Name: large-scale reconstructions and independent, unbiased clustering based on morphological metrics to classify neurons in selective populations
Uploaded: 2017-02-15T12:30:00.000+00:00
Duration: 12 min 27 s
Description: Watch this Scientific Journal Video about Large-scale Reconstructions and Independent, Unbiased Clustering Based on Morphological Metrics to Classify Neurons in Selective Populations at JoVE.com

We would like to thank Drs. Ed Callaway and Marty Usrey for allowing us to use the tissue prepared as a part of a prior study and Libby Fairless and Shiyuan Liu for help with neuronal reconstructions. This work was funded by the NIH (NEI: EY018683) and the Whitehall Foundation.

<ol>
	<li>Cajal, S. R. y. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=.">.</a> Histologie du systeme nerveaux de l'homme et des vertebres. , (1911).</li><li>Seung, H. S. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Toward+functional+connectomics.">Toward functional connectomics.</a> Nature. 471, 170-172 (2011).</li><li>Callaway, E. M. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Transneuronal+circuit+tracing+with+neurotropic+viruses.">Transneuronal circuit tracing with neurotropic viruses.</a> Current Opinion in Neurobiology. 18, 1-7 (2009).</li><li>Wickersham, I. R., Finke, S., Conzelmann, K. K., Callaway, E. M. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Retrograde+neuronal+tracing+with+a+deletion-mutant+rabies+virus.">Retrograde neuronal tracing with a deletion-mutant rabies virus.</a> Nature Methods. 4, 47-49 (2007).</li><li>Briggs, F., Kiley, C. W., Callaway, E. M., Usrey, W. M. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Morphological+substrates+for+parallel+streams+of+corticogeniculate+feedback+originating+in+both+V1+and+V2+of+the+macaque+monkey.">Morphological substrates for parallel streams of corticogeniculate feedback originating in both V1 and V2 of the macaque monkey.</a> Neuron. 90, 388-399 (2016).</li><li>Cauli, B., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Classification+of+fusiform+neocortical+interneurons+based+on+unsupervised+clustering.">Classification of fusiform neocortical interneurons based on unsupervised clustering.</a> PNAS. 97, 6144-6149 (2000).</li><li>Bragg, E. M., Fairless, E. A., Liu, S., Briggs, F. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Morphology+of+visual+sector+thalamic+reticular+neurons+in+the+macaque+monkey+sugests+retinotopically-specialized,+parallel+stream-mixed+input+to+the+lateral+geniculate+nucleus.">Morphology of visual sector thalamic reticular neurons in the macaque monkey sugests retinotopically-specialized, parallel stream-mixed input to the lateral geniculate nucleus.</a> J. Comparative Neurology. 525 (5), 1273-1290 (2017).</li><li>Osakada, F., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=New+rabies+virus+varients+for+monitoring+and+manipulating+activity+and+gene+expression+in+defined+neural+circuits.">New rabies virus varients for monitoring and manipulating activity and gene expression in defined neural circuits.</a> Neuron. 71, 617-631 (2011).</li><li>Callaway, E. M., Luo, L. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Monosynaptic+circuit+tracing+with+glycoprotein-deleted+rabies+viruses.">Monosynaptic circuit tracing with glycoprotein-deleted rabies viruses.</a> Journal of Neuroscience. 35, 8979-8985 (2015).</li><li>Nhan, H. L., Callaway, E. M. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Morphology+of+superior+colliculus-+and+middle+temporal+area-projecting+neurons+in+primate+primary+visual+cortex.">Morphology of superior colliculus- and middle temporal area-projecting neurons in primate primary visual cortex.</a> J. Comparative Neurology. 520, 52-80 (2012).</li><li>Pinault, D. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=The+thalamic+reticular+nucleus:+structure,+function+and+concept.">The thalamic reticular nucleus: structure, function and concept.</a> Brain Research Reviews. 46, 1-31 (2004).</li><li>Gage, G. J., Kipke, D. R., Shain, W. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Whole+Animal+Perfusion+Fixation+for+Rodents.">Whole Animal Perfusion Fixation for Rodents.</a> Journal of Visualized Experiments. 65, e3564 (2012).</li><li>Wong-Riley, M. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Changes+in+the+visual+system+of+monocularly+sutured+or+enucleated+cats+demonstrable+with+cytochrome+oxidase+histochemistry.">Changes in the visual system of monocularly sutured or enucleated cats demonstrable with cytochrome oxidase histochemistry.</a> Brain Research. 171, 11-28 (1979).</li><li>Gerfen, C. R. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Basic+neuroanatomical+methods.">Basic neuroanatomical methods.</a> Current Protocols in Neuroscience. Chapter 1, Unit 1.1 (2003).</li><li>Thorndike, R. L. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Who+belongs+in+the+family?.">Who belongs in the family?.</a> Psychometrika. 18, 267-276 (1953).</li><li>Talebi, V., Baker, C. I. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Categorically+distinct+types+of+receptive+fields+in+early+visual+cortex.">Categorically distinct types of receptive fields in early visual cortex.</a> Journal of Neurophysiology. 115, 2556-2576 (2016).</li><li>Helmstaedter, M., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Connectomic+reconstruction+of+the+inner+plexiform+layer+in+the+mouse+retina.">Connectomic reconstruction of the inner plexiform layer in the mouse retina.</a> Nature. 500, 168-174 (2013).</li></ol>

The authors have nothing to disclose.

