Neuronal circuits in the brain are subject to synaptic plasticity mechanisms induced by sensory experience (Ko et al., 2013) and learning (Chen et al., 2015; Peters et al., 2017) while exhibiting a critical ability to maintain network activity within a normal operating range during perturbations such as sensory deprivation (Hengen et al., 2016; Turrigiano, 2012). The homeostatic regulation of neuronal activity has been demonstrated in vivo, where chronic visual deprivation through monocular eyelid suture induces an initial decline in activity of pyramidal neurons in the visual cortex that is eventually restored to baseline (Hengen et al., 2013; Hengen et al., 2016). Other studies using enucleation to deprive the vision show that the recovery of neuronal activity is accompanied by an increase in spine size (Barnes et al., 2015; Keck et al., 2013). However, the molecular mechanisms underlying this homeostatic regulation of neuronal activity in vivo have not been much investigated. Further, whether these homeostatic mechanisms occur homogenously across the individual neuron or are specific to individual dendritic compartments remains elusive.

AMPA receptors (AMPARs) are the principle postsynaptic glutamate receptors mediating fast excitatory synaptic transmission, and regulation of AMPAR trafficking is critical for synaptic plasticity and brain function (Diering and Huganir, 2018; Volk et al., 2015). One of the major forms of homeostatic regulation of neuronal activity involves the modulation of AMPAR expression at synapses and this has been extensively characterized in vitro and ex vivo (Desai et al., 2002; Goel et al., 2006; Goel and Lee, 2007; O'Brien et al., 1998; Turrigiano, 2012; Turrigiano et al., 1998), where chronic visual deprivation induces up-regulation of synaptic AMPARs in the primary visual cortex (V1). However, whether sensory-deprivation-induced homeostatic regulation of AMPAR trafficking occurs in vivo is unknown.

Here, by using longitudinal in vivo two-photon imaging of fluorescently labeled synaptic AMPAR expression, which is a good proxy for postsynaptic strength, we sought to investigate the underlying molecular mechanisms of homeostatic plasticity in vivo and examine how sensory deprivation affects real-time AMPAR dynamics with single-synapse resolution in awake, unanesthetized animals. We observed that AMPAR expression within individual spines in V1 is highly dynamic under normal conditions and that this dynamic is cell type-specific and visual experience-dependent. Visual deprivation by binocular enucleation initially decreased synaptic AMPARs on apical dendrites of V1 layer 2/3 (L2/3) neurons, but the expression then recovered and subsequently underwent further increase with prolonged deprivation. The later increase of AMPARs induced by deprivation is absent on apical dendrites of L5 neurons, indicating that the homeostatic regulation of AMPARs is cell type-specific. Further, within L2/3 neurons, the increases of synaptic AMPARs on basal dendrites following visual deprivation are earlier and larger than on apical dendrites, suggesting a depth-dependent mechanism. Finally, we show that the up-regulation of synaptic AMPARs induced by deprivation is dependent on expression of the AMPAR-binding protein GRIP1. Collectively, our study reveals detailed spatiotemporal dynamics of AMPARs within live animals in response to visual deprivation.

In the present study, we chronically monitored AMPAR expression in individual synapses within live animals with or without binocular enucleation in order to investigate how experience shapes neural circuits in the adult brain. We found that under baseline conditions in mice with normal experience, sGluA1 expression levels in individual spines are very dynamic. Upon visual deprivation, basal dendrites of V1 L2/3 neurons enhance sGluA1 earlier than apical dendrites and deep spines increase more than superficial spines. The changes induced by deprivation are specific to V1 L2/3 neurons but not L5 neurons and the increase in spine sGluA1 is dependent on GRIP1 expression (Figure 7F). To our knowledge, this work is the first to longitudinally examine synapse strength in different layers of cortical neurons in unanesthetized awake mice. Our in vivo imaging data with single-synapse resolution provide unprecedented levels of spatiotemporal information regarding synaptic AMPAR dynamics and reveal that neurons exhibit a tremendous heterogeneity of synaptic changes under both normal and sensory deprivation conditions.

AMPAR trafficking is critical for synaptic plasticity and brain function (Anggono and Huganir, 2012; Huganir and Nicoll, 2013; Volk et al., 2015). We show that synaptic expression of AMPARs is highly dynamic in mice with normal experience despite the relatively stable overall expression, suggesting that strength of individual synapses is continuously being modified. Intriguingly, the degree of variation of spine sGluA1 expression in L2/3 neurons is larger than that in L5 neurons, indicating that spines in L2/3 neurons are more dynamic than those in L5 neurons. Since L2/3 neurons receive more feedforward inputs from L4 than L5 neurons, the higher dynamic observed in L2/3 neurons suggests that sensory input is a drive for synaptic AMPAR dynamic. Moreover, depriving the visual input greatly changes the dynamics of V1, shifting from L2/3 to L5, which further supports that this dynamic is driven by sensory inputs.

Synaptic strength and spine size are well correlated, and synaptic potentiation or depression is usually accompanied with an increase or decrease in spine size, respectively (Bosch et al., 2014; Makino and Malinow, 2009). In agreement with previous studies, overall we observed a strong correlation between spine sGluA1 level and spine size in live animals. Further, we found that the change of spine sGluA1 and change in spine size change were also highly correlated. Nevertheless, we do find a small population of spines that exhibit a divergence of spine form and function. Moreover, synaptic sGluA1 shows greater changes than spine size following visual deprivation. These suggest that differences in synaptic strength may be underestimated or even not correctly determined when using spine size as a measurement. Indeed, the dissociation of spine size and synaptic strength has been reported many times (Lee et al., 2012). Spine number or volume is not changed at all at cerebellar Purkinje cell synapses during LTD (Sdrulla and Linden, 2007). Insulin-induced endocytosis of AMPARs is not accompanied by spine shrinkage (Wang et al., 2007). Spine size and AMPAR function are coupled through some common signaling mechanisms, but there is a divergence in the downstream signaling pathways that regulate these two processes. For instance, it has been shown that spine shrinkage is mediated by cofilin while LTD is dependent on protein phosphatase one although both require calcium influx through NMDA receptors and enhanced calcineurin activity (Zhou et al., 2004). Thus, spine size, in certain conditions, is not a good indication of synaptic strength, and our imaging of synaptic AMPAR expression provides a direct and accurate way to monitor functional changes at synapses.

Neurons receive and process information from thousands of inputs at synapses that are distributed throughout the extensively branching dendrites. To overcome the filtering and attenuation caused by the cable properties of dendrites (Rall, 1962a; Rall, 1962b), synapses express a varying number of AMPARs that increases with distance from the soma, a phenomenon known as distance-dependent scaling (Andrasfalvy and Magee, 2001). While this phenomenon has been extensively studied in hippocampal CA1 pyramidal neurons (Menon et al., 2013; Nicholson et al., 2006; Shipman et al., 2013), it is unknown whether this scaling occurs in cortical pyramidal neurons. Here, we find that both L2/3 and L5 cortical pyramidal neurons display a distance-dependent baseline expression of sGluA1, wherein distal spines have higher levels of sGluA1 than proximal spines. These results suggest that the distance-dependent scaling of AMPARs might be a general phenomenon across brain regions. The molecular mechanisms of this phenomenon and whether the underlying mechanisms are same or different between hippocampal and cortical pyramidal neurons require further investigations.

Many studies using ex vivo acute slices have shown that chronic visual deprivation leads to synaptic enrichment of AMPARs in V1 L2/3 neurons through synaptic scaling mechanisms (Goel and Lee, 2007; He et al., 2012). These reports used whole-cell recordings to determine AMPAR-mediated miniature excitatory postsynaptic current (mEPSC) amplitude or biochemistry methods to measure synaptic AMPAR expression (Goel et al., 2006; Goel and Lee, 2007), which both reflect average AMPAR level of all synapses from a cell or a large number of cells in V1, and thus lacks crucial spatial information describing the dynamic behavior of individual synapses. Using in vivo imaging of fluorescently labeled AMPARs, we were able to track individual spine changes before and after visual deprivation, thus providing unprecedented levels of spatiotemporal information regarding synaptic AMPAR dynamics. We determined that changes in spine sGluA1 expression induced by visual deprivation are highly heterogeneous in live animals and only a subset of spines undergoes potentiation. These potentiated synapses are spatially organized and tend to be located deep within L2/3 neurons, predominately on basal rather than apical dendrites, and this depth-dependent mechanism even applies to spines along the same dendrite, indicating that deprivation-induced changes are compartment- and input-specific, as apical dendrites and basal dendrites of L2/3 neurons receive different inputs (Ko et al., 2011; Makino and Komiyama, 2015; Petrus et al., 2015; Zhang et al., 2014). In addition, it has been shown that the expression pattern of neurotransmitter receptors varies in different layers, and thus other extracellular factors like neuromodulators could also contribute to this depth-dependent plasticity (Brombas et al., 2014; Ji et al., 2015). Although homeostatic plasticity is generally thought to occur globally throughout the neuron, it can also occur locally (Béïque et al., 2011; Turrigiano, 2012). Our results reveal that the visual deprivation-induced homeostatic up-regulation of synaptic AMPARs is synapse-specific.

