all-or-nothing recruitment of inactive postsynaptic receptor clusters, in that it assesses the responsiveness to exogenous ligands but may include additional extrasynaptic receptors. Thus, to focus more specifically on synaptic receptors, the authors provoked vesicular release of glutamate from presynaptic nerve terminals by brief hypertonic challenges and measured the mini frequency. They hypothesized that if potentiation arises from recruitment of these inactive receptor clusters, the response to enhanced vesicular release should be potentiated to the same degree as the base mini frequency. However, because this was not the case, they argued that the mini frequency potentiation originates presynaptically. Finally, the authors examined the role of calcium through external Ca2+ removal on mini frequency before and after the induction of potentiation and found that presynaptic enhancement was not reduced by calcium removal, suggesting an alternate mechanism for mini frequency potentiation.
Evidence for Silent Synapses: Implications for the Expression of LTP
Isaac et al. also tackle the question of whether the mechanism of increased synaptic strength at hippocampal CA3-CA1 synapses from LTP initiation is predominantly presynaptic or postsynaptic. They take a different approach in exploring the existence of silent synapses in hippocampal CA1 pyramidal cells as a potential postsynaptic mechanism. First, the authors wanted to establish the existence of silent synapses that express NMDARs but not functional AMPARs, as it could be the case that the stimulus strength is so low that they simply are not activated. They did this by holding the postsynaptic cell at -60 mV, and after obtaining a small EPSC, decreased the stimulus intensity so that no EPSCs were detected. They then depolarized the cell to +30 mV and stimulated at the same intensity, which produced EPSCs that were completely blocked by D-APV (indicating that they were mediated by NMDARs). They then hyperpolarized the cell to -60 mV, finding an absence of EPSCs that was not affected by CNQX (AMPAR antagonist). The authors then examined silent synapse recruitment, where they depolarized the cell to +30 mV, and then increased the stimulation strength in small increments to note the EPSC amplitudes at early (EPSCE) and late (EPSCL) time points that presumably reflect AMPAR and NMDAR EPSCs, respectively. At first, the increase in stimulus strength had no effect on EPSCE, even though EPSCL increased. But later, at higher stimulus strengths, both components of the EPSC increased. They argue that the different effects of stimulation strength on the two EPSCs show that higher stimulus strengths is necessary to activate synapses with both NMDARs and AMPARs. Now having found that there are synapses on CA1 cells that contain NMDARs but no AMPARs, the authors wanted to know whether LTP could modify these synapses. They held the cell at -60 mV, reducing stimulus intensity to a level at which no AMPAR EPSCs were observed, and then depolarized the cell to -10 mV while maintaining the same stimulation rate, and then hyperpolarized the cell back to to -60 mV, again maintaining a constant stimulation rate. Through this pairing LTP induced protocol, they found the appearance of AMPAR EPSCs. Finally, it is worthy to note that the authors performed several controls, such as ensuring that AMPAR EPSCS did not simply appear spontaneously during the course of the experiment due to a change in stimulation conditions, and using D-APV during the pairing protocol to block NMDA receptors (and thus the induction of LTP), which prevented the appearance of AMPAR EPSCs.
Isaac et al.
The context of this paper is the debate about the mechanisms of LTP (and LTD) induction and maintenance. The authors of the paper identify an agreed-upon result: that the failure rate of synaptic transmission decreases with LTP. However, this decrease could be due to either a presynaptic increase in release probability or the insertion of AMPARs into the postsynaptic membrane. The authors hypothesize that the latter is the mechanism of LTP induction and moreover, that this hypothesis would require the existence of NMDAR-only synapses, or “silent” synapses. The authors test this hypothesis by using low stimulation strengths to stimulate only silent synapses at -60mV, resulting in no EPSP. Then, depolarization of the postsynaptic cell (to +30mv) paired with low stimulation removed the Mg++ block and resulted in an EPSC, which was subsequently blocked by APV application. The base level of stimulation did not change when the cell was again hyperpolarized and CNQX was applied, showing that AMPARs are still not present. This shows that silent – NMDAR only – synapses do exist. They next induced LTP using a pairing protocol and this induced a new current, presumably a current from the newly-inserted AMPARs. Additionally, they did several control experiments, the most important of which was the blocking of LTP by APV application before the pairing. I agree with these findings because they are convincing in and of themselves and also agree with previous data about the changing of CVs during LTP. On the other hand, I do not think the existence of postsynaptic changes during LTP induction precludes the possibility of presynaptic changes as well.
Malgaroli et al.
This paper is addressing the same debate as Isaac et al., namely if LTP works through a pre- or postsynaptic mechanism. This paper agrees that the potentiation is initiated through a postsynaptic mechanism but also asserts that its maintenance is through a presynaptic mechanism. They found that with LTP, the mini frequency increases but the amplitude is steady, which could reflect either enhanced release or the recruitment of receptors to the postsynaptic membrane. They then tested the postsynaptic cell’s responsiveness by puffing in receptor agonists to already potentiated synapses, which did not increase postsynaptic current, as it should have if there were increased numbers of receptors postsynaptically. Instead, they conclude, the mechanism for mini potentiation is increases vesicular release. They also show that it requires calcium entry and is dependent on NMDARs. I agree with their results but wonder about how universal they are. They are careful to say that mini potentiation seems to be presynatpic in their experimental setup, and I wonder if there are other experiments that see, rather than an increase in mini frequency, an increase in mini amplitude. I also wonder what mechanism could communicate the induction of LTP in the postsynaptic cells to the presynaptic cell for maintenance.