Comment on the article “A Super-Resolved View of the Alzheimer’s Disease-Related Amyloidogenic Pathway in Hippocampal Neurons”

9 September 2021

In the discussion of Yu et al. [1], the authors indicate one of the main messages as follows:
“In this context we would like to point out that we used STED microscopy, which enables a resolution of around 20–30 nm in the x,y-plane and around 100 nm in the z-axis. Due to the lower resolution along the z-axis, it is important to avoid overlap along the x-axis. A previous study using STORM microscopy suggested that AβPP is present in both the pre- and post-synapse [50]. However, this is probably an artefact due the fact that the axons and dendrites were positioned on top of each other and the resolution along the z-axis is limited even with STORM microscopy. Another study, using SIM microscopy and pre-and postsynaptic markers separated in the x-y plane showed presynaptic localization [51], in line with our findings.”

Our concerns regarding the central idea on the selective localization of AβPP are obtained from Fig. 4 in the manuscript [1] are listed below:

  • The most glaring issue is in Fig. 4, which connects to the central theme of the discussion arguing for the absence of full-length AβPP and CTF. AβPP-CT staining in panels A and B would indicate the difference in the use of non-uniform scaling where the dynamic ranges of AβPP-CT imaging is quite different in both panels. This is very much evident since the frequency of occurrence of AβPP-CT when colocalized with synaptophysin is many-fold more than PSD95 in the same neuronal processes. The images in the panels indicate that there was either an issue with staining or an issue with segmentation and choosing of synapses since the density of detected AβPP-CT staining is very different from the upper to lower panel. This is very important since this qualitative analysis forms the crux of the message.
  • On analyzing the colocalization of AβPP-CT, it could be observed that a lot of independent AβPP-CT staining is aligned laterally opposite to the synaptophysin tag further confirming the presence of AβPP in postsynaptic compartments. This strengthens the observation that is described in previous comments, which deals with selective dynamic scaling of the data to represent the authors hypothesis than what is present in the data.
  • Standardization was not performed. Leica 3D STED can be operated in 2D mode with high lateral resolution and 3D mode with improved axial and lateral resolution. In the 2D mode, the expected experimental resolution is 50-60 nm at best, and in the 3D mode, X, Y, and Z resolution changes to a near isotropic PSF of around 90-100 nm in 3D. Resolution should be confirmed by using a marker for Point Spread Function (a fluorescent object less than 30 nm) when they compare it with techniques of comparable resolution.
  • The paper should inform the readers about segmentation protocols or how/from where these spines were selected/annotated without a bias.
  • A simple follow-through on Kedia at al. [2] shows that they have used several super resolution paradigms and exhaustive quantification to rule out this artefact occurring. They have relied on multiple super resolution imaging paradigms like PALM, STORM, and UPAINT along with both 2D and stereographic analysis of 3D STED in the cited paper [2]. Additionally, Kedia et al. [2] observed both endogenous and ectopically expressed molecules to interpret data in the same manuscript. The PALM studies show that ectopically expressed AβPP molecules are laterally diffused into morphologically characterized spines (Fig. 2 and Figure S4 of [2]). It was also supplemented by UPAINT data where they label only the AβPP present on the surface (Figure S5). Secondly, Kedia et al. [2] have used both 2D STED and 3D STED (Figure S9) to show that AβPP counter labelled by C Terminal antibody is present in postsynaptic density and other compartments of the excitatory synapse. In line with observation of STORM data in Kedia et al. [2], other paradigms confirm the presence of AβPP in the spine and functional zones of the synapse. Lastly semiautomated sampling and thresholding from 1000s of synapses Kedia et al. [2] used is robust and unbiased.
  • The argument from the Rice et al. [3] paper is that there is less AβPP in post- compared to presynapse (Fig. S1 of [3]). Rice et al. [3] indicate differential expression of AβPP in pre-and postsynapse but not absence of the same.
  • Presence of AβPP in postsynaptic compartment is in consensus with several biochemical, electrophysiological, and microscopic observations putting full-length AβPP in both pre- and postsynaptic compartments [4-17]. Of note, the seminal works put full length AβPP in dendritic microdomains, endosomes and excitatory synapses, presynaptic compartment, and dendritic spines [18-21]. The most rigorous set of studies by Helm et al. and Wilhelm et al., to date, confirm the presence of full-length AβPP in both mushroom and stubby spines as well as the synaptic membrane of presynaptic compartment [20, 21]. Helm et al. [21] and Wilhelm et al. [20] used a combination of ensemble based super resolution microscopy (STED), quantitative biochemical methods, electron microscopy, and molecular modelling to extract copy numbers of AβPP in neuronal cells and different kinds of spines. The work from Helm et al. [21] follows up their seminal work in quantifying AβPP numbers in the presynaptic compartment where they showed full length AβPP on the surface of the presynaptic membrane. Additionally, several layers of additional evidence that cannot be summarized in this letter puts full length AβPP in pre/post synapses, and we should also consider observations that highlight interaction between pre/post synaptic AβPP in synapses, surface staining assays, and how expression of detrimental forms of AβPP affected AMPA receptor recruitment and diffusion, linking their post synaptic localization and function.

