Functional spreading of hyperexcitability induced by human and synthetic intracellular Aβ oligomers

Article activity feed

1. 

   eLife

   Feb 12, 2021

   Reviewer #3:

   This is an interesting manuscript in which the authors have investigated the effect of intracellular injection of oligomeric beta-amyloid into hippocampal neurons both in cultures and adult animals. They find that starting from 500 pM, intracellular injection of oligomeric beta-amyloid rapidly increases the frequency of synaptic currents and higher concentrations potentiate the AMPA receptor-mediated current. Both effects were PKC-dependent. Furthermore, they find that following PKC activation there is release of NO which in turn increases release of neurotransmitter not only in the nearby pre-synaptic site, but also in neighboring cells. This suggests that intracellular injections of oligomeric beta-amyloid into the postsynaptic neuron can increase network excitability at a distance. The effect on neuronal excitability …

   Reviewer #3:

   This is an interesting manuscript in which the authors have investigated the effect of intracellular injection of oligomeric beta-amyloid into hippocampal neurons both in cultures and adult animals. They find that starting from 500 pM, intracellular injection of oligomeric beta-amyloid rapidly increases the frequency of synaptic currents and higher concentrations potentiate the AMPA receptor-mediated current. Both effects were PKC-dependent. Furthermore, they find that following PKC activation there is release of NO which in turn increases release of neurotransmitter not only in the nearby pre-synaptic site, but also in neighboring cells. This suggests that intracellular injections of oligomeric beta-amyloid into the postsynaptic neuron can increase network excitability at a distance. The effect on neuronal excitability would involve AMPA-driven synaptic activity without altering membrane intrinsic properties. The conclusions are sound. However, there are two main aspects of the observed phenomenon that have not been taken adequately into account, or have been avoided by the authors. The authors have not investigated the effects of application of oligomeric beta-amyloid into the extracellular space and the presynaptic neurons, two other compartments of the synapse. They might have performed experiments comparing findings from experiments with intracellular injections of oligomeric beta-amyloid into the post-synaptic neurons, with effects of extracellular application and those of injections into the presynaptic neuron.

   Additional minor concerns are related to the following issues:

   a) The raw data on Figure 3 suggest that not only excitatory transmission is affected but also inhibitory transmission is somewhat modified. Measurement of the charge might be misleading.

   b) This reviewer is not clear on the meaning of the following sentence in the discussion "Contrary to previously published data using extracellular Aβ or with more chronic application models [45-50], we did not find any synaptic deficits". The current work shows synaptic changes!

   c) There is a mistake in the numbering of figures in the discussion. The paper has no figure 11. When referring to figure 10, they must mean something else.

   d) The model on Figure 10 needs work. The authors should explain what various elements of the drawing mean, or better label them directly on the figure.

   Read the original source
2. 

   eLife

   Feb 12, 2021

   Reviewer #2:

   Epilepsy is often an early sign observed in Alzheimer patients and there are several mechanisms that may contribute to this hyperexcitability. In this study, the authors focused on an important observation suggesting that intracellular Amyloid beta, a protein often found in plaques in the brain, is found early on inside neurons of the hippocampus, the learning and memory center of the brain. Interestingly, when unique early forms of Ab named oligomers were introduced inside neurons, the cells and surrounded circuits became hyperexcitable. This increased excitability was mediated mainly by the release of glutamate on AMPA glutamate receptors. Remarkedly, these excitatory effects were triggered by intracellular amyloid oligomers through a retrograde signal named nitrous oxide. This manuscript suggests that early stages of the …

   Reviewer #2:

   Epilepsy is often an early sign observed in Alzheimer patients and there are several mechanisms that may contribute to this hyperexcitability. In this study, the authors focused on an important observation suggesting that intracellular Amyloid beta, a protein often found in plaques in the brain, is found early on inside neurons of the hippocampus, the learning and memory center of the brain. Interestingly, when unique early forms of Ab named oligomers were introduced inside neurons, the cells and surrounded circuits became hyperexcitable. This increased excitability was mediated mainly by the release of glutamate on AMPA glutamate receptors. Remarkedly, these excitatory effects were triggered by intracellular amyloid oligomers through a retrograde signal named nitrous oxide. This manuscript suggests that early stages of the disease may comprise significant increases in network excitability that may trigger a cascade of synaptic dysfunction and cognitive deficits such as memory loss.