The increase in spine sGluA1 on the apical dendrites of L2/3 neurons occurred after 7 days of visual deprivation while in V1 region from the acute brain slice we saw a significant elevation of synaptic GluA1 after 2 days of deprivation. That discrepancy led us to investigate the changes happening on the basal dendrites of L2/3 neurons. Indeed, we observed that the amount of synaptic sGluA1 on the basal dendrites of L2/3 neurons was significantly increased after one day of deprivation and remained elevated afterwards. Therefore, the increase observed after 2 days ex vivo is primarily caused by the increase of synaptic sGluA1 on basal dendrites of L2/3 neurons. As far as we know, our study is the first report to show that synaptic sGluA1 on the apical dendrites of L2/3 neurons decreases first following visual deprivation. The dendrites that show the initial decrease are located in the superficial region and they probably receive top-down inputs from regions like retrosplenial cortex and cingulate (Makino and Komiyama, 2015; Roth et al., 2016; Zhang et al., 2014). Visual deprivation may induce LTD that results in reduced synaptic AMPAR level in those connections. Extracellular factors like neuromodulators could also contribute to this (Brombas et al., 2014; Ji et al., 2015).

Visual-deprivation-induced homeostatic plasticity has been reported to be lamina-specific and age-dependent in V1. For instance, L4 neurons have an early critical period from P16 to P21 during which visual loss homeostatically up-regulates excitatory synaptic transmission (Desai et al., 2002). In L2/3, homeostatic plasticity is expressed after P21 and persists into adulthood (Goel and Lee, 2007). In L6, dark exposure initiated before but not after P21 increases average amplitude of mEPSC (Petrus et al., 2011). Regarding L5 neurons, visual deprivation induces an increase in spine size and mEPSC amplitude of L5 neurons in adult mice (Barnes et al., 2017; Keck et al., 2013). However, in the same study they also found that such deprivation did not cause a homeostatic response of L2/3 neurons, which is contrary to other ex vivo studies (Barnes et al., 2015; Goel and Lee, 2007). These discrepancies could result from different deprivation paradigms being used, such as monocular or binocular enucleation, dark exposure, eyelid suture (Whitt et al., 2014). Notably, a study has carefully and systematically investigated how varying visual deprivation paradigms affect plasticity in V1 and demonstrates that a complete loss of visually driven cortical activity is required to elicit homeostatic plasticity in V1 L2/3 pyramidal neurons (He et al., 2012). Here, we used binocular enucleation to completely deprive visual inputs and showed that this paradigm successfully induces up-regulation of synaptic sGluA1 in V1 L2/3 neurons, which is consistent with previous ex vivo studies. However, we did not see any increase of synaptic sGluA1 on apical dendrites of L5 neurons following the same deprivation in adult animals. The increase in mEPSC amplitude of L5 neurons induced by visual deprivation could be contributed by the basal dendrites, as here we only imaged the apical dendrites of L5 neurons. Due to their deep locations, we are not able to image the basal dendrites of L5 neurons with our two-photon system. In addition, it has been shown that in the barrel cortex there is distinct plasticity triggered by sensory changes in specific subtypes of L5 neurons (Greenhill et al., 2015; Holtmaat et al., 2006). In the primary visual cortex, we also observed two populations of L5 cells showing different responses to visual deprivation, with one increasing sGluA1 and the other one not. Future studies are necessary to examine whether these two different responses result from distinct subtypes of the cells. Also, it will be interesting to know whether the homeostatic plasticity in L5 neurons is development-dependent or not.

Emerging evidence has shown that loss of a sensory input leads to widespread changes across brain areas, including the deprived sensory cortex and spared sensory cortices (Ibrahim et al., 2016; Lee and Whitt, 2015; Petrus et al., 2015). For example, visual deprivation not only induces homeostatic plasticity in primary visual cortex, but also produces compensatory changes in other spared sensory cortices, such as somatosensory cortex and auditory cortex, with decreases of mEPSC amplitude in L2/3 pyramidal neurons (Lee and Whitt, 2015; Petrus et al., 2015). However, how visual deprivation alters non-V1 visual cortex remains unknown. In the study, we showed that loss of vision caused a reduction in synaptic sGluA1 of L2/3 neurons within non-V1 visual cortex. Notably, unlike somatosensory and auditory cortices, non-V1 visual cortex exhibited a much earlier decrease. Therefore, the adaptation of brain circuits within visual cortex is different from those in spared sensory cortices.

Overall, by tracking the strength of individual synapses with longitudinal imaging of AMPAR dynamics in living animals during visual deprivation, we reveal that synaptic inputs to distinct cortical layers are differentially modulated in response to sensory experience. Our study supports the notion that the adult brain remains remarkably malleable and is continuously reshaped by experience.

All experimental protocols were conducted according to the National Institutes of Health guidelines for animal research and were approved by the Animal Care and Use Committee at Johns Hopkins University School of Medicine.

Fluorescence intensity analysis

Request a detailed protocol

Signal intensity in spines was analyzed using a custom-written software MapManager (https://mapmanager.net) in Igor Pro as previously described (Zhang et al., 2015). Briefly, each spine was assigned two regions of interest (ROIs) with a spine_ROI enclosed the spine head and a shaft_ROI enclosing the dendritic shaft adjacent to that spine. A background_ROI (same shape and number of pixels as the spine_ROI and shaft_ROI) was translated in x/y to a nearby region of the image that was representative of the background fluorescence. Intensity of SEP-GluA1 represents surface sGluA1 expression as SEP signal is pH-dependent, whereby acidic intracellular environments quenches the fluorescence. To compare intensity values between imaging sessions, the background subtracted spine_ROI from either the SEP-GluA1 or dsRed channel was normalized to background subtracted the dsRed signal on the adjacent dendritic shaft_ROI. Further forms of normalizations were performed for different analyses as described in the following paragraph.

In Figure 1G–I, spine surface SEP-GluA1 (sGluA1) level of each spine was normalized to its individual day one intensity and the geometric mean of spine change per dendrite was calculated.

In Figure 2B,C,H, relative spine sGluA1 level was calculated by normalizing to the average level of the dendrite where the spine was located. In Figure 2F, Figure 2—figure supplement 1A,B, the intensities of newly formed spines were measured when the spines were first detected; the intensities of eliminated spines were measured when the spines were last detected. In Figure 2H, Figure 2—figure supplement 1D, relative spine size (y axis) or sGluA1 (x axis) was calculated by normalizing to the average level of the dendrite where the spine was located. In Figure 2I, Figure 2—figure supplement 1E, spine size (y axis) or sGluA1 (x axis) daily change was calculated by normalization to the level on the previous day.

In Figure 3B,C,E,F; Figure 4D,F; Figure 5F,G; Figure 6F, H; Figure 7E; Figure 3—figure supplement 1A,C; Figure 4—figure supplement 1C; Figure 5—figure supplement 1B, C, K; Figure 6—figure supplement 1B,C,E,F; Figure 7—figure supplement 1A,B, individual spine sGluA1/size levels were normalized to the average of three baselines (BL1-BL3) and the geometric mean of spine change per dendrite was calculated. In Figure 5H, the change ratio on each imaging session was calculated by normalizing changes of spine sGluA1 level on basal dendrites to the change in spine sGluA1 expression on apical dendrites of the same neuron. In Figure 4D,E; Figure 6H, dendrites were defined as decrease dendrites if they had a significant decrease of at least 18% (standard deviation of the percent change in the sham group) in spine sGluA1 at VD1; Otherwise, they were defined as ‘no decrease’ dendrites. In Figure 4—figure supplement 1C, spines were defined as decrease spines if they had a significant decrease of at least 43.5% (standard deviation of the percent change in the sham group) in spine sGluA1 at VD1; Otherwise, they were defined as ‘no decrease’ spines. In Figure 6H, cells were defined as decrease cells if their apical dendrites had a significant decrease of at least 9.8% (standard deviation of the percent change in sham group) in spine sGluA1 at VD1; Otherwise, they were defined as ‘no decrease’ cells.

In Figure 5—figure supplement 1E; Figure 6—figure supplement 1H, the depth of dendrite was calculated by averaging all spines on the dendrite. In each imaging ROI, the relative depth of dendrites was calculated as the Z distance relative to the most superficial dendrite (depth = 0) and the deepest dendrite (depth = 1). In Figure 5—figure supplement 1F–J, apical dendrites and basal dendrites were combined together. In Figure 6—figure supplement 1I,J, apical dendrites were analyzed. As shown in Figure 5—figure supplement 1D, for depth analysis, spine depth from the pia was calculated as the z distance relative to the most superficial spine (depth = 0) and the deepest spine (depth = 1) on the same dendrite. For distance analysis, on each dendrite, spine distance was calculated relative to the most proximal spine to the branch point (defined as 0) and the most distal spine (defined as 1). Spine sGluA1 change was calculated by normalization to the average sGluA1 change of all persistent spines on the same dendrite. The dendrites were defined as ‘ascending’ dendrites if the dendrites were extending towards pia and ΔDepth / ΔDistance was larger than 0.1 (Scheme 1). The dendrites were defined as ‘descending’ dendrites if the dendrites were extending away from pia and ΔDepth / ΔDistance was larger than 0.1 (Scheme 1). For all imaging analysis, the averages were calculated per dendrite.

Scheme 1

Download asset Open asset

All data generated or analyzed (changes of fluorescence over time) during this study are included in the manuscript and supporting files. We also provide one set of raw imaging data (Supplementary files 1 and 2). Due to the large volume of imaging data sets, all raw imaging data are on a local secure server from Huganir lab and will be available upon request. We provide the raw source data generated by our analysis software from the raw images (Supplementary file 3). Signal intensity in spines was analyzed using a custom-written software MapManager (https://mapmanager.net) in Igor Pro. Full documentation and source code download is available at https://github.com/mapmanager.