Neuroanatomical studies have remained a pillar of neuroscience and recent interest in connectomics and structure-function relationships has renewed enthusiasm for detailed morphological characterization of selective neuronal populations. Traditionally, neuroanatomical studies have relied on qualitative classifications of neurons into morphologically distinct classes of neurons defined by expert neuroanatomists. With advances in the techniques for reconstructing neurons and extracting morphological data, it is now possible to utilize more sophisticated and quantitative data analysis methods to classify morphologically distinct neuronal classes in an unbiased manner. In this study, step-by-step methods are outlined for 1) selectively labeling hundreds of individual neurons using virus-mediated retrograde labeling methods; 2) systematically reconstructing hundreds of individual neurons to form a comprehensive and therefore representative sample of neurons within a selective population; and 3) independent clustering analyses with statistical evaluation to determine optimal clusters of neurons within a selective population based on morphological metrics. These analysis methods are verified by applying different clustering methods to an existing dataset of TRN neurons and we demonstrate based on three separate evaluation methods that TRN neurons are optimally clustered into three morphologically distinct subpopulations.
A major advantage of the protocols outlined in this study is their flexibility. The G-deleted rabies virus has been effective in driving EGFP expression in hundreds of retrogradely labeled neurons following injection into a target structure. The same rabies virus strain is similarly effective in multiple species including macaque monkeys5,9, ferrets (Hasse &#38; Briggs, in revision), and rodents7,8. Accordingly, modified rabies virus can be utilized to selectively label the complete dendritic arborization patterns of any population of neurons following injection into a target structure. Antibody staining against GFP followed by DAB/peroxide reaction is recommended because this causes permanent staining of labeled neurons and enables tissue sections to be visualized under a microscope without the aid of fluorescence microscopy (which bleaches fluorescent label with longer exposure). However, alternative staining methods are available and/or fluorescence can be directly visualized if desired. An advantage of measuring raw fluorescence is that different viruses driving expression of different fluorescent molecules can be combined, for example to determine whether neurons project to multiple target brain structures. Additionally, the steps outlined above may be applied to all available morphological data (for example axons and dendrites) to enable increasingly robust morphological classification of neurons. It is also important to note that smaller quantities of rabies virus can also be injected in order to generate sparser retrograde labeling, which could be advantageous for reconstructions of axons in addition to dendrites.
An additional advantage of the analysis methods outlined in this study is that multiple different clustering approaches can be utilized and compared. At each step in the cluster analysis, different options are available to calculate between-neuron distances in parameter space as well as different algorithms to generate clusters. Additionally, multiple evaluation methods are available. The latter is especially useful as a means to verify optimal clustering, as outlined in the Representative Results section. Undoubtedly, additional algorithms for generating clusters and for evaluating optimal clustering will become available as the field moves forward. Together, these analysis methods will continue to improve quantitative approaches in the field of neuroanatomy and enhance the robustness of findings.
There is growing interest in automation of neuronal reconstructions because by-hand reconstructions of individual labeled neurons are time consuming, require training and experience, and are still somewhat subjective. Because computer-based automatic neuronal reconstructions are still error-prone, electron microscopy-based neuroanatomical data are still analyzed by people (see for example17). However, it is likely that automated neuronal reconstruction will be available in the near future. Automated neuronal reconstruction will require even more scrutiny on the data analysis end, thus the clustering analyses and evaluation methods outlined here will become even more important as the field continues to advance technologically.