   Here are my comments to strengthen the manuscript. Overall this is a strong study with an interesting take on the role of intracellular amyloid and how it contributes to increased network excitability in AD.

   There is an interest to determine the mechanisms responsible for the hyperexcitability often associated with familial and sporadic forms of Alzheimer's disease. Many have focused on possible reduction in inhibitory interneuron function as essential drivers of the increased excitability of the network. Although there exist a large number of investigations determining the effects of extracellular Ab on synaptic transmission, the intracellular effects of Ab and its contribution to disruptions of synaptic transmission remains less well understood. A couple of studies have shown that intracellular application of Ab (Ab42) induces decreases in long-term potentiation and basal synaptic transmission. In this study, the authors have investigated how intracellular Ab oligomers (iAbo) contribute to enhanced excitability in the CA1 region of the hippocampus. To do so, they have intracellularly applied human brain-derived and synthetic Ab oligomers through the patch-pipette in principal neurons recorded in vitro and in vivo.

   In this study, the authors show that intracellular application of intracellular Ab oligomers increased the frequency and the amplitude of excitatory currents and spiking in ex vivo hippocampal slices. Effects that were mimicked by human oligomers. The intracellular amyloid mediated effects were through the amplification of AMPAergic spontaneous activity and currents, and, to a lesser extent, spontaneous GABAA mediated currents. Miniature frequency and amplitude of AMPA-mediated EPSCs were also increased and were sensitive to PKC blockers. Interestingly, since intracellular Ab increased the frequency of EPSCs, which is a presynaptic effect, a signaling molecule is likely to be released postsynaptically to modulate presynaptic terminals. The hypothesis that the retrograde signal NO was involved by determining the sensitivity of NOS inhibitor L-NAME. L-NAME reduced the increased iAbo mediated frequency of spontaneous post-synaptic excitatory currents in cultured neurons. The L-NAME compound was shown to reduce the iAbo -mediated No from both the recorded and neighboring neurons providing further evidence that intracellular Ab oligomers triggered NO release and increased glutamate release. Increases in the excitability of CA1 pyramidal cells were also observed in vivo by intracellular application of AB oligomer. Overall, this is a well written study that demonstrates a novel perspective of the effects of intracellular Ab oligomers on CA1 principal neurons and suggests possible mechanisms underlying hyperexcitability.

   Novelty:

   1. use intracellular oligomers, synthetics and humans

   2. Showing that iAb oligo increased post and presynaptic AMPA-mediated EPSCs.

   3. The presynaptic increases in EPSCs were mediated by NOS and NO, this could potentially spread widely across the network.

   4. spontaneous IPSCs were also increased (through an undetermined mechanism).

   5. the iAbo increase in excitation was also observed in vivo.

   Questions:

   Intracellular Ab produces both an increase in EPSCs and IPSCs. However, in Fig 3, the IPSCs, measures using a charge transfer quantification, did not show a significant change in response to iAbo, in contrast to EPSCs. This spontaneous inhibition here was measured as charge transfer which depends on the amount of charges in time. I wonder why this was not significant since this measurement should have picked up a possible increase in spontaneous IPSCs?

   With regard to the inhibition, In the schematic on Fig. 10, I find this incomplete and slightly inaccurate since it shows one terminal releasing both glutamate and GABA with NO increasing both. While this is obviously an oversimplification, it's slightly inaccurate since NO was not directly shown to increase sIPSCs. Were NOS blockers able to disrupt the increase in sIPSCs? Moreover, there are many papers that have shown that PKC can also phosphorylate GABA receptors and increase their conductance. What could be the reason that this was not involved here? This needs to be discussed.

   The experiments were done in cultured neurons, in slices and in vivo. It's not always easily discernible in what conditions the experiments were done when reading the manuscript, especially when looking at the figures and figure legends. This should be at least stated in the figure legends. To help the reader, the conditions in which the currents were recorded (GABA and or excitatory receptor blockers, other ion blockers could be indicated in the figure legends to ease the comprehension of how the experiments were done and what was measured). In relation to this, was the sIPSC iAbo-mediated increases also blocked by L-NAME?