1. 1. Aerts J
   2. Nys J
   3. Arckens L

   (2014) A highly reproducible and straightforward method to perform in vivo ocular enucleation in the mouse after eye opening

   Journal of Visualized Experiments 6:e51936.

   https://doi.org/10.3791/51936

   • Google Scholar
2. 1. Andrasfalvy BK
   2. Magee JC

   (2001) Distance-dependent increase in AMPA receptor number in the dendrites of adult hippocampal CA1 pyramidal neurons

   The Journal of Neuroscience 21:9151–9159.

   https://doi.org/10.1523/JNEUROSCI.21-23-09151.2001

   • PubMed
   • Google Scholar
3. 1. Anggono V
   2. Huganir RL

   (2012) Regulation of AMPA receptor trafficking and synaptic plasticity

   Current Opinion in Neurobiology 22:461–469.

   https://doi.org/10.1016/j.conb.2011.12.006

   • PubMed
   • Google Scholar
4. 1. Barnes SJ
   2. Sammons RP
   3. Jacobsen RI
   4. Mackie J
   5. Keller GB
   6. Keck T

   (2015) Subnetwork-Specific homeostatic plasticity in mouse visual cortex in vivo

   Neuron 86:1290–1303.

   https://doi.org/10.1016/j.neuron.2015.05.010

   • PubMed
   • Google Scholar
5. 1. Barnes SJ
   2. Franzoni E
   3. Jacobsen RI
   4. Erdelyi F
   5. Szabo G
   6. Clopath C
   7. Keller GB
   8. Keck T

   (2017) Deprivation-Induced homeostatic spine scaling in Vivo Is Localized to Dendritic Branches that Have Undergone Recent Spine Loss

   Neuron 96:871–882.

   https://doi.org/10.1016/j.neuron.2017.09.052

   • PubMed
   • Google Scholar
6. 1. Béïque JC
   2. Na Y
   3. Kuhl D
   4. Worley PF
   5. Huganir RL

   (2011) Arc-dependent synapse-specific homeostatic plasticity

   PNAS 108:816–821.

   https://doi.org/10.1073/pnas.1017914108

   • PubMed
   • Google Scholar
7. 1. Bosch M
   2. Castro J
   3. Saneyoshi T
   4. Matsuno H
   5. Sur M
   6. Hayashi Y

   (2014) Structural and molecular remodeling of dendritic spine substructures during long-term potentiation

   Neuron 82:444–459.

   https://doi.org/10.1016/j.neuron.2014.03.021

   • PubMed
   • Google Scholar
8. 1. Bridi MCD
   2. de Pasquale R
   3. Lantz CL
   4. Gu Y
   5. Borrell A
   6. Choi SY
   7. He K
   8. Tran T
   9. Hong SZ
   10. Dykman A
   11. Lee HK
   12. Quinlan EM
   13. Kirkwood A

   (2018) Two distinct mechanisms for experience-dependent homeostasis

   Nature Neuroscience 21:843–850.

   https://doi.org/10.1038/s41593-018-0150-0

   • PubMed
   • Google Scholar
9. 1. Brombas A
   2. Fletcher LN
   3. Williams SR

   (2014) Activity-dependent modulation of layer 1 inhibitory neocortical circuits by acetylcholine

   Journal of Neuroscience 34:1932–1941.

   https://doi.org/10.1523/JNEUROSCI.4470-13.2014

   • PubMed
   • Google Scholar
10. 1. Chen SX
    2. Kim AN
    3. Peters AJ
    4. Komiyama T

    (2015) Subtype-specific plasticity of inhibitory circuits in motor cortex during motor learning

    Nature Neuroscience 18:1109–1115.

    https://doi.org/10.1038/nn.4049

    • PubMed
    • Google Scholar
11. 1. Desai NS
    2. Cudmore RH
    3. Nelson SB
    4. Turrigiano GG

    (2002) Critical periods for experience-dependent synaptic scaling in visual cortex

    Nature Neuroscience 5:783–789.

    https://doi.org/10.1038/nn878

    • PubMed
    • Google Scholar
12. 1. Diering GH
    2. Huganir RL

    (2018) The AMPA receptor code of synaptic plasticity

    Neuron 100:314–329.

    https://doi.org/10.1016/j.neuron.2018.10.018

    • PubMed
    • Google Scholar
13. 1. Dong H
    2. O'Brien RJ
    3. Fung ET
    4. Lanahan AA
    5. Worley PF
    6. Huganir RL

    (1997) GRIP: a synaptic PDZ domain-containing protein that interacts with AMPA receptors

    Nature 386:279–284.

    https://doi.org/10.1038/386279a0

    • Google Scholar
14. 1. Gainey MA
    2. Tatavarty V
    3. Nahmani M
    4. Lin H
    5. Turrigiano GG

    (2015) Activity-dependent synaptic GRIP1 accumulation drives synaptic scaling up in response to action potential blockade

    PNAS 112:E3590–E3599.

    https://doi.org/10.1073/pnas.1510754112

    • PubMed
    • Google Scholar
15. 1. Garrett ME
    2. Nauhaus I
    3. Marshel JH
    4. Callaway EM

    (2014) Topography and areal organization of mouse visual cortex

    Journal of Neuroscience 34:12587–12600.

    https://doi.org/10.1523/JNEUROSCI.1124-14.2014

    • PubMed
    • Google Scholar
16. 1. Goel A
    2. Jiang B
    3. Xu LW
    4. Song L
    5. Kirkwood A
    6. Lee HK

    (2006) Cross-modal regulation of synaptic AMPA receptors in primary sensory cortices by visual experience

    Nature Neuroscience 9:1001–1003.

    https://doi.org/10.1038/nn1725

    • PubMed
    • Google Scholar
17. 1. Goel A
    2. Lee HK

    (2007) Persistence of experience-induced homeostatic synaptic plasticity through adulthood in superficial layers of mouse visual cortex

    Journal of Neuroscience 27:6692–6700.

    https://doi.org/10.1523/JNEUROSCI.5038-06.2007

    • PubMed
    • Google Scholar
18. 1. Greenhill SD
    2. Ranson A
    3. Fox K

    (2015) Hebbian and Homeostatic Plasticity Mechanisms in Regular Spiking and Intrinsic Bursting Cells of Cortical Layer 5

    Neuron 88:539–552.

    https://doi.org/10.1016/j.neuron.2015.09.025

    • Google Scholar
19. 1. He K
    2. Petrus E
    3. Gammon N
    4. Lee H-K

    (2012) Distinct sensory requirements for unimodal and Cross-Modal homeostatic synaptic plasticity

    Journal of Neuroscience 32:8469–8474.

    https://doi.org/10.1523/JNEUROSCI.1424-12.2012

    • Google Scholar
20. 1. Hengen KB
    2. Lambo ME
    3. Van Hooser SD
    4. Katz DB
    5. Turrigiano GG

    (2013) Firing rate homeostasis in visual cortex of freely behaving rodents

    Neuron 80:335–342.

    https://doi.org/10.1016/j.neuron.2013.08.038

    • PubMed
    • Google Scholar
21. 1. Hengen KB
    2. Torrado Pacheco A
    3. McGregor JN
    4. Van Hooser SD
    5. Turrigiano GG

    (2016) Neuronal firing rate homeostasis is inhibited by sleep and promoted by wake

    Cell 165:180–191.

    https://doi.org/10.1016/j.cell.2016.01.046

    • PubMed
    • Google Scholar
22. 1. Holtmaat AJ
    2. Trachtenberg JT
    3. Wilbrecht L
    4. Shepherd GM
    5. Zhang X
    6. Knott GW
    7. Svoboda K

    (2005) Transient and persistent dendritic spines in the neocortex in vivo

    Neuron 45:279–291.

    https://doi.org/10.1016/j.neuron.2005.01.003

    • PubMed
    • Google Scholar
23. 1. Holtmaat A
    2. Wilbrecht L
    3. Knott GW
    4. Welker E
    5. Svoboda K

    (2006) Experience-dependent and cell-type-specific spine growth in the neocortex

    Nature 441:979–983.

    https://doi.org/10.1038/nature04783

    • Google Scholar
24. 1. Holtmaat A
    2. Svoboda K

    (2009) Experience-dependent structural synaptic plasticity in the mammalian brain

    Nature Reviews Neuroscience 10:647–658.

    https://doi.org/10.1038/nrn2699

    • PubMed
    • Google Scholar
25. 1. Huganir RL
    2. Nicoll RA

    (2013) AMPARs and synaptic plasticity: the last 25 years

    Neuron 80:704–717.

    https://doi.org/10.1016/j.neuron.2013.10.025

    • Google Scholar
26. 1. Ibrahim LA
    2. Mesik L
    3. Ji XY
    4. Fang Q
    5. Li HF
    6. Li YT
    7. Zingg B
    8. Zhang LI
    9. Tao HW

    (2016) Cross-Modality sharpening of visual cortical processing through Layer-1-Mediated inhibition and disinhibition

    Neuron 89:1031–1045.

    https://doi.org/10.1016/j.neuron.2016.01.027

    • PubMed
    • Google Scholar
27. 1. Ji W
    2. Gămănuţ R
    3. Bista P
    4. D'Souza RD
    5. Wang Q
    6. Burkhalter A