Neuroanatomy is one of the foundations of neuroscience1 and recent interest in &#34;connectomics&#34; has renewed enthusiasm for understanding the morphological diversity of neuronal populations and the connections between specific neurons2. Methods for labeling and reconstructing neurons have greatly improved with recent innovations, including genetic and virus-mediated circuit tracing approaches3,4, enabling more comprehensive morphological surveys of neuronal populations5. In addition to improvements in labeling individual neurons, quantitative data analysis techniques have also emerged that enable independent and unbiased classification of neurons into distinct subpopulations based on morphological data5,6. These unbiased techniques are an improvement upon more traditional qualitative classification methods that have been the standard in the field for over a century. The goal of this study is to outline, step-by-step, the combination of virus-mediated labeling of neurons within a selective population, large-scale reconstructions of a comprehensive sample of these neurons, and quantitative data analysis based on independent clustering with statistical evaluation. By combining these methods, we outline a novel approach toward the collection and analysis of neuroanatomical data to facilitate comprehensive sampling and unbiased classification of morphologically unique neuronal types within a selective neuronal population.
As an example of these methods, we describe our analysis of a large population of neurons within a single sector of the thalamic reticular nucleus (TRN) of the macaque monkey. These data are from a prior study7. Methods for selectively labeling TRN neurons projecting to the dorsal lateral geniculate nucleus of the thalamus (dLGN) using surgical injection of modified rabies virus encoding EGFP4,8 (see Table of Specific Materials/Equipment, row 2) are outlined. This modified rabies virus lacks the gene encoding an essential coat protein, eliminating trans-synaptic movement of the virus. Once the virus enters axon terminals at the injection site, it acts like a traditional retrograde tracer with the important benefit of driving EGFP expression throughout the full dendritic arborization of infected neurons5,9,10. Accordingly, this G-deleted rabies virus can be utilized to selectively infect and label any neuronal population following injection and retrograde transport.
In order to perform a comprehensive analysis of a specific neuronal population, it is important to sample from a broad distribution of neurons within the population. Because the virus-mediated labeling technique produces complete intracellular, &#34;Golgi-like&#34; fills of many neurons with axons at the virus injection site, it is possible to reconstruct a very large sample of neurons within the full extent of a brain structure. Additionally, because the modified rabies virus is so effective at infecting and labeling large numbers of neurons, it is possible to reconstruct hundreds of neurons per animal. Procedures for sampling 160 neurons throughout the visual sector of the TRN11 in order to generate a comprehensive sample of dLGN-projecting TRN neurons are outlined. The process of reconstructing individual neurons using a neuron reconstruction system including a microscope, camera, and reconstruction software is described. Also described are methods to determine positions of individual neurons within a brain structure (in this case within the TRN) and to verify virus injection site volume and location within a structure (in this case within the dLGN) using volumetric contour reconstructions. Steps to export morphological data and perform independent cluster analyses based on morphological metrics measured for each neuron are outlined. There are limitations to clustering methods and there are also a variety of different clustering algorithms available. Accordingly, these options and the benefits of some of the more commonly used algorithms are described. The cluster analysis does not provide statistical verification of the uniqueness of clusters. Therefore, additional steps are outlined to verify optimal clustering as well as the relationships between morphological data within and across clusters. Statistical methods for evaluating clusters for the TRN dataset to confirm that TRN neurons are grouped into three unique clusters based on 10 independent morphological metrics are described.
Thus, by outlining steps for selectively labeling, reconstructing, and analyzing morphological data from a specific neuronal population, we describe methods for quantifying morphological differences among neurons within a population. Prior findings of distinct neuronal types within the visual sector of the macaque monkey TRN are confirmed with separate statistical evaluation methods. Together, we hope these techniques will be broadly applicable to neuroanatomical datasets and help establish quantitative classification of the diversity of neuronal populations through the brain.

<table><tbody><tr><td>SAD&amp;Delta;G-EGFP</td><td>E.M. Callaway Laboratory, Salk Institute</td><td></td><td>Prepared by Dr. F. Osakada. G-deleted rabies virus available through the Salk Institute Viral Core</td></tr><tr><td>Recording electrode: platinum/iridium or tungsten</td><td>FHC</td><td>UEPSGGSE1N2M</td><td>Visit website (www.fh-co.com) for alternative order specifications</td></tr><tr><td>Nanoject II</td><td>Drummod Scientific</td><td>3-000-204, 110V</td><td>Alternatives: picospritzer, Hamilton syringe</td></tr><tr><td>Freezing microtome</td><td>Thermo Scientific</td><td></td><td></td></tr><tr><td>DAB</td><td>Sigma Aldrich</td><td>D5905-50TAB</td><td>3,3&#039;-Diaminobenzidine tetrahydrochloride, tablet, 10 mg substrate per tablet. Caution: carcinogen - must be bleached before discarding</td></tr><tr><td>Cytochrome C</td><td>Sigma Aldrich</td><td>C2037-100MG</td><td></td></tr><tr><td>Catalase</td><td>Sigma-Aldich</td><td>C9322-5G</td><td></td></tr><tr><td>Rabbit anti-GFP</td><td>Life Technologies/Thermo Fisher</td><td>#A-11122&amp;nbsp;</td><td>Primary antibody</td></tr><tr><td>Biotinylated goat anti-rabbit</td><td>Vector Laboratories</td><td>#BA-1000</td><td>Secondary antibody</td></tr><tr><td>Neurolucida System&amp;nbsp;</td><td>MicroBrightField</td><td></td><td>Software for neuron tracing and analysis. http://www.mbfbioscience.com/neurolucida</td></tr><tr><td>Neurolucida Explorer</td><td>MicroBrightField</td><td></td><td>Data export software</td></tr><tr><td>Microfire Camera&amp;nbsp;</td><td>Optronics</td><td></td><td>2-Megapixel true color microscope camera. http://www.simicroscopes.com/pdfs/microfire.pdf</td></tr><tr><td>Nikon E800 Microscope</td><td>Nikon Instruments Inc.</td><td></td><td>Biological research microscope. http://www.microscopyu.com/museum/eclipseE800.html</td></tr><tr><td>Matlab</td><td>The MathWorks Inc.&amp;nbsp;</td><td></td><td>Matrix-based computational mathematics software. http://www.mathworks.com</td></tr><tr><td>Microsoft Office Excel</td><td>Microsoft</td><td></td><td>Spreadsheet program</td></tr></tbody></table>