   In other studies, investigating intracellular application of Ab, such as the Ripoli et al., 2014 paper, showed that iAb produced significant reductions in EPSCs in their hippocampal neurons. What are the differences explaining this? This should be discussed. Similarly, Gulisano et al., 2019, showed that extracellular, but not intracellular oligo Ab had effects on excitability when it was applied extracellularly but not intracellularly. This should also be discussed.

   In the introduction, it's mentioned that the nature of hyperexcitability is unknown. I agree that it's incompletely known, but what is known is that there is a large variety of possible causes. For example, changes in GABAergic interneuron function (see Hijazi et Al 2019) is well known to be a contributing factor. There are many studies that have shown possible contributing causes of hyperexcitability, therefore, something IS known, and this should be identified in the introduction.

   How do these increases in synaptic transmission by applying pM concentrations of oligomers fit with the data showing that extracellular Ab oligomers of comparable concentrations decrease synaptic transmission through presynaptic reductions in glutamate release? This needs to be put into context and discussed.

   Read the original source
3. 

   eLife

   Feb 12, 2021

   Reviewer #1:

   This is an interesting study of the effects of intracellularly-applied amyloid beta (Ab) in primary hippocampal cultures of embryonic rats or in area CA1 of hippocampal slices or anesthetized rats that are less than 35 days old (therefore prepubertal). In vivo, whole cell recordings were made of CA1 neurons which is difficult and therefore a strength. Both synthetic Ab and human-derived Ab were applied by adding them to the internal solution of a patch electrode. Several interesting effects were documented, such as increased evoked and miniature EPSCs (mEPSCs) as well as some effects on IPSCs and neuronal properties. A major question is whether these effects were pharmacological or physiological.

   An intriguing finding was that the increased EPSCs was reduced by inhibiting a PKC-mediated effect of nitric oxide (NO). …

   Reviewer #1:

   This is an interesting study of the effects of intracellularly-applied amyloid beta (Ab) in primary hippocampal cultures of embryonic rats or in area CA1 of hippocampal slices or anesthetized rats that are less than 35 days old (therefore prepubertal). In vivo, whole cell recordings were made of CA1 neurons which is difficult and therefore a strength. Both synthetic Ab and human-derived Ab were applied by adding them to the internal solution of a patch electrode. Several interesting effects were documented, such as increased evoked and miniature EPSCs (mEPSCs) as well as some effects on IPSCs and neuronal properties. A major question is whether these effects were pharmacological or physiological.

   An intriguing finding was that the increased EPSCs was reduced by inhibiting a PKC-mediated effect of nitric oxide (NO). Furthermore, the effect of intracellular Ab on the recorded cell had effects on neighboring cells. Whether those were due to diffusion of NO, synaptic inputs from the recorded cell on neighboring cells, or release of Ab from the recorded cell was not clear. The authors suggested this is 'functional spreading of hyperexcitabiliity' similar to the way prions are spread transynaptically (actually this has been suggested for Ab too; see work by Karen Duff or Brad Hyman's groups) although this seems premature because the work that has been done with prions and Ab involves spread over a long time and a long distance relative to the results of the present study. Still the results are interesting and could be relevant in some way to the development of the disease or hyperexcitability.

   MAJOR CONCERNS

   One major issue is whether the results are relevant to Alzheimer's disease (AD) or represent interesting pharmacological data about what Ab can potentially do in some of its forms in normal tissue. The cultures are from embryonic rats and it is not clear how well they can predict what occurs in aged humans with AD. This issue is not only a question related to the preparation of tissue but the use of Ab intracellularly. It is not clear that synthetic or human Ab that is prepared outside the animal and used to fill electrodes to dialyze a cell is similar to the Ab generated in a cell of a person with AD. Independent of the methods to determine whether it is oligomeric outside the cell, once dialyzed it is not clear how it may change and where it would go. In AD Ab has a particular location and precursor where it forms and how it travels to the external milieu. As a product of its precursor APP, several peptides are produced besides Ab and many labs think they are as important as Ab in the disease. Although a strength to use atomic force microscopy to attempt to verify the form of Ab being used, it is not clear what form was actually in the dialyzed cell and how that compared to the form in AD.

   How this work relates to other studies that are similar is important. It seems that few other studies that have applied Ab are mentioned because few have studied it intracellularly. However, they are relevant because adding Ab has been shown to cause an increase in hippocampal neurons of excitatory activity at low concentration but at higher concentrations synaptic transmission is weakened. Many studies of mouse models of AD pathology suggest reduced synaptic transmission and plasticity, although many others show hyperexcitability, often without adding Ab at all.

   PKC and NO do a lot of things throughout the brain and body. How do the effects the authors have identified relate to all these other effects. For example, if PKC is activated by another mechanism, would it occlude the effects of Ab? What are the changes in PKC and NO in AD?

   ADDITIONAL CONCERNS

   I am not sure of the validation of Ab using the anti amyloid or 6E10 antibodies. The western blot shows a large region that both antibodies detect and the 6E10 antibody shows an even greater band. It is not clear what the large range of bands that are shown imply except nonspecificity. The antigen that the antibodies recognize should be stated exactly.

   Clarifying sample sizes throughout the study is needed.

   Do the cultures include interneurons? Are the excitatory and inhibitory neurons interconnected? This information will help interpret the results.

   The external solution for cultures contains 5.4 mM K+ which is quite high, and can induce hyperexcitability. Therefore it is important to be sure controls did not show hyperexcitability even after persistent recordings. Similarly, the use of 100uM AMPA and GABA seem very high. Justifying these high concentrations is important. They should lead to hyperexcitability and toxicity (AMPA) over time. Another point of concern is that the concentration of K+ for the slice work is 3 mM, much different than cultures. There are also differences in Mg2+ and Ca2+, making data hard to compare in the two preparations.

   Line 295 mentions 2 min recording periods were used to acquire sufficient events. One wants to know if this was done throughout the paper and if so, how many events per 2 min was considered sufficient?

   Terms related to intrinsic membrane properties and firing need to be explained much more because each lab has a slightly different method.

   In the statistics part of the Methods, why is Welch's ANOVA (followed by Games-Howell) used when variance was unequal. Usually the test to determine inequality is provided, so it is clear it was done objectively and with a reasonable test. Then if the data are unequal there is often a choice for a non parametric test, which is common. Some groups transform the data such as taking the log of all data values. If this reduces the variance between groups, sufficient to pass the test to determine inequality, it leads to a parametric test like a one-way ANOVA followed by Tukey's posthoc test.

   In the Results, Line 331 suggests that the authors think they know what a low concentration is for Ab. I don't think it is known in AD what is low and what is high. In other studies of Ab, low concentrations were picomolar (Puzzo et al., listed in the references). So it is not clear the term low is justified for 50 nM.

   The bursts of activity are not quantified. What was defined as a burst? What was the burst frequency and did it change over the recording period?

   In the section about mPSCs in culture, starting on Line 348, were these events EPSCs or IPSCs? It is important because in the section starting on Line 383 there were changes in IPSCs but the authors conclude a major role of EPSCs only. For example, Line 400 suggests that the effects of Ab were on AMPA receptor-mediated activity but it seems from the data there were also some effects on IPSCs.

   Line 434. Provide evidence that the fluorescent probe accurately measures NO.

   At the top of page 19 there is a section that needs to be moved earlier because it relates to the work in cultures. That earlier section needs to be reinterpreted given changes in membrane properties occurred. Also, if there is increased synaptic activity in cells dialyzed with Ab, TTX needs to be added to be sure of intrinsic properties. The increase in excitability the authors discuss could be due to the synaptic activity or changes in properties, or both and this needs clarification.

   The last paragraph on page 20 is not useful because DRG neurons are so different from hippocampal neurons. One could have effects in DRG but not hippocampus, and vice-versa. The paragraph starting on Line 616 should be revised. It is not a series of compelling arguments in its present form. For example, saying that AMPAR are linked to epilepsy seems quite obvious, and does not mean that the work presented here is like epilepsy because AMPAR events increased in several assays. Increased AMPAR events also occur when there is a change in behavioral state, plasticity, etc.

   In the conclusions, I don't think the data suggest a synaptic change in AMPAR alone. There are intrinsic changes and changes in GABAergic events. Many sites in the brain could have different effects but were not studied. It is not clear effects of NO were coordinated in the way they affected adjacent neurons to the recorded cell. NO simply could have diffused to an area around the recorded cell. I may have missed evidence to the contrary, but effects could have been mediated by axons of the recorded cell and not NO.

   In Figure 1b, there is a representative example. Could the neurons be shown? Then one knows the relationship of the signal to the location of neurons.

   Graphs should show points. This is one way to clarify sample size easily also.

   MINOR POINTS

   Line 169 mentions stable access resistance and one usually provides a number indicating how little it increased over time, such as 10-20%. Similarly the way synaptic events were discriminated by noise is not provided (line 291). Instead, a brief description is provided.

   Line 292 mentions noise ~2 pA but it is much higher in the data shown in the figures.

   Solvents of drugs are not listed at all, and controls that show no effect of vehicle need clarification in some cases.

   On Line 371, Ab-mediated neurotransmission is used. I believe this needs to be modulated rather than mediated, or an explanation is needed.

   On Line 381, how do the authors know that EPSCs are mediated primarily by AMPA receptors in this preparation?

   On Line 393, what is the comparison of AMPA-mediated events to [where it is stated they are what is mostly changing]?

   In all of the sections where drugs were applied, abbreviations need to be spelled out before the first use, concentrations need to be confirmed as specifically action on the intended receptor, and indirect effects on other cells need to be discussed if bath-applied.

   The sentence starting on Line 417 is a repetition of a prior sentence on the previous page.

   Line 433. Clarify what low concentrations mean here.

   Line 444. mPSCs are referred to here. One needs to know what were the values for E and IPSCs.

   In this section it is often stated that there is a decrease but actually the dialyzed cells are compared to controls so different language is needed.

   Line 461. It is not clear that the hippocampus is the first site to be affected in AD. The entorhinal cortex is earlier in the studies of some, and in the mouse models it is usually the cortex that gets plaque first. In humans, the locus coeruleus may be earlier than the entorhinal cortex.

   How the plots of current vs. spikes were done is important. If there were differences in membrane potential, that could affect the spike output. If there were differences in input resistance or threshold, that also could play a role. One can control for these potential confounds, so explanations are needed.

   Line 472. Vm does not generate fluctuations in this case. Vm changes, and synaptic potentials get larger or smaller, add new components or lose them, etc.

   Line 476. It is not clear why cells are firing at membrane potentials so hyperpolarized to threshold.

   The streptavidin/calbindin labeling is good but the morphology of the cell is not like a pyramidal cell of area CA1 because there is a major branch of the dendrites at almost a right angle to the apical dendrites. The electrophysiology of this cell might be like an interneuron, and two of the figures show firing with a large afterhyperpolarization similar to an interneuron.

   In Figure 3, what are EPSCs and what are spikes would be helpful to point out. The concentration, 500 nm, may never be reached in the brain of an individual with AD, or do the authors have evidence that concentration is relevant in vivo?

   There are typos in figure headings, such as Contro instead of Control and in figure 4g, AMPAergic has the c below AMPAergi

   Read the original source
4. 

   eLife

   Feb 12, 2021

   Summary: This study provides new information about how amyloid beta (Ab) oligomers (Abo) may contribute to hyperexcitability which is important because Abo and hyperexcitability have been suggested to occur early in the development of Alzheimer's disease (AD). The authors added Abo intracellularly (iAbo) using dialysis from a patch pipette. Their data suggest iAbo led to increased synaptic excitation mediated presynaptically by retrograde signalling of nitric oxide (NO). Furthermore, they present data suggesting that there is spread of this increase in excitation to neighboring neurons.

   Major Comments:

   1. The nature of the described effects of intracellular iAbo are quite unexpected, occurring within a minute of obtaining intracellular recording configuration, which contrasts with at least on previous study. While some controls for …

   Summary: This study provides new information about how amyloid beta (Ab) oligomers (Abo) may contribute to hyperexcitability which is important because Abo and hyperexcitability have been suggested to occur early in the development of Alzheimer's disease (AD). The authors added Abo intracellularly (iAbo) using dialysis from a patch pipette. Their data suggest iAbo led to increased synaptic excitation mediated presynaptically by retrograde signalling of nitric oxide (NO). Furthermore, they present data suggesting that there is spread of this increase in excitation to neighboring neurons.

   Major Comments:

   1. The nature of the described effects of intracellular iAbo are quite unexpected, occurring within a minute of obtaining intracellular recording configuration, which contrasts with at least on previous study. While some controls for intracellular application of oligomers are provided, with reverse iAbo failing to reproduce the effect (Fig 2S1) and the effect being blocked by the antibody A11 (Fig 2S2), further controls are necessary to explain this rapid effect, which seems faster than that for the diffusion of the fluorescent tag into the cell (Fig 1S1). Note that Pusch and Neher (Pflug Arch 1988) determined diffusion time for different substances. That paper or others should be cited, and then some estimation of equilibrium time based on diffusibility of ab oligomers should be provided. Equations 17 and 18 in that paper provide some estimates based on molecular weight or diffusion coefficient. One point in Pusch and Neher is there is extreme variability between access times across cells and that it depends on access resistance, of course. Finally, the Pusch and Neher calculations were for small spherical cells - diffusion into spatially extended cells with long dendrites where the synapses are will take even longer. This is especially critical, as one of the major papers of precedent for this work is that of Ripoli, et al. 2014 (cited in the manuscript) in which the authors of that work examined effects of patch applied Ab42 over the course of 20 minutes, with internal controls showing differences between initial responses, right after break in, and 20 minutes later when the oligomer and/or monomers will have had a chance to equilibrate with the intracellular contents. It is not clear how such a rapid effect as indicated in the figures could be achieved by such a large molecule as Ab. The data suggest a time to effect of seconds to minutes, and the peak effect occurs before the fluorescence peaks, which seems hard to explain.

   2. The data need reorganization in terms of their results using h-iAbo or iAbo. There needs to be a clear demonstration of why both were used if the results are generalized with both (or not) and if they can actually use both interchangeably.

   3. The authors need to clearly indicate whether the experiments were done in culture or in slices. The authors need to provide a rationale on why specific experiments were done in culture and others in slices.

   4. There are aspects of the observed phenomenon that have not been taken adequately into account. For example, the authors have not investigated the effects of application of oligomeric beta-amyloid to either the extracellular space or the presynaptic neurons, two other compartments of the synapse.

   5. Aspects of the data raise questions: 1) Western blots appear to have multiple bands 2) evidence that the fluorescent probe accurately measures NO. 3) The bursts of activity are not quantified. What was defined as a burst? What was the burst frequency and did it change over the recording period? 4)The external solution for cultures contain 5.4 mM K+ which is quite high, and can induce hyperexcitability. Similarly, the use of 100uM AMPA and GABA seem very high. Justifying these high concentrations is important. They should lead to hyperexcitability and toxicity (AMPA) over time. Another point of concern is that the concentration of K+ for the slice work is 3 mM, much different than cultures. There are also differences in Mg2+ and Ca2+, making data hard to compare in the two preparations. 5) sample sizes are unclear 6) Intracellular Ab produces increases in both EPSCs and IPSCs. However, in Fig 3, the IPSC measures using a charge transfer quantification, did not show a significant change in response to iAbo, in contrast to EPSCs. 7) With regard to the inhibition, In the schematic on Fig. 10, I find this incomplete and slightly inaccurate since it shows one terminal releasing both glutamate and GABA with NO increasing both. While this is obviously an oversimplification, it's slightly inaccurate since NO was not directly shown to increase sIPSCs. Were NOS blockers able to disrupt the increase in sIPSCs? Moreover, there are many papers that have shown that PKC can also phosphorylate GABA receptors and increase their conductance. What could be the reason that this was not involved here? This needs to be discussed.

   6. How this work relates to other studies is necessary. For example, how this study is related to others about Ab exposure is lacking. Also, regarding hyperexcitability, many possible causes exist. These should be summarized in the introduction and the authors should comment how their results fit with these studies. Regarding PKC and NO, PKC and NO have several known actions throughout the brain and body. How do the effects the authors have identified relate to all these other effects? For example, if PKC is activated by another mechanism, would it occlude effects of Ab? What are the changes in PKC and NO in AD? Regarding the ability of the data to address AD, a major issue is whether the results are relevant to AD or represent interesting pharmacological data about what Ab can potentially do in some of its forms in normal tissue.

   Reviewer #2 opted to reveal their name to the authors in the decision letter after review.

   Read the original source
5. 

   Version published to 10.1101/2020.10.16.332445v1 on bioRxiv

   Oct 17, 2020