    (2015) Modularity in the organization of mouse primary visual cortex

    Neuron 87:632–643.

    https://doi.org/10.1016/j.neuron.2015.07.004

    • PubMed
    • Google Scholar
28. 1. Kalatsky VA
    2. Stryker MP

    (2003) New paradigm for optical imaging: temporally encoded maps of intrinsic signal

    Neuron 38:529–545.

    https://doi.org/10.1016/s0896-6273(03)00286-1

    • PubMed
    • Google Scholar
29. 1. Keck T
    2. Keller GB
    3. Jacobsen RI
    4. Eysel UT
    5. Bonhoeffer T
    6. Hübener M

    (2013) Synaptic scaling and homeostatic plasticity in the mouse visual cortex in vivo

    Neuron 80:327–334.

    https://doi.org/10.1016/j.neuron.2013.08.018

    • PubMed
    • Google Scholar
30. 1. Ko H
    2. Hofer SB
    3. Pichler B
    4. Buchanan KA
    5. Sjöström PJ
    6. Mrsic-Flogel TD

    (2011) Functional specificity of local synaptic connections in neocortical networks

    Nature 473:87–91.

    https://doi.org/10.1038/nature09880

    • PubMed
    • Google Scholar
31. 1. Ko H
    2. Cossell L
    3. Baragli C
    4. Antolik J
    5. Clopath C
    6. Hofer SB
    7. Mrsic-Flogel TD

    (2013) The emergence of functional microcircuits in visual cortex

    Nature 496:96–100.

    https://doi.org/10.1038/nature12015

    • PubMed
    • Google Scholar
32. 1. Lee KF
    2. Soares C
    3. Béïque JC

    (2012) Examining form and function of dendritic spines

    Neural Plasticity 2012:704103.

    https://doi.org/10.1155/2012/704103

    • PubMed
    • Google Scholar
33. 1. Lee HK
    2. Whitt JL

    (2015) Cross-modal synaptic plasticity in adult primary sensory cortices

    Current Opinion in Neurobiology 35:119–126.

    https://doi.org/10.1016/j.conb.2015.08.002

    • PubMed
    • Google Scholar
34. 1. Liao D
    2. Zhang X
    3. O'Brien R
    4. Ehlers MD
    5. Huganir RL

    (1999) Regulation of morphological postsynaptic silent synapses in developing hippocampal neurons

    Nature Neuroscience 2:37–43.

    https://doi.org/10.1038/4540

    • PubMed
    • Google Scholar
35. 1. Majewska A
    2. Sur M

    (2003) Motility of dendritic spines in visual cortex in vivo: changes during the critical period and effects of visual deprivation

    PNAS 100:16024–16029.

    https://doi.org/10.1073/pnas.2636949100

    • PubMed
    • Google Scholar
36. 1. Makino H
    2. Komiyama T

    (2015) Learning enhances the relative impact of top-down processing in the visual cortex

    Nature Neuroscience 18:1116–1122.

    https://doi.org/10.1038/nn.4061

    • PubMed
    • Google Scholar
37. 1. Makino H
    2. Malinow R

    (2009) AMPA receptor incorporation into synapses during LTP: the role of lateral movement and exocytosis

    Neuron 64:381–390.

    https://doi.org/10.1016/j.neuron.2009.08.035

    • PubMed
    • Google Scholar
38. 1. Makino H
    2. Malinow R

    (2011) Compartmentalized versus global synaptic plasticity on dendrites controlled by experience

    Neuron 72:1001–1011.

    https://doi.org/10.1016/j.neuron.2011.09.036

    • PubMed
    • Google Scholar
39. 1. Mao L
    2. Takamiya K
    3. Thomas G
    4. Lin DT
    5. Huganir RL

    (2010) GRIP1 and 2 regulate activity-dependent AMPA receptor recycling via exocyst complex interactions

    PNAS 107:19038–19043.

    https://doi.org/10.1073/pnas.1013494107

    • PubMed
    • Google Scholar
40. 1. Mejias R
    2. Adamczyk A
    3. Anggono V
    4. Niranjan T
    5. Thomas GM
    6. Sharma K
    7. Skinner C
    8. Schwartz CE
    9. Stevenson RE
    10. Fallin MD
    11. Kaufmann W
    12. Pletnikov M
    13. Valle D
    14. Huganir RL
    15. Wang T

    (2011) Gain-of-function glutamate receptor interacting protein 1 variants alter GluA2 recycling and surface distribution in patients with autism

    PNAS 108:4920–4925.

    https://doi.org/10.1073/pnas.1102233108

    • PubMed
    • Google Scholar
41. 1. Menon V
    2. Musial TF
    3. Liu A
    4. Katz Y
    5. Kath WL
    6. Spruston N
    7. Nicholson DA

    (2013) Balanced synaptic impact via distance-dependent synapse distribution and complementary expression of AMPARs and NMDARs in hippocampal dendrites

    Neuron 80:1451–1463.

    https://doi.org/10.1016/j.neuron.2013.09.027

    • PubMed
    • Google Scholar
42. 1. Nauhaus I
    2. Ringach DL

    (2007) Precise alignment of micromachined electrode arrays with V1 functional maps

    Journal of Neurophysiology 97:3781–3789.

    https://doi.org/10.1152/jn.00120.2007

    • PubMed
    • Google Scholar
43. 1. Nicholson DA
    2. Trana R
    3. Katz Y
    4. Kath WL
    5. Spruston N
    6. Geinisman Y

    (2006) Distance-dependent differences in synapse number and AMPA receptor expression in hippocampal CA1 pyramidal neurons

    Neuron 50:431–442.

    https://doi.org/10.1016/j.neuron.2006.03.022

    • PubMed
    • Google Scholar
44. 1. Nusser Z
    2. Lujan R
    3. Laube G
    4. Roberts JD
    5. Molnar E
    6. Somogyi P

    (1998) Cell type and pathway dependence of synaptic AMPA receptor number and variability in the Hippocampus

    Neuron 21:545–559.

    https://doi.org/10.1016/S0896-6273(00)80565-6

    • PubMed
    • Google Scholar
45. 1. O'Brien RJ
    2. Kamboj S
    3. Ehlers MD
    4. Rosen KR
    5. Fischbach GD
    6. Huganir RL

    (1998) Activity-dependent modulation of synaptic AMPA receptor accumulation

    Neuron 21:1067–1078.

    https://doi.org/10.1016/S0896-6273(00)80624-8

    • PubMed
    • Google Scholar
46. 1. Oku Y
    2. Huganir RL

    (2013) AGAP3 and Arf6 regulate trafficking of AMPA receptors and synaptic plasticity

    Journal of Neuroscience 33:12586–12598.

    https://doi.org/10.1523/JNEUROSCI.0341-13.2013

    • PubMed
    • Google Scholar
47. 1. Peters AJ
    2. Liu H
    3. Komiyama T

    (2017) Learning in the rodent motor cortex

    Annual Review of Neuroscience 40:77–97.

    https://doi.org/10.1146/annurev-neuro-072116-031407

    • PubMed
    • Google Scholar
48. 1. Petrus E
    2. Anguh TT
    3. Pho H
    4. Lee A
    5. Gammon N
    6. Lee HK

    (2011) Developmental switch in the polarity of experience-dependent synaptic changes in layer 6 of mouse visual cortex

    Journal of Neurophysiology 106:2499–2505.

    https://doi.org/10.1152/jn.00111.2011

    • PubMed
    • Google Scholar
49. 1. Petrus E
    2. Rodriguez G
    3. Patterson R
    4. Connor B
    5. Kanold PO
    6. Lee HK

    (2015) Vision loss shifts the balance of feedforward and intracortical circuits in opposite directions in mouse primary auditory and visual cortices

    Journal of Neuroscience 35:8790–8801.

    https://doi.org/10.1523/JNEUROSCI.4975-14.2015

    • PubMed
    • Google Scholar
50. 1. Pfennig S
    2. Foss F
    3. Bissen D
    4. Harde E
    5. Treeck JC
    6. Segarra M
    7. Acker-Palmer A

    (2017) GRIP1 binds to ApoER2 and EphrinB2 to induce Activity-Dependent AMPA receptor insertion at the synapse

    Cell Reports 21:84–96.

    https://doi.org/10.1016/j.celrep.2017.09.019

    • PubMed
    • Google Scholar
51. 1. Pologruto TA
    2. Sabatini BL
    3. Svoboda K

    (2003) ScanImage: flexible software for operating laser scanning microscopes

    Biomedical Engineering Online 2:13.

    https://doi.org/10.1186/1475-925X-2-13

    • PubMed
    • Google Scholar
52. 1. Rall W

    (1962a) Electrophysiology of a dendritic neuron model

    Biophysical Journal 2:145–167.

    https://doi.org/10.1016/S0006-3495(62)86953-7

    • PubMed
    • Google Scholar
53. 1. Rall W

    (1962b) Theory of physiological properties of dendrites

    Annals of the New York Academy of Sciences 96:1071–1092.

    https://doi.org/10.1111/j.1749-6632.1962.tb54120.x

    • PubMed
    • Google Scholar
54. 1. Roth MM
    2. Dahmen JC
    3. Muir DR
    4. Imhof F
    5. Martini FJ
    6. Hofer SB

    (2016) Thalamic nuclei convey diverse contextual information to layer 1 of visual cortex

    Nature Neuroscience 19:299–307.