large-scale reconstructions and independent, unbiased clustering based on morphological metrics to classify neurons in selective populations

This protocol outlines large-scale reconstructions of neurons combined with the use of independent and unbiased clustering analyses to create a comprehensive survey of the morphological characteristics observed among a selective neuronal population. Combination of these techniques constitutes a novel approach for the collection and analysis of neuroanatomical data. Together, these techniques enable large-scale, and therefore more comprehensive, sampling of selective neuronal populations and establish unbiased quantitative methods for describing morphologically unique neuronal classes within a population.
The protocol outlines the use of modified rabies virus to selectively label neurons. G-deleted rabies virus acts like a retrograde tracer following stereotaxic injection into a target brain structure of interest and serves as a vehicle for the delivery and expression of EGFP in neurons. Large numbers of neurons are infected using this technique and express GFP throughout their dendrites, producing &#34;Golgi-like&#34; complete fills of individual neurons. Accordingly, the virus-mediated retrograde tracing method improves upon traditional dye-based retrograde tracing techniques by producing complete intracellular fills.
Individual well-isolated neurons spanning all regions of the brain area under study are selected for reconstruction in order to obtain a representative sample of neurons. The protocol outlines procedures to reconstruct cell bodies and complete dendritic arborization patterns of labeled neurons spanning multiple tissue sections. Morphological data, including positions of each neuron within the brain structure, are extracted for further analysis. Standard programming functions were utilized to perform independent cluster analyses and cluster evaluations based on morphological metrics. To verify the utility of these analyses, statistical evaluation of a cluster analysis performed on 160 neurons reconstructed in the thalamic reticular nucleus of the thalamus (TRN) of the macaque monkey was made. Both the original cluster analysis and the statistical evaluations performed here indicate that TRN neurons are separated into three subpopulations, each with unique morphological characteristics.

We have shown previously that large-scale reconstructions of neurons within a selective population is feasible following injection of modified rabies virus into the dLGN5. Recently, the same tissue was utilized to reconstruct 160 neurons in the visual sector of the TRN (Bragg et al., in review; Figure 2A-B) following the detailed methodological steps described above. In the TRN study, three unique clusters of TRN neurons were identified based on independent cluster analysis of 10 morphological metrics: cell body area, cell body roundness, medial-lateral position relative to the center of the TRN, dorsal-ventral position within the TRN, number of dendritic trees, average dendritic distance to nodes, average length of 3rd and higher order dendrites, average angle of 1st order dendrites, average phase angle of the total dendritic arborization, and angular deviation of the total dendritic arborization. TRN neurons were equivalently distributed across these three clusters (Figure 2C). Separate statistical comparisons of all 10 morphological metrics across neurons in the three clusters yielded statistically significant differences across clusters for all but two of the morphological metrics included in the original cluster analysis. Thus, based on the hierarchical tree dendrogram generated from the cluster analysis (Figure 2C) and the separate statistical comparisons of morphological metrics across clusters, TRN neurons are classified into three unique groups.
In this study, we specifically wanted to apply statistical methods to separately evaluate clusters determined with a cluster analysis. The TRN dataset from the prior study was evaluated to determine whether the three clusters observed were optimal. Separate additional cluster analyses, namely PCA and GMM clustering, were performed to verify the prior clustering methods (Figure 1). Three key outcomes separately verify the prior cluster analysis. Firstly, when the original cluster analysis method was evaluated using the &#39;evalclusters&#39; function specifying the &#39;linkage&#39; function to generate clusters, the optimal number of clusters was 3, matching the original conclusion based on the hierarchical tree dendrogram (Figure 2C). Secondly, PCA was performed on the original data matrix of 10 morphological metrics for 160 TRN neurons as an alternative means of grouping neurons based on contributions of morphological metrics. Three GMMs were generated, assuming TRN neurons were grouped into 1, 2, or 3 unique clusters. The negative log likelihood (NLL) and Akaike information criterion (AIC) yielded the lowest, and therefore most informative, values and both values were optimal when the GMM used 3 clusters to describe TRN morphological data (Figure 3). Thirdly, the GMM-based clustering was separately evaluated using the &#39;evalclusters&#39; function and the optimal number of clusters was 3. Thus observations of the hierarchical tree dendrogram, statistical comparisons of morphological metrics across neurons separated into the three clusters, PCA and GMM-based clustering, and separate evaluations of each clustering method (linkage and GMM), all yielded the same outcome that TRN neurons are optimally separated into 3 clusters based on the 10 morphological metrics utilized.
<img alt="Figure 1" src="/files/ftp_upload/55133/55133fig1.jpg" /> 
Figure 1: Data Format. Example Data matrix with n rows corresponding to the number of neurons and m = 5 columns corresponding to the total number of independent variables to include in the cluster analysis (do not include headers in Data matrix).
<img alt="Figure 2" src="/files/ftp_upload/55133/55133fig2.jpg" /> 
Figure 2: Independent Clustering of 160 TRN Neurons. A. Left, photograph of single coronal section through the dLGN of one animal, stained for cytochrome oxidase activity to visualize dLGN layers and against GFP to visualize virus injection. Arrow indicates regions of dense retrogradely labeled TRN neurons. Section orientation follows the dorsal-ventral/medial-lateral (DV/ML) compass below and scale bar represents 500 &#956;m for all panels in A. Second from left, contour outlines of dLGN (red) and TRN (maroon) for all sections containing virus injection (black contours). Yellow stars indicate injection site centers. Orientation and scale bars as in A. Third from left, 3-d renderings of contours and injection site, orientation and scale bars as in A. Right, maps of locations of reconstructed TRN neurons, color-coded according to cluster assignment (C) within a single aggregate TRN contour (maroon). Orientation and scale bars as in A. B. Left, aggregate contours of TRN (black) and dLGN (grey) with 5 reconstructed TRN neurons colored warm to cool according to their medial-lateral position within the TRN. Orientation according to the DV/ML compass in A, scale bar represents 500 &#956;m. Right, photographs of the same 5 TRN neurons with color-matched scale bars representing 100 &#956;m. C. Hierarchical tree dendrogram following cluster analysis illustrating linkage distances between 160 reconstructed TRN neurons based on 10 independent morphological metrics. Three clusters illustrated in blue, green, and red. All portions of this figure are adapted from figures included in Bragg et al. (in review). <a href="https://cloudflare-jove-com.remotexs.ntu.edu.sg/files/ftp_upload/55133/55133fig2large.jpg" target="_blank">Please click here to view a larger version of this figure.</a>
<img alt="Figure 3" src="/files/ftp_upload/55133/55133fig3.jpg" /> 
Figure 3: Cluster Evaluation. Plot of two different criterion values, negative log likelihood (NLL, red) and Akaike information criterion (AIC, blue) generated from three GMMs assuming 1, 2, or 3 clusters. GMMs with 3 clusters provide the lowest criterion values.
Supplemental Code File: Example code, written in a Matlab-compatible language, to perform cluster analysis on an example Data matrix followed by cluster evaluation using PCA and GMM approaches. <a href="https://cloudflare-jove-com.remotexs.ntu.edu.sg/files/ftp_upload/55133/Briggs&amp;BraggSupplementaryCodeFile.m" target="_blank">Please click here to download this file.</a>

Watch this Scientific Journal Video about Large-scale Reconstructions and Independent, Unbiased Clustering Based on Morphological Metrics to Classify Neurons in Selective Populations at JoVE.com

Large-scale Reconstructions and Independent, Unbiased Clustering Based on Morphological Metrics to Classify Neurons in Selective Populations

This protocol describes large-scale reconstructions of selective neuronal populations, labeled following retrograde infection with a modified rabies virus expressing fluorescent markers, and independent, unbiased cluster analyses that enable comprehensive characterization of morphological metrics among distinct neuronal subclasses.

This protocol outlines large-scale reconstructions of neurons combined with the use of independent and unbiased clustering analyses to create a ...

large-scale-reconstructions-independent-unbiased-clustering-based-on

Nanyang Technological University

Research

JoVE Journal

Neuroscience

6.9K Views.  Geisel School of Medicine at Dartmouth. This protocol describes large-scale reconstructions of selective neuronal populations, labeled following retrograde infection with a modified rabies virus expressing fluorescent markers, and independent, unbiased cluster analyses that enable comprehensive characterization of morphological metrics among distinct neuronal subclasses.