The evidence we list here, along with the issues observed in Fig. 4, point out that the conclusion that “neither full-length AβPP nor APP-CTF was present at the post synapse” cannot be confirmed with the paradigm Yu et al. [1] used in the paper nor does it confirm the observations in Kedia at al. [2] is an artefact nor their statement about Rice et al. [3] confirming the absence of AβPP or AβPP-CTF in post synapse.

Deepak Nair
Centre for Neuroscience
Indian Institute of Science
deepak@iisc.ac.in

REFERENCES
[1] Yu Y, Gao Y, Winblad B, Tjernberg L, Schedin Weiss S (2021) A super-resolved view of the Alzheimer's disease-related amyloidogenic pathway in hippocampal neurons. J Alzheimers Dis, doi: 10.3233/JAD-215008.
[2] Kedia S, Ramakrishna P, Netrakanti PR, Jose M, Sibarita JB, Nadkarni S, Nair D (2020) Real-time nanoscale organization of amyloid precursor protein. Nanoscale 12, 8200-8215.
[3] Rice HC, de Malmazet D, Schreurs A, Frere S, Van Molle I, Volkov AN, Creemers E, Vertkin I, Nys J, Ranaivoson FM, Comoletti D, Savas JN, Remaut H, Balschun D, Wierda KD, Slutsky I, Farrow K, De Strooper B, de Wit J (2019) Secreted amyloid-beta precursor protein functions as a GABABR1a ligand to modulate synaptic transmission. Science 363, eaao4827.
[4] Montagna E, Dorostkar MM, Herms J (2017) The role of APP in structural spine plasticity. Front Mol Neurosci 10, 136.
[5] Muller UC, Deller T, Korte M (2017) Not just amyloid: physiological functions of the amyloid precursor protein family. Nat Rev Neurosci 18, 281-298.
[6] Klevanski M, Herrmann U, Weyer SW, Fol R, Cartier N, Wolfer DP, Caldwell JH, Korte M, Muller UC (2015) The APP intracellular domain is required for normal synaptic morphology, synaptic plasticity, and hippocampus-dependent behavior. J Neurosci 35, 16018-16033.
[7] Soba P, Eggert S, Wagner K, Zentgraf H, Siehl K, Kreger S, Lower A, Langer A, Merdes G, Paro R, Masters CL, Muller U, Kins S, Beyreuther K (2005) Homo- and heterodimerization of APP family members promotes intercellular adhesion. EMBO J 24, 3624-3634.
[8] Rodrigues DI, Gutierres J, Pliassova A, Oliveira CR, Cunha RA, Agostinho P (2014) Synaptic and sub-synaptic localization of amyloid-beta protein precursor in the rat hippocampus. J Alzheimers Dis 40, 981-992.
[9] Pliassova A, Lopes JP, Lemos C, Oliveira CR, Cunha RA, Agostinho P (2016) The association of amyloid-beta protein precursor with alpha- and beta-secretases in mouse cerebral cortex synapses is altered in early Alzheimer's disease. Mol Neurobiol 53, 5710-5721.
[10] Hoe HS, Fu Z, Makarova A, Lee JY, Lu C, Feng L, Pajoohesh-Ganji A, Matsuoka Y, Hyman BT, Ehlers MD, Vicini S, Pak DT, Rebeck GW (2009) The effects of amyloid precursor protein on postsynaptic composition and activity. J Biol Chem 284, 8495-8506.
[11] DeBoer SR, Dolios G, Wang R, Sisodia SS (2014) Differential release of beta-amyloid from dendrite- versus axon-targeted APP. J Neurosci 34, 12313-12327.
[12] Wei W, Nguyen LN, Kessels HW, Hagiwara H, Sisodia S, Malinow R (2010) Amyloid beta from axons and dendrites reduces local spine number and plasticity. Nat Neurosci 13, 190-196.
[13] Wang Z, Wang B, Yang L, Guo Q, Aithmitti N, Songyang Z, Zheng H (2009) Presynaptic and postsynaptic interaction of the amyloid precursor protein promotes peripheral and central synaptogenesis. J Neurosci 29, 10788-10801.
[14] Shigematsu K, McGeer PL, McGeer EG (1992) Localization of amyloid precursor protein in selective postsynaptic densities of rat cortical neurons. Brain Res 592, 353-357.
[15] Tampellini D, Rahman N, Gallo EF, Huang Z, Dumont M, Capetillo-Zarate E, Ma T, Zheng R, Lu B, Nanus DM, Lin MT, Gouras GK (2009) Synaptic activity reduces intraneuronal Abeta, promotes APP transport to synapses, and protects against Abeta-related synaptic alterations. J Neurosci 29, 9704-9713.
[16] Tyan SH, Shih AY, Walsh JJ, Maruyama H, Sarsoza F, Ku L, Eggert S, Hof PR, Koo EH, Dickstein DL (2012) Amyloid precursor protein (APP) regulates synaptic structure and function. Mol Cell Neurosci 51, 43-52.
[17] Weingarten J, Weingarten M, Wegner M, Volknandt W (2017) APP-A novel player within the presynaptic active zone Proteome. Front Mol Neurosci 10, 43.
[18] Das U, Scott DA, Ganguly A, Koo EH, Tang Y, Roy S (2013) Activity-induced convergence of APP and BACE-1 in acidic microdomains via an endocytosis-dependent pathway. Neuron 79, 447-460.
[19] Das U, Wang L, Ganguly A, Saikia JM, Wagner SL, Koo EH, Roy S (2016) Visualizing APP and BACE-1 approximation in neurons yields insight into the amyloidogenic pathway. Nat Neurosci 19, 55-64.
[20] Wilhelm BG, Mandad S, Truckenbrodt S, Krohnert K, Schafer C, Rammner B, Koo SJ, Classen GA, Krauss M, Haucke V, Urlaub H, Rizzoli SO (2014) Composition of isolated synaptic boutons reveals the amounts of vesicle trafficking proteins. Science 344, 1023-1028.
[21] Helm MS, Dankovich TM, Mandad S, Rammner B, Jahne S, Salimi V, Koerbs C, Leibrandt R, Urlaub H, Schikorski T, Rizzoli SO (2021) A large-scale nanoscopy and biochemistry analysis of postsynaptic dendritic spines. Nat Neurosci 24, 1151-1162.

Comments

Dr. Deepak Kumaran Nair, we are proud of our study, which has been carefully and thoroughly conducted and we stand firmly behind all the data presented. Despite your aggressive tone and several incorrect statements, we thank the Editors for giving us the opportunity to clarify the points raised. In fact, we believe that these points are very important to highlight.

Regarding the performance of the STED microscope used, we slightly modified the 6th paragraph in the discussion and specified the calibrated resolution upon correction of the proofs. This, however, does not change our conclusions, if anything, it rather strengthens them.

One important point emphasized in this study [1], as well as in two of our previous studies [2, 3], is the need to take into consideration the biological 3D direction of the synapse in order to separate the pre- and postsynaptic sides, even with the super-resolution microscopy techniques used here. We have spent several years imaging neurons with synaptic markers and actin staining by super-resolution microscopy and know how difficult it can be to find synapses in the preparations where the pre- and postsynapse can be visualized clearly separated, without overlap along the z-axis. In the primary hippocampal neuron cultures, axons are intimately surrounding the dendrites in an intricate manner to form synaptic contacts, where the synaptic cleft is only around 25 nm wide. The synapses in these cultures are present in various directions. Thus, in order to determine the pre- versus postsynaptic side, we select synapses for which the pre- and postsynapse are separated in the x,y plane and use 2D STED, the mode for which we obtain the best possible resolution in the x,y plane (which is not possible to obtain along the z-axis even with 3D STED) to image them. Notably, using super-resolution microscopy techniques with even higher resolution, such as MINFLUX and iPALM, we will be able to overcome such issues in future approaches.

Importantly, as specified in our article, synapses were selected in an unbiased manner:
Actin staining was used to visualize the morphology of the neurons, enabling selection of synapses where the pre- and postsynaptic sides are located side-by-side with the spine extended in the xy plane (to avoid the risk of false positive overlap along the z-axis). Healthy neurons were selected from the actin staining in confocal mode”.

On top of that, we would like to emphasize that the synapses were selected with the AβPP channel off in order to perform the selection in an unbiased manner with respect to the AβPP staining.

We believe that, in order to determine the synaptic side, our approach is more exact than approaches using automatic detection of thousands of synapses where the resolution is not sufficient or the direction of the synapse is not known. This is verified by the clear-cut data we obtained from 60 synapses in Fig 4D. We could have calculated a higher number of synapses if we deemed it to be necessary. However, there is such a clear difference between the AβPP C-terminal staining in the pre- versus postsynapse, that the statistical evaluation is very certain with n=60. Please note that these data were from three different experiments (three different batches of mice) stained and imaged on different days.

Regarding Fig. 4, you seem to have missed the fact that the A and B panel were taken from different images from different staining combinations. Scaling and dynamic ranges were carefully set. The reason that there is more AβPP-CT staining in the A-panel is that the level of axons and presynapses is higher in this panel than in the B-panel. In images where we wanted to determine the synaptic localization, we used only two STED channels, since it is increasingly difficult to obtain high resolution when more STED channels are added. The actin staining helps to outline the shape of the dendrites, but the staining is much weaker in the axons [3]. Thus, it is easier to ensure the direction of the synapse in cases where a presynaptic marker and actin staining are combined than when actin staining and postsynaptic marker are combined. In panel 4B, we point at regions where we are sure of the direction since there is not such a high density of axons in this region. Figure 4A and B are two original images out of several showing the same localization. We have imaged many samples and taken many images and can assure that these ones are representative. As you can see in Fig. 4D, we analyzed 60 synapses from each staining combination and the results are very clear.

As we have written in the paper, and as you correctly noted, not all AβPP-CT colocalized with the synaptic vesicle marker synaptophysin in Fig. 4A. Some colocalized with other types of vesicles. We also found AβPP-CT in other parts of neurons (axons, soma, and dendrites) but not in the dendritic spines. It is obvious that AβPP-CT did not colocalize with dendritic spines, since the spines are clearly stained by actin. We believe that most of the AβPP-CT not colocalizing with synaptophysin are still in axons but some could also be in dendrites but not the postsynapse.

We cited the studies by Kedia et al. [5] and Rice et al. [4] and compared them to our study due to the fact that these studies are the ones we are aware of that have used super-resolution microscopy to study the pre- versus postsynaptic localization of AβPP-CT in fixed primary hippocampal neurons. Two of these studies ([1] and [4]) show similar results, but not [5]. Figure S1 in [4] shows a SIM image with the pre- and postsynaptic markers clearly separated. The SIM image clearly shows AβPP overlapping with the pre- but not the postsynaptic marker.

Having said that, we are well aware of the fact that there is controversy with regards to the synaptic localization of AβPP. As summarized in our recent review article, studying the subcellular localization of AβPP and its fragments in neurons is highly challenging for several reasons [6]. The subject is extensive, and it is not possible to refer to all excellent articles available in this field, which are plenty-fold more than the ones cited by Dr. Nair in his Letter to Editor. There may be differences between different model systems used (e.g., different cell types and different regions in the brain). AβPP and its fragments play many roles in neurons, and they are certainly not only present inside the neuron but also released and bind to receptors on the neuronal membrane, reportedly both in the pre- and postsynaptic region, and can be transferred from neuron to neuron and thus affect synaptic functions in many ways. Our study focused on intracellular AβPP with the purpose of elucidating AβPP processing along the amyloidogenic pathway. In summary, our results are valid for the model system we used, i.e., mouse primary hippocampal neurons.

On behalf of all authors,
Sophia Schedin Weiss

REFERENCES

[1] Yu Y, Gao Y, Winblad B, Tjernberg L, Schedin Weiss S (2021) A super-resolved view of the Alzheimer's disease-related amyloidogenic pathway in hippocampal neurons. J Alzheimers Dis, doi: 10.3233/JAD-215008.

[2] Schedin-Weiss S, Caesar I, Winblad B, Blom H, Tjernberg LO (2016) Super-resolution microscopy reveals gamma-secretase at both sides of the neuronal synapse. Acta Neuropathol Commun 4, 29.

[3] Yu Y, Jans DC, Winblad B, Tjernberg LO, Schedin-Weiss S (2018) Neuronal Abeta42 is enriched in small vesicles at the presynaptic side of synapses. Life Sci Alliance 1, e201800028.

[4] Rice HC, de Malmazet D, Schreurs A, Frere S, Van Molle I, Volkov AN, Creemers E, Vertkin I, Nys J, Ranaivoson FM, Comoletti D, Savas JN, Remaut H, Balschun D, Wierda KD, Slutsky I, Farrow K, De Strooper B, de Wit J (2019) Secreted amyloid-beta precursor protein functions as a GABABR1a ligand to modulate synaptic transmission. Science 363, eaao4827.

[5] Kedia S, Ramakrishna P, Netrakanti PR, Jose M, Sibarita JB, Nadkarni S, Nair D (2020) Real-time nanoscale organization of amyloid precursor protein. Nanoscale 12, 8200-8215.

[6] Lin T, Tjernberg LO, Schedin-Weiss S (2021) Neuronal trafficking of the amyloid precursor protein-What do we really know? Biomedicines 9, 801.