    https://doi.org/10.1038/nn.4197

    • PubMed
    • Google Scholar
55. 1. Saito T
    2. Nakatsuji N

    (2001) Efficient gene transfer into the embryonic mouse brain using in vivo electroporation

    Developmental Biology 240:237–246.

    https://doi.org/10.1006/dbio.2001.0439

    • PubMed
    • Google Scholar
56. 1. Schneider CA
    2. Rasband WS
    3. Eliceiri KW

    (2012) NIH image to ImageJ: 25 years of image analysis

    Nature Methods 9:671–675.

    https://doi.org/10.1038/nmeth.2089

    • PubMed
    • Google Scholar
57. 1. Sdrulla AD
    2. Linden DJ

    (2007) Double dissociation between long-term depression and dendritic spine morphology in cerebellar purkinje cells

    Nature Neuroscience 10:546–548.

    https://doi.org/10.1038/nn1889

    • PubMed
    • Google Scholar
58. 1. Shipman SL
    2. Herring BE
    3. Suh YH
    4. Roche KW
    5. Nicoll RA

    (2013) Distance-dependent scaling of AMPARs is cell-autonomous and GluA2 dependent

    Journal of Neuroscience 33:13312–13319.

    https://doi.org/10.1523/JNEUROSCI.0678-13.2013

    • PubMed
    • Google Scholar
59. 1. Suresh A
    2. Dunaevsky A

    (2017) Relationship between synaptic AMPAR and spine dynamics: impairments in the FXS mouse

    Cerebral Cortex 27:4244–4256.

    https://doi.org/10.1093/cercor/bhx128

    • PubMed
    • Google Scholar
60. 1. Takamiya K
    2. Mao L
    3. Huganir RL
    4. Linden DJ

    (2008) The glutamate receptor-interacting protein family of GluR2-binding proteins is required for long-term synaptic depression expression in cerebellar purkinje cells

    Journal of Neuroscience 28:5752–5755.

    https://doi.org/10.1523/JNEUROSCI.0654-08.2008

    • PubMed
    • Google Scholar
61. 1. Takumi Y
    2. Ramírez-León V
    3. Laake P
    4. Rinvik E
    5. Ottersen OP

    (1999) Different modes of expression of AMPA and NMDA receptors in hippocampal synapses

    Nature Neuroscience 2:618–624.

    https://doi.org/10.1038/10172

    • PubMed
    • Google Scholar
62. 1. Tan HL
    2. Queenan BN
    3. Huganir RL

    (2015) GRIP1 is required for homeostatic regulation of AMPAR trafficking

    PNAS 112:10026–10031.

    https://doi.org/10.1073/pnas.1512786112

    • PubMed
    • Google Scholar
63. 1. Turrigiano GG
    2. Leslie KR
    3. Desai NS
    4. Rutherford LC
    5. Nelson SB

    (1998) Activity-dependent scaling of quantal amplitude in neocortical neurons

    Nature 391:892–896.

    https://doi.org/10.1038/36103

    • PubMed
    • Google Scholar
64. 1. Turrigiano G

    (2012) Homeostatic synaptic plasticity: local and global mechanisms for stabilizing neuronal function

    Cold Spring Harbor Perspectives in Biology 4:a005736.

    https://doi.org/10.1101/cshperspect.a005736

    • PubMed
    • Google Scholar
65. 1. Volk L
    2. Chiu SL
    3. Sharma K
    4. Huganir RL

    (2015) Glutamate synapses in human cognitive disorders

    Annual Review of Neuroscience 38:127–149.

    https://doi.org/10.1146/annurev-neuro-071714-033821

    • PubMed
    • Google Scholar
66. 1. Wang XB
    2. Yang Y
    3. Zhou Q

    (2007) Independent expression of synaptic and morphological plasticity associated with long-term depression

    Journal of Neuroscience 27:12419–12429.

    https://doi.org/10.1523/JNEUROSCI.2015-07.2007

    • PubMed
    • Google Scholar
67. 1. Whitt JL
    2. Petrus E
    3. Lee HK

    (2014) Experience-dependent homeostatic synaptic plasticity in neocortex

    Neuropharmacology 78:45–54.

    https://doi.org/10.1016/j.neuropharm.2013.02.016

    • PubMed
    • Google Scholar
68. 1. Zhang S
    2. Xu M
    3. Kamigaki T
    4. Hoang Do JP
    5. Chang WC
    6. Jenvay S
    7. Miyamichi K
    8. Luo L
    9. Dan Y

    (2014) Selective attention long-range and local circuits for top-down modulation of visual cortex processing

    Science 345:660–665.

    https://doi.org/10.1126/science.1254126

    • PubMed
    • Google Scholar
69. 1. Zhang Y
    2. Cudmore RH
    3. Lin DT
    4. Linden DJ
    5. Huganir RL

    (2015) Visualization of NMDA receptor-dependent AMPA receptor synaptic plasticity in vivo

    Nature Neuroscience 18:402–407.

    https://doi.org/10.1038/nn.3936

    • PubMed
    • Google Scholar
70. 1. Zhou Q
    2. Homma KJ
    3. Poo MM

    (2004) Shrinkage of dendritic spines associated with long-term depression of hippocampal synapses

    Neuron 44:749–757.

    https://doi.org/10.1016/j.neuron.2004.11.011

    • PubMed
    • Google Scholar

Acceptance summary:

Understanding synaptic plasticity is a major goal of developmental neuroscience. Tan et al. have demonstrated techniques for following hundreds of identified dendritic spines in specific neuron classes over the critical period for ocular dominance of mouse visual cortex, with and without visual deprivation by enucleation. The authors demonstrate that spine dynamics depend on the cell type and dendrite type (apical vs. basal), suggesting that the mechanisms of synaptic refinement are not uniform but vary across specific elements and sub-elements of neural circuits.

Decision letter after peer review:

Thank you for submitting your article "Lamina-specific AMPA receptor dynamics following visual deprivation in vivo" for consideration by eLife. Your article has been reviewed by three peer reviewers, one of whom is a member of our Board of Reviewing Editors, and the evaluation has been overseen by Ronald Calabrese as the Senior Editor. The following individual involved in review of your submission has agreed to reveal their identity: Corette J. Wierenga (Reviewer #2).

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

Summary:

The reviewers were united in admiring the quality of the raw data collected and its contribution to our understanding how synaptic spines are influenced by plasticity. At the same time, many concerns were raised about the interpretations and conclusions drawn in the text. The reviewers were united in suggesting a major revision of the paper for consideration for publication in eLife.

Essential revisions:

1) "Hot" and "cold" dendrites: The authors cite the existence of dendrites that exhibit "hot" (most synapses changed) and "cold" (most synapses did not change) dynamics. How do we know that spines with changes are more likely than chance to be located next to each other? There is a certain probability that such a situation could arise from chance, and no effort was made to see if the rate of occurrence of these hot and cold dendrites was greater or less than expected by a random arrangement of spines onto the dendrites (which could be examined by making 1000s of surrogate "location-shuffled" datasets and calculating percentiles). We appreciate that the authors have spent a long time looking at the data and may have some reason to believe that the hot and cold dendrites might be a real phenomenon, but at present there is no evidence in the paper that convince us it is real.

2) Cell-type-specificity vs. dendrite-specificity: The authors claim that layer 5 cells exhibit different plasticity than layer 2/3 neurons. We do understand that the plasticity was different in the observed layer 2/3 and layer 5 spines. However, there seem to be other equally parsimonious explanations. For instance, the following alternative hypothesis: basal dendrites exhibit prolonged homeostatic plasticity while apical dendrites and their branches do not. I imagine that layer 5 spines were only examined in apical dendrites, and so this possibility could not be rejected. Instead of claiming that cell types exhibit different plasticity, a broader set of possibilities should be enumerated in the Abstract, Results, and Discussion. Further, can the authors link their findings to possible VD-induced changes in V1 circuitry? Which synapses are changing and which not? Do L4 inputs target mostly apical or basal dendrites of L2/3 cells? Why would L5 cells not be affected?

3) The observed increase in sGluA1 only occurs after 7 days, while the increase is already observed after 2 days in vitro. Also the initial decrease in sGluA1 observed in vivo has never been reported in vitro (as far as we know). What do the author think is happening at dendrites that show the initial decrease? The authors should discuss this remarkable time course.

4) While GluA1 is correlated with synaptic strength (Figure 1D), the two quantities are not equal; there is considerable variation around the linear correlation. Please rephrase.

5) No account is taken of the well documented differences in homeostatic response and mechanism involved in subtypes of layer 5 neuron.

6) The GRIP KO does show an increase in GluA1 between time VD3 and VD7 points in complete contrast to the conclusion in the text.

7) Unfortunately, the paper does not pay much attention to the quite extensive literature on the mechanism of homeostatic plasticity, nor a good deal of literature conducted on GluA1 knockouts in visual cortex. Neither does it frame the experiments within the appropriate context, because it equates monocular deprivation, dark exposure and eye enucleation experiments within and outside the critical period for development of binocular vision, and thereby cloaks differences between the methods under the misleading term (here) of "sensory deprivation".

Reviewer #2:

In this manuscript the authors have performed longitudinal two-photon imaging to follow changes in spine size and AMPA receptor content in vivo during visual deprivation. They provide evidence that not all synapses respond in a similar manner and interpret this to reflect a depth-dependent plasticity. These new in vivo data are very valuable and the findings are intriguing. I have listed some suggestions to further improve manuscript below.

1) The authors claim that the synaptic changes depend on cortical depth, but I find this a rather vague concept, which is not much discussed. Can the authors speculate if this is linked to a specific input or to a extracellular factor? Can the authors link their findings to possible VD-induced changes in V1 circuitry? Which synapses are changing and which not? Do L4 inputs target mostly apical or basal dendrites of L2/3 cells? Why would L5 cells not be affected?

2) The authors make a strong claim that VD-induced synaptic changes do not occur in L5 (Discussion, sixth paragraph) and that L5 dendrites, in contrast to L2/3 cells, do not show depth-dependent changes after deprivation (e.g. subsection “Lamina-specific increases of spine sGluA1 level after deprivation”, last paragraph). However, they only looked in a small subset of dendrites in L1 and they cannot exclude changes in basal dendrites or dendrites closer to the soma. Furthermore, they actually found a small decrease in sGluA1 in L5 cells after VD (Figure 7E). In addition, the effect in L2/3 as shown in Figure 5—figure supplement 1D is pretty subtle and the number of observations is a lot lower in Figure 6—figure supplement 1I. The most important finding in L2/3, the difference between basal and apical dendrites (or distal/proximal dendrites), was not assessed in L5 cells. I therefore think this claim is too strong and I would ask the authors to either provide additional evidence or to tone this statement down.

Reviewer #3:

This is an interesting paper in many ways, extending the literature on imaging spines size longitudinally, to look at the role played by GluA1 and GRIP in the process of homeostatic response to eye enucleation. Several of the results are of great interest and well founded in the data, for example:

1) The difference in GluA1 dynamics and response to enucleation of apical versus basal dendrites of layer 2/3.

2) The heterogeneous response of (apical?) dendrites of layer 5 neurones.

However, in a number of places, the paper also includes interpretations of the data that are either not supported by the data at all, or comprise such minor components of the highly variable data that they are probably of little biological significance, even if true. There are numerous examples of the two types of issue, but in summary they include:

1) That GluA1 fluorescence does not equate to synaptic strength.

2) Purported difference in fluorescence by depth measurement along ascending or descending dendrites are based on data where depth is a tiny contributor to the large variance.

[Editors' note: further revisions were suggested prior to acceptance, as described below.]

Thank you for resubmitting your article "Lamina-specific AMPA receptor dynamics following visual deprivation in vivo" for consideration by eLife. Your article has been reviewed by two peer reviewers, one of whom is a member of our Board of Reviewing Editors, and the evaluation has been overseen by Ronald Calabrese as the Senior Editor. The following individual involved in review of your submission has agreed to reveal their identity: Corette J. Wierenga (Reviewer #2).

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

Summary and Essential revisions:

The new manuscript is improved, but reviewers were concerned that a few very weak trends were highlighted along with the manuscripts strongest findings. Please remove the part about hot dendrites, which were not upheld in statistical analysis, and move Figure 2J-K, Figure 5I-K to supplement or leave them out completely, and update the text to reflect the changes.

Reviewer #1:

The manuscript by Tan et al. is improved over the previous version. However, there is still one big problem.

The statistical shuffling analysis shows that the likelihood of observing the number of "hot" dendrites that were observed in these studies is about 30% (Figure 2—figure supplement 1A, the authors report 53.21% but adding 30% and 40% for 0 or 1 seems more like 70% total for 0 and 1), while the likelihood of observing 3 cold dendrites was very unlikely (3.82%). There is therefore no significant evidence for "hot" dendrites. The authors ought to remove such a reference. The case is better for cold dendrites; there is some evidence for a few cold dendrites. If it were me, I might remove this bit entirely because the numbers are small, but the authors can decide if they want to leave in the part about cold dendrites.

The other changes were satisfactory.

Reviewer #2:

The authors have addressed most of my comments in a positive manner. The manuscript has clearly improved and I thank the authors for their complete answer to my and the other reviewers' comments. I particularly appreciate the (perhaps somewhat speculative) discussion on how the depth-dependence of synaptic changes after binocular enucleation could reflect changes in specific inputs. This emphasizes the biological relevance of their findings.

I have the feeling that the manuscript would improve by a stronger focus on the strongest and most interesting results. In my opinion, the presentation of small effects and weak correlations of which the biological relevance are not directly clear is diluting the impact of the main results, which actually are of high interest. The authors may consider to move some of the minor findings (Figure 2J-K, Figure 5I-K) to the supplementary information (or leave them out completely).

Essential revisions:

1) "Hot" and "cold" dendrites: The authors cite the existence of dendrites that exhibit "hot" (most synapses changed) and "cold" (most synapses did not change) dynamics. How do we know that spines with changes are more likely than chance to be located next to each other? There is a certain probability that such a situation could arise from chance, and no effort was made to see if the rate of occurrence of these hot and cold dendrites was greater or less than expected by a random arrangement of spines onto the dendrites (which could be examined by making 1000s of surrogate "location-shuffled" datasets and calculating percentiles). We appreciate that the authors have spent a long time looking at the data and may have some reason to believe that the hot and cold dendrites might be a real phenomenon, but at present there is no evidence in the paper that convince us it is real.

We have now examined the probability whether the observed “hot” or “cold” dendrites arise from chance or not by randomly shuffling the position of our spines 10,000 times. We found that both the number of “cold” dendrites and the sum of “hot” and” cold” dendrites (representing all dynamic dendrites) observed here were significantly greater than would occur by chance. The number of “hot” dendrites was also higher than the mean number of the shuffled data although it was not significant. That is probably due to the small number of dendrites we imaged. Nevertheless, these data clearly demonstrate that there are spatially clustered dendritic subregions where AMPARs are more readily down- or up- regulated. We have now added this comparison with shuffled data as Figure 2—figure supplement 1A.

2) Cell-type-specificity vs. dendrite-specificity: The authors claim that layer 5 cells exhibit different plasticity than layer 2/3 neurons. We do understand that the plasticity was different in the observed layer 2/3 and layer 5 spines. However, there seem to be other equally parsimonious explanations. For instance, the following alternative hypothesis: basal dendrites exhibit prolonged homeostatic plasticity while apical dendrites and their branches do not. I imagine that layer 5 spines were only examined in apical dendrites, and so this possibility could not be rejected. Instead of claiming that cell types exhibit different plasticity, a broader set of possibilities should be enumerated in the Abstract, Results, and Discussion. Further, can the authors link their findings to possible VD-induced changes in V1 circuitry? Which synapses are changing and which not? Do L4 inputs target mostly apical or basal dendrites of L2/3 cells? Why would L5 cells not be affected?

Due to their deep locations, we are not able to image the basal dendrites of L5 neurons with our two-photon system. We agree with the reviewer that we cannot exclude the possibility that the basal dendrites of layer 5 neurons could increase synaptic AMPARs following visual deprivation. We have changed our Abstract and Results, claiming that visual deprivation specifically increases synaptic AMPARs on apical dendrites of L2/3 neurons but not on apical dendrites of L5 neurons. In the Discussion, we also discuss the possibility that basal dendrites of L5 neurons could show synaptic enrichment of AMPARs after visual deprivation.

Our data show that the changes in synaptic sGluA1 induced by visual deprivation in V1 L2/3 neurons are significantly correlated with the depth (distance from the pia) of the dendrites or spines, wherein deep dendrites or spines are potentiated more than superficial ones (Figure 5; Figure 5—figure supplement 1). There are many possible mechanisms. First, as the reviewer mentioned, it could be input-specific, as apical dendrites and basal dendrites of L2/3 neurons receive different inputs. The basal dendrites primarily receive feedforward inputs from L4 and nearby L2/3 neurons (Ko et al., 2011; Lee et al., 2016; Park et al., 2019), while the apical dendrites receive feedback inputs from regions like retrosplenial cortex, cingulate, and thalamus (Makino and Komiyama, 2015; Roth et al., 2016; Zhang et al., 2014). In addition, the inhibitory projections to apical dendrites and basal dendrites of L2/3 neurons are distinct as well (Fino et al., 2013; Ma et al., 2014). Therefore, the depth-dependent changes could be driven by specific inputs. Second, as the reviewer also mentioned, it could be caused by extracellular neuromodulators, such as acetylcholine. For example, previous studies have shown that there is a distinct expression pattern of M2 muscarinic acetylcholine receptor in V1 (Ji et al., 2015). The expression is patchy in L1 where the apical dendrites of L2/3 neurons are located, but the expression of M2 muscarinic acetylcholine receptor in L2/3 where the basal dendrites of L2/3 neurons lie is less intense and uniform (Ji et al., 2015). It has been reported that acetylcholine can influence the excitability of interneurons in a cell-class dependent manner (Brombas et al., 2014). These distinct expression patterns of receptors in different depths could contribute to the depth-dependent changes we observed following visual deprivation. We have discussed these possible mechanisms in the Discussion.

Based on our results and previous literature, we think the feedforward inputs from L4 onto the basal dendrites of L2/3 neurons are potentiated in the beginning (one day after binocular enucleation) and show further potentiation afterwards. The top-down inputs from other regions onto the apical dendrites of L2/3 neurons are weakened first (one day after deprivation), but then recover and eventually undergo potentiation with prolonged deprivation (7 days).

In contrast to V1 L2/3 neurons, we did not see a significant increase of synaptic sGluA1 on apical dendrites of V1 L5 neurons after visual deprivation (Figure 6E, F). Many factors could account for this difference. First, L5 neurons and L2/3 neurons receive very different inputs. The canonical cortical microcircuit in V1 is that thalamic input drives activity in a feedforward and sequential fashion from L4 to L2/3 to L5 and out to other regions although numerous examples of alternate connections exist (Adesnik and Naka, 2018). L5 neurons are considered as one of the main integrators in the cortical column as their dendrites span all cortical layers and thus receive inputs from all layers (Briggs and Callaway, 2005). Further, L5 and L2/3 neurons differ in their dendritic arborization (Rojo et al., 2016; Spruston, 2008). L2/3 neurons have more confined dendritic trees compared to L5 neurons and apical dendrites of L5 neurons extend a greater distance than those of L2/3 neurons to reach the pia surface (Spruston, 2008). Indeed, there have been many studies demonstrating that L2/3 neurons and L5 neurons behave or function differently in response to changes in experience, such as whisker trimming, auditory stimulus, and motor learning (Holtmaat et al., 2006; Holtmaat et al., 2005; Sakata and Harris, 2009; Tjia et al., 2017).

We have now expanded our Discussion to include these circuit and cell-type specific effects of visual deprivation.

3) The observed increase in sGluA1 only occurs after 7 days, while the increase is already observed after 2 days in vitro. Also the initial decrease in sGluA1 observed in vivo has never been reported in vitro (as far as we know). What do the author think is happening at dendrites that show the initial decrease? The authors should discuss this remarkable time course.

The increase in spine sGluA1 on the apical dendrites of L2/3 neurons occurred after 7 days of visual deprivation while in V1 region from the acute brain slice we saw a significant elevation of synaptic GluA1 after 2 days of deprivation (Figure 3B; Figure 5A, B). That discrepancy led us to investigate the changes happening on the basal dendrites of L2/3 neurons. Indeed, we observed that the amount of synaptic sGluA1 on the basal dendrites of L2/3 neurons was significantly increased after one day of deprivation and remained elevated afterwards (Figure 5F-H). Therefore, the increase observed after 2 days in vitro is primarily caused by the increase of synaptic sGluA1 on basal dendrites of L2/3 neurons.

As far as we know, our study is the first report to show that synaptic sGluA1 on the apical dendrites of L2/3 neurons decreases first following visual deprivation and we confirmed that in vivo image data with in vitro biochemical experiments by micro-dissecting out L1 of V1 region (Figure 5C). This reduction only occurs on the apical dendrites of L2/3 neurons but not on their basal dendrites. Previous studies using whole-cell recordings or biochemical methods to examine L2/3 neuron or the whole V1 region are not able to detect that since they measure the average AMPAR level of all synapses from a cell or a larger number of cells in V1. This also highlights the advantage of our in vivo imaging technique that allows us to track individual spine changes during visual deprivation and provides unprecedented levels of spatiotemporal information regarding synaptic AMPAR dynamics.

The dendrites that show the initial decrease are located in the superficial region and they probably receive top-down inputs from regions like retrosplenial cortex and cingulate as we mentioned above. Visual deprivation may induce LTD that results in reduced synaptic AMPAR level in those connections. Extracellular factors like acetylcholine could also contribute to this.

We have now expanded our Discussion to include this remarkable time course.

4) While GluA1 is correlated with synaptic strength (Figure 1D), the two quantities are not equal; there is considerable variation around the linear correlation. Please rephrase.

We agree that spine sGluA1 level does not equate to synaptic strength despite the strong positive correlation detected. We have rephrased the sentences and claim that we are measuring spine sGluA1 level not synaptic strength.

5) No account is taken of the well documented differences in homeostatic response and mechanism involved in subtypes of layer 5 neuron.

We have reanalyzed our data based on the reviewer’s suggestion to look at different populations of layer 5 neurons. L5 basal dendrites are too deep to image in live animals with our two-photon system, and thus we were only able to image apical dendrites of L5 neurons that are located in L1. As a result, we do not have the whole structure of the cells we imaged. It is not reliable to tell which cells are thick tufted and which ones are thin tufted just based on their dendrites in L1. As an alternate way, we categorized the cells based on their responses following binocular enucleation: whether spine sGluA1 on apical dendrites decreased or not at day 1 after enucleation. We found that there are two populations of cells: one population showed decrease in the first few days following deprivation and then gradually recovered while the other group did not decrease at day 1 but exhibited a graduate increase despite that the increase is not significant (Figure 6H). As the reviewer mentioned, based on some previous studies (Greenhill et al., 2015; Holtmaat et al., 2006), the first population of cells might be RS thin tufted L5 cells and the second population might be IB thick tufted L5 cells. Further studies are necessary to confirm the identities of those cells. We have changed the Results and Discussion based on these new analyses.

6) The GRIP KO does show an increase in GluA1 between time VD3 and VD7 points in complete contrast to the conclusion in the text.

There is a recovery of spine GluA1 expression between VD3 and VD7. This recovery could be mediated through some compensatory regulations of other proteins involved in exocytosis of AMPARs, such as GRIP2 (Anggono and Huganir, 2012). We have rephrased our sentences and discussed that recovery. Nevertheless, there is no significant increase of spine sGluA1 following 7 days of enucleation compared to baseline level in GRIP1 KO animals, while in WT control animals synaptic sGluA1 is significantly increased after 7 days’ enucleation. We have compared the effects of genotypes (Figure 7—figure supplement 1) and there is a significant difference between WT control animals and GRIP1 KO animals. These results provide strong evidence that the increase of sGluA1 induced by enucleation is dependent on GRIP1.

7) Unfortunately, the paper does not pay much attention to the quite extensive literature on the mechanism of homeostatic plasticity, nor a good deal of literature conducted on GluA1 knockouts in visual cortex. Neither does it frame the experiments within the appropriate context, because it equates monocular deprivation, dark exposure and eye enucleation experiments within and outside the critical period for development of binocular vision, and thereby cloaks differences between the methods under the misleading term (here) of "sensory deprivation".

We have clarified our usage of “sensory deprivation” in the manuscript and added some discussion regarding the different effects and mechanisms of visual deprivation. Meanwhile, the homeostatic regulation in vivo in our manuscript refers to the homeostatic recovery of neuronal activity of pyramidal neurons in the visual cortex following visual deprivation in intact animals and this has been demonstrated by using either monocular eyelid suture or enucleation. (Barnes et al., 2015; Hengen et al., 2013; Hengen et al., 2016; Keck et al., 2013). Using chronic multielectrode recordings, Hengen et al., showed that the firing rate of L2/3 pyramidal neurons in monocular V1 region dropped first but recovered later following monocular lid suture (Hengen et al., 2013; Hengen et al., 2016). Similarly, the neuronal activity of L2/3 and L5 neurons in V1 measured by calcium imaging decreased after binocular lesions but restored later despite prolonged deprivation (Keck et al., 2013). The underlying mechanisms might be different, but both deprivation paradigms could induce homeostatic regulation of neuronal activity in the visual cortex. The molecular mechanisms underlying this recovery of neuronal activity in vivo remain largely unexplored, and here we examined AMPAR changes in live animals following binocular enucleation to provide molecular insights into this homeostatic plasticity.

As the reviewer pointed out, there are many differences between monocular visual deprivation and binocular deprivation, and different deprivation paradigms, such as enucleation, dark exposure, and eyelid suture, have distinct effects on the visual pathways and plasticity. Further, the plasticity induced by visual deprivation has been shown to be age-dependent (Whitt et al., 2014). We are aware of those differences and thus we confirmed some previous data in our hands with binocular enucleation before we went for further study. For example, when we found that binocular enucleation did not increase synaptic GluA1 on apical dendrites of V1 L2/3 neurons after 2 days. We confirmed that 2 days of binocular enucleation did increase synaptic GluA1 in V1 with ex vivo biochemical experiments, which led us to the hypothesis of depth-dependent changes. More importantly, in our study, all experiments were done under same conditions (same age of animals and same way to deprive the vision), and we did not combine different ways and mix the results. Therefore, we believe that our conclusions are well supported by our data.

Regarding the explanations of our data and how our data fit in with previous studies, we have discussed them in detail in the Discussion (fifth to eighth paragraphs). Differences between our study and previous findings could result from different paradigms used to deprive the vision or different ages of the animals being used. This is a very complex and still on-going topic and beyond the focus of our paper. We wished to study how visual deprivation through binocular enucleation affected real-time AMPAR dynamics in awake mice. Therefore, we did not spend much space in explaining the differences in the very beginning of the Introduction in order to avoid confusion and distraction from our main point. We have rephrased our sentences to make this clear. The GluA1 KO study (Ranson et al., 2013) is very interesting although the ocular dominance plasticity and homeostatic plasticity of neuronal activity are not exactly the same. It still implicates the importance of GluA1 in the visual plasticity and our observation of dynamic changes of GluA1-containing AMPARs during visual deprivation further supports that and provide more mechanistic details.

Reviewer #2:

[…]

1) The authors claim that the synaptic changes depend on cortical depth, but I find this a rather vague concept, which is not much discussed. Can the authors speculate if this is linked to a specific input or to a extracellular factor? Can the authors link their findings to possible VD-induced changes in V1 circuitry? Which synapses are changing and which not? Do L4 inputs target mostly apical or basal dendrites of L2/3 cells? Why would L5 cells not be affected?

Our data show that the changes in synaptic sGluA1 induced by visual deprivation in V1 L2/3 neurons are significantly correlated with the depth (distance from the pia) of the dendrites or spines, wherein deep dendrites or spines are potentiated more than superficial ones (Figure 5; Figure 5—figure supplement 1). There are many possible mechanisms. First, as the reviewer mentioned, it could be input-specific, as apical dendrites and basal dendrites of L2/3 neurons receive different inputs. The basal dendrites primarily receive feedforward inputs from L4 and nearby L2/3 neurons (Ko et al., 2011; Lee et al., 2016; Park et al., 2019), while the apical dendrites receive feedback inputs from regions like retrosplenial cortex, cingulate, and thalamus (Makino and Komiyama, 2015; Roth et al., 2016; Zhang et al., 2014). In addition, the inhibitory projections to apical dendrites and basal dendrites of L2/3 neurons are distinct as well (Fino et al., 2013; Ma et al., 2014). Therefore, the depth-dependent changes could be driven by specific inputs. Second, as the reviewer also mentioned, it could be caused by extracellular neuromodulators, such as acetylcholine. For example, previous studies have shown that there is a distinct expression pattern of M2 muscarinic acetylcholine receptor in V1 (Ji et al., 2015). The expression is patchy in L1 where the apical dendrites of L2/3 neurons are located, but the expression of M2 muscarinic acetylcholine receptor in L2/3 where the basal dendrites of L2/3 neurons lie is less intense and uniform (Ji et al., 2015). It has been reported that acetylcholine can influence the excitability of interneurons in a cell-class dependent manner (Brombas et al., 2014). These distinct expression patterns of receptors in different depths could contribute to the depth-dependent changes we observed following visual deprivation. We have discussed these possible mechanisms in the Discussion.

Based on our results and previous literature, we think the feedforward inputs from L4 onto the basal dendrites of L2/3 neurons are potentiated in the beginning (one day after binocular enucleation) and show further potentiation afterwards. The top-down inputs from other regions onto the apical dendrites of L2/3 neurons are weakened first (one day after deprivation), but then recover and eventually undergo potentiation with prolonged deprivation (7 days).

In contrast to V1 L2/3 neurons, we did not see a significant increase of synaptic sGluA1 on apical dendrites of V1 L5 neurons after visual deprivation (Figure 6E, F). Many factors could account for this difference. First, L5 neurons and L2/3 neurons receive very different inputs. The canonical cortical microcircuit in V1 is that thalamic input drives activity in a feedforward and sequential fashion from L4 to L2/3 to L5 and out to other regions although numerous examples of alternate connections exist (Adesnik and Naka, 2018). L5 neurons are considered as one of the main integrators in the cortical column as their dendrites span all cortical layers and thus receive inputs from all layers (Briggs and Callaway, 2005). Further, L5 and L2/3 neurons differ in their dendritic arborization (Rojo et al., 2016; Spruston, 2008). L2/3 neurons have more confined dendritic trees compared to L5 neurons and apical dendrites of L5 neurons extend a greater distance than those of L2/3 neurons to reach the pia surface (Spruston, 2008). Indeed, there have been many studies demonstrating that L2/3 neurons and L5 neurons behave or function differently in response to changes in experiences, such as whisker trimming, auditory stimulus, and motor learning (Holtmaat et al., 2006; Holtmaat et al., 2005; Sakata and Harris, 2009; Tjia et al., 2017).

We have now expanded our Discussion to include these circuit and cell-type specific effects of visual deprivation.

2) The authors make a strong claim that VD-induced synaptic changes do not occur in L5 (Discussion, sixth paragraph) and that L5 dendrites, in contrast to L2/3 cells, do not show depth-dependent changes after deprivation (e.g. subsection “Lamina-specific increases of spine sGluA1 level after deprivation”, last paragraph). However, they only looked in a small subset of dendrites in L1 and they cannot exclude changes in basal dendrites or dendrites closer to the soma. Furthermore, they actually found a small decrease in sGluA1 in L5 cells after VD (Figure 7E). In addition, the effect in L2/3 as shown in Figure 5—figure supplement 1D is pretty subtle and the number of observations is a lot lower in Figure 6—figure supplement 1I. The most important finding in L2/3, the difference between basal and apical dendrites (or distal/proximal dendrites), was not assessed in L5 cells. I therefore think this claim is too strong and I would ask the authors to either provide additional evidence or to tone this statement down.

Due to the deep location of L5 basal dendrites, we were not able to image those populations of dendrites with our two-photon imaging system. We agree that we cannot exclude the possibility that L5 basal dendrites could increase synaptic sGluA1 upon visual deprivation. We have changed our Abstract and Result, claiming that visual deprivation specifically increases synaptic AMPARs on the apical dendrites of L2/3 neurons but not on the apical dendrites of L5 neurons. In the Discussion, we also discuss the possibility that the basal dendrites of L5 neurons could show synaptic enrichment of AMPARs after visual deprivation.

Reviewer #3:

[…]

1) That GluA1 fluorescence does not equate to synaptic strength.

We agree with the reviewer that the GluA1 signal we imaged in the study does not equate to synaptic strength, but we think it is a good indicator or synaptic strength. Our data as well as many previous studies have observed a strong positive correlation between spine GluA1 level and synaptic strength (Makino and Malinow, 2011).

2) Purported difference in fluorescence by depth measurement along ascending or descending dendrites are based on data where depth is a tiny contributor to the large variance.

The biggest difference we saw is between apical dendrites and basal dendrites, which leads to the hypothesis that the increase induced by enucleation is depth-dependent. We thus examined whether this depth-dependent mechanism applied to dendrites in the same imaging region or even spines along the same dendrite. The correlation between the depth of the dendrites and the increases of sGuA1 induced by enucleation is small but significant (P = 0.006), and the correlation between the depth of the spines along the same dendrite and the increases of sGluA1 induced by enucleation is smaller but still significant (P = 0.0386) while in the control sham-surgery animals none of the correlations was significant. We acknowledge that the effects are small and the small r² indicates that most likely they are not linearly correlated. However, we think these data are very interesting and could be biologically meaningful. Indeed, when we further categorized spines by dendritic compartment and then compared the sGluA1 intensity in the deepest 15% of spines with the most superficial 15% of spines along the same dendritic segment, we found that in both apical and basal dendrites spines positioned more deeply exhibited greater deprivation-induced changes in synaptic sGluA1 expression than spines more superficial poised along the same dendrite (Figure 5K). We could remove them if necessary.

[Editors' note: further revisions were suggested prior to acceptance, as described below.]

Reviewer #1:

The manuscript by Tan et al. is improved over the previous version. However, there is still one big problem.

The statistical shuffling analysis shows that the likelihood of observing the number of "hot" dendrites that were observed in these studies is about 30% (Figure 2—figure supplement 1A, the authors report 53.21% but adding 30% and 40% for 0 or 1 seems more like 70% total for 0 and 1), while the likelihood of observing 3 cold dendrites was very unlikely (3.82%). There is therefore no significant evidence for "hot" dendrites. The authors ought to remove such a reference. The case is better for cold dendrites; there is some evidence for a few cold dendrites. If it were me, I might remove this bit entirely because the numbers are small, but the authors can decide if they want to leave in the part about cold dendrites.

Thank you very much for your insightful comments. We have deleted this cluster result in the revised manuscript as suggested.

The other changes were satisfactory.

Reviewer #2:

The authors have addressed most of my comments in a positive manner. The manuscript has clearly improved and I thank the authors for their complete answer to my and the other reviewers' comments. I particularly appreciate the (perhaps somewhat speculative) discussion on how the depth-dependence of synaptic changes after binocular enucleation could reflect changes in specific inputs. This emphasizes the biological relevance of their findings.

I have the feeling that the manuscript would improve by a stronger focus on the strongest and most interesting results. In my opinion, the presentation of small effects and weak correlations of which the biological relevance are not directly clear is diluting the impact of the main results, which actually are of high interest. The authors may consider to move some of the minor findings (Figure 2J-K, Figure 5I-K) to the supplementary information (or leave them out completely).

We thank the reviewer for the great comments and suggestions. We agree with the reviewer and have moved Figure 2J-K, Figure 5I-K to the supplement as suggested.

Author details

1. Han L Tan

   Solomon H Snyder Department of Neuroscience and Kavli Neuroscience Discovery Institute, Johns Hopkins University School of Medicine, Baltimore, United States

   Contribution

   Conceptualization, Data curation, Formal analysis, Investigation, Visualization, Methodology

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0001-5163-7720

2. Richard H Roth

   Solomon H Snyder Department of Neuroscience and Kavli Neuroscience Discovery Institute, Johns Hopkins University School of Medicine, Baltimore, United States

   Contribution

   Data curation, Investigation

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-6855-999X

3. Austin R Graves

   Solomon H Snyder Department of Neuroscience and Kavli Neuroscience Discovery Institute, Johns Hopkins University School of Medicine, Baltimore, United States

   Contribution

   Data curation

   Competing interests

   No competing interests declared

4. Robert H Cudmore

   Department of Physiology and Membrane Biology, University of California School of Medicine, Davis, United States

   Contribution

   Resources, Software

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-0440-1704

5. Richard L Huganir

   Solomon H Snyder Department of Neuroscience and Kavli Neuroscience Discovery Institute, Johns Hopkins University School of Medicine, Baltimore, United States

   Contribution

   Conceptualization, Resources, Supervision, Funding acquisition, Methodology, Project administration

   For correspondence

   rhuganir@jhmi.edu

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0001-9783-5183

• 1,836

  Page views

• 258

  Downloads

• 11

  Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.