Video: Large-scale Reconstructions and Independent, Unbiased Clustering Based on Morphological Metrics to Classify Neurons in Selective Populations

Automated Sholl Analysis of Digitized Neuronal Morphology at Multiple Scales

Neuronal morphology plays a significant role in determining how neurons function and communicate1-3. Specifically, it affects the ability of neurons to receive inputs from other cells2 and contributes to the propagation of action potentials4,5. The morphology of the neurites also affects how information is processed. The diversity of dendrite morphologies facilitate local and long range signaling and allow individual neurons or groups of neurons to carry out specialized functions within the neuronal network6,7. Alterations in dendrite morphology, including fragmentation of dendrites and changes in branching patterns, have been observed in a number of disease states, including Alzheimer's disease8, schizophrenia9,10, and mental retardation11. The ability to both understand the factors that shape dendrite morphologies and to identify changes in dendrite morphologies is essential in the understanding of nervous system function and dysfunction.
Neurite morphology is often analyzed by Sholl analysis and by counting the number of neurites and the number of branch tips. This analysis is generally applied to dendrites, but it can also be applied to axons. Performing this analysis by hand is both time consuming and inevitably introduces variability due to experimenter bias and inconsistency. The Bonfire program is a semi-automated approach to the analysis of dendrite and axon morphology that builds upon available open-source morphological analysis tools. Our program enables the detection of local changes in dendrite and axon branching behaviors by performing Sholl analysis on subregions of the neuritic arbor. For example, Sholl analysis is performed on both the neuron as a whole as well as on each subset of processes (primary, secondary, terminal, root, etc.) Dendrite and axon patterning is influenced by a number of intracellular and extracellular factors, many acting locally. Thus, the resulting arbor morphology is a result of specific processes acting on specific neurites, making it necessary to perform morphological analysis on a smaller scale in order to observe these local variations12.
The Bonfire program requires the use of two open-source analysis tools, the NeuronJ plugin to ImageJ and NeuronStudio. Neurons are traced in ImageJ, and NeuronStudio is used to define the connectivity between neurites. Bonfire contains a number of custom scripts written in MATLAB (MathWorks) that are used to convert the data into the appropriate format for further analysis, check for user errors, and ultimately perform Sholl analysis. Finally, data are exported into Excel for statistical analysis. A flow chart of the Bonfire program is shown in Figure 1.

We have developed a computer program to analyze neuronal morphology. In combination with two existing open source analysis tools, our program performs Sholl analysis and determines the number of neurites, branch points, and neurite tips. The analyses are performed so that local changes in neurite morphology can be observed.

Neuronal morphology plays a significant role in determining how neurons function and communicate1-3.  Specifically, it affects the ability of neurons to ...

Morphology-Based Distinction Between Healthy and Pathological Cells Utilizing Fourier Transforms and Self-Organizing Maps

The appearance and the movements of immune cells are driven by their environment. As a reaction to a pathogen invasion, the immune cells are recruited to the site of inflammation and are activated to prevent a further spreading of the invasion. This is also reflected by changes in the behavior and the morphological appearance of the immune cells. In cancerous tissue, similar morphokinetic changes have been observed in the behavior of microglial cells: intra-tumoral microglia have less complex 3-dimensional shapes, having less-branched cellular processes, and move more rapidly than those in healthy tissue. The examination of such morphokinetic properties requires complex 3D microscopy techniques, which can be extremely challenging when executed longitudinally. Therefore, the recording of a static 3D shape of a cell is much simpler, because this does not require intravital measurements and can be performed on excised tissue as well. However, it is essential to possess analysis tools that allow the fast and precise description of the 3D shapes and allows the diagnostic classification of healthy and pathogenic tissue samples based solely on static, shape-related information. Here, we present a toolkit that analyzes the discrete Fourier components of the outline of a set of 2D projections of the 3D cell surfaces via Self-Organizing Maps. The application of artificial intelligence methods allows our framework to learn about various cell shapes as it is applied to more and more tissue samples, whilst the workflow remains simple.

Here, we provide a workflow that allows the identification of healthy and pathological cells based on their 3-dimensional shape. We describe the process of using 2D projection outlines based on the 3D surfaces to train a Self-Organizing Map that will provide objective clustering of the investigated cell populations.

The appearance and the movements of immune cells are driven by their environment. As a reaction to a pathogen invasion, the immune cells are recruited to ...

Cancer Research

Electrophysiological and Morphological Characterization of Neuronal Microcircuits in Acute Brain Slices Using Paired Patch-Clamp Recordings

The combination of patch clamp recordings from two (or more) synaptically coupled neurons (paired recordings) in acute brain slice preparations with simultaneous intracellular biocytin filling allows a correlated analysis of their structural and functional properties. With this method it is possible to identify and characterize both pre- and postsynaptic neurons by their morphology and electrophysiological response pattern. Paired recordings allow studying the connectivity patterns between these neurons as well as the properties of both chemical and electrical synaptic transmission. Here, we give a step-by-step description of the procedures required to obtain reliable paired recordings together with an optimal recovery of the neuron morphology. We will describe how pairs of neurons connected via chemical synapses or gap junctions are identified in brain slice preparations. We will outline how neurons are reconstructed to obtain their 3D morphology of the dendritic and axonal domain and how synaptic contacts are identified and localized. We will also discuss the caveats and limitations of the paired recording technique, in particular those associated with dendritic and axonal truncations during the preparation of brain slices because these strongly affect connectivity estimates. However, because of the versatility of the paired recording approach it will remain a valuable tool in characterizing different aspects of synaptic transmission at identified neuronal microcircuits in the brain.

Patch-clamp recordings and simultaneous intracellular biocytin filling of synaptically coupled neurons in acute brain slices allow a correlated analysis of their structural and functional properties. The aim of this protocol is to describe the essential technical steps of electrophysiological recording from neuronal microcircuits and their subsequent morphological analysis.

The combination of patch clamp recordings from two (or more) synaptically coupled neurons (paired recordings) in acute brain slice preparations with ...

Analyzing Dendritic Morphology in Columns and Layers

In many regions of the central nervous systems, such as the fly optic lobes and the vertebrate cortex, synaptic circuits are organized in layers and columns to facilitate brain wiring during development and information processing in developed animals. Postsynaptic neurons elaborate dendrites in type-specific patterns in specific layers to synapse with appropriate presynaptic terminals. The fly medulla neuropil is composed of 10 layers and about 750 columns; each column is innervated by dendrites of over 38 types of medulla neurons, which match with the axonal terminals of some 7 types of afferents in a type-specific fashion. This report details the procedures to image and analyze dendrites of medulla neurons. The workflow includes three sections: (i) the dual-view imaging section combines two confocal image stacks collected at orthogonal orientations into a high-resolution 3D image of dendrites; (ii) the dendrite tracing and registration section traces dendritic arbors in 3D and registers dendritic traces to the reference column array; (iii) the dendritic analysis section analyzes dendritic patterns with respect to columns and layers, including layer-specific termination and planar projection direction of dendritic arbors, and derives estimates of dendritic branching and termination frequencies. The protocols utilize custom plugins built on the open-source MIPAV (Medical Imaging Processing, Analysis, and Visualization) platform and custom toolboxes in the matrix laboratory language. Together, these protocols provide a complete workflow to analyze the dendritic routing of Drosophila medulla neurons in layers and columns, to identify cell types, and to determine defects in mutants.

Here, we show how to analyze dendritic routing of Drosophila medulla neurons in columns and layers. The workflow includes a dual-view imaging technique to improve the image quality and computational tools for tracing, registering dendritic arbors to the reference column array and for analyzing the dendritic structures in 3D space.

In many regions of the central nervous systems, such as the fly optic lobes and the vertebrate cortex, synaptic circuits are organized in layers and ...

Large-scale Three-dimensional Imaging of Cellular Organization in the Mouse Neocortex

The mammalian neocortex is composed of many types of excitatory and inhibitory neurons, each with specific electrophysiological and biochemical properties, synaptic connections, and in vivo functions, but their basic functional and anatomical organization from cellular to network scale is poorly understood. Here we describe a method for the three-dimensional imaging of fluorescently-labeled neurons across large areas of the brain for the investigation of the cortical cellular organization. Specific types of neurons are labeled by the injection of fluorescent retrograde neuronal tracers or expression of fluorescent proteins in transgenic mice. Block brain samples, e.g., a hemisphere, are prepared after fixation, made transparent with tissue clearing methods, and subjected to fluorescent immunolabeling of the specific cell types. Large areas are scanned using confocal or two-photon microscopes equipped with large working distance objectives and motorized stages. This method can resolve the periodic organization of the cell type-specific microcolumn functional modules in the mouse neocortex. The procedure can be useful for the study of three-dimensional cellular architecture in the diverse brain areas and other complex tissues.

Here we describe a procedure for tissue clearing, fluorescent labeling, and large-scale imaging of mouse brain tissue which, thereby, enables visualization of the three-dimensional organization of cell types in the neocortex.

The mammalian neocortex is composed of many types of excitatory and inhibitory neurons, each with specific electrophysiological and biochemical ...

Biocytin Recovery and 3D Reconstructions of Filled Hippocampal CA2 Interneurons

How cortical network activity processes information is of importance to a large number of basic and clinical scientific questions. The protocol described here identifies the basic building blocks of this circuitry. The in-depth studies of cortical regions will ultimately provide other scientists with the circuit components needed for an understanding of how the brain acquires, processes and stores information and what goes wrong in disease, while the electrophysiological and morphological data are widely used by computational neuroscientists in the construction of model networks that explore information processing. The protocol outlined here describes how biocytin-filled cells recorded in the CA2 region of the hippocampus are recovered and then reconstructed in&#160;3D. Additionally, the protocol describes the demonstration of calcium binding protein or peptide content in recorded interneurons.

The protocol outlined here describes the immunofluorescence analysis, biocytin recovery and high-quality reconstructions of hippocampal CA2 interneurons following the intracellular electrophysiological recordings in vitro, allowing neuronal characterization and ultimately fine neuronal anatomy to be studied.

How cortical network activity processes information is of importance to a large number of basic and clinical scientific questions. The protocol described ...

A Method for 3D Reconstruction and Virtual Reality Analysis of Glial and Neuronal Cells

Serial sectioning and subsequent high-resolution imaging of biological tissue using electron microscopy (EM) allow for the segmentation and reconstruction of high-resolution imaged stacks to reveal ultrastructural patterns that could not be resolved using 2D images. Indeed, the latter might lead to a misinterpretation of morphologies, like in the case of mitochondria; the use of 3D models is, therefore, more and more common and applied to the formulation of morphology-based functional hypotheses. To date, the use of 3D models generated from light or electron image stacks makes qualitative, visual assessments, as well as quantification, more convenient to be performed directly in 3D. As these models are often extremely complex, a virtual reality environment is also important to be set up to overcome occlusion and to take full advantage of the 3D structure. Here, a step-by-step guide from image segmentation to reconstruction and analysis is described in detail.

The described pipeline is designed for the segmentation of electron microscopy datasets larger than gigabytes, to extract whole-cell morphologies. Once the cells are reconstructed in 3D, customized software designed around individual needs can be used to perform a qualitative and quantitative analysis directly in 3D, also using virtual reality to overcome view occlusion.

Serial sectioning and subsequent high-resolution imaging of biological tissue using electron microscopy (EM) allow for the segmentation and reconstruction ...

Ballistic Labeling of Pyramidal Neurons in Brain Slices and in Primary Cell Culture

It has been reported that the size and shape of dendritic spines is related to their structural plasticity. To identify the morphological structure of pyramidal neurons and dendritic spines, a ballistic labeling technique can be utilized. In the present protocol, pyramidal neurons are labeled with DilC18(3) dye and analyzed using neuronal reconstruction software to assess neuronal morphology and dendritic spines. To investigate neuronal structure, dendritic branching analysis and Sholl analysis are performed, allowing researchers to draw inferences about dendritic branching complexity and neuronal arbor complexity, respectively. The evaluation of dendritic spines is conducted using an automatic assisted classification algorithm integral to the reconstruction software, which classifies spines into four categories (i.e., thin, mushroom, stubby, filopodia). Furthermore, an additional three parameters (i.e., length, head diameter, and volume) are also chosen to assess alterations in dendritic spine morphology. To validate the potential of wide application of the ballistic labeling technique, pyramidal neurons from in vitro cell culture were successfully labeled. Overall, the ballistic labeling method is unique and useful for visualizing neurons in different brain regions in rats, which in combination with sophisticated reconstruction software, allows researchers to elucidate the possible mechanisms underlying neurocognitive dysfunction.

We present a protocol to label and analyze pyramidal neurons, which is critical for evaluating potential morphological alterations in neurons and dendritic spines that may underlie neurochemical and behavioral abnormalities.

It has been reported that the size and shape of dendritic spines is related to their structural plasticity. To identify the morphological structure of ...

Automatic Identification of Dendritic Branches and their Orientation

The structure of neuronal dendritic trees plays a key role in the integration of synaptic inputs in neurons. Therefore, characterization of the morphology of dendrites is essential for a better understanding of neuronal function. However, the complexity of dendritic trees, both when isolated and especially when located within neuronal networks, has not been completely understood. We developed a new computational tool, SOA (Segmentation and Orientation Analysis), which allows automatic measurement of the orientation of dendritic branches from fluorescence images of 2D neuronal cultures. SOA, written in Python, uses segmentation to distinguish dendritic branches from the image background and accumulates a database on the spatial direction of each branch. The database is then used to calculate morphological parameters such as the directional distribution of dendritic branches in a network and the prevalence of parallel dendritic branch growth. The data obtained can be used to detect structural changes in dendrites in response to neuronal activity and to biological and pharmacological stimuli.

Presented is a computational tool that allows simple and direct automatic measurement of orientations of neuronal dendritic branches from 2D fluorescence images.

The structure of neuronal dendritic trees plays a key role in the integration of synaptic inputs in neurons. Therefore, characterization of the morphology ...

A Tissue Clearing Method for Neuronal Imaging from Mesoscopic to Microscopic Scales

A detailed protocol is provided here to visualize neuronal structures from mesoscopic to microscopic levels in brain tissues. Neuronal structures ranging from neural circuits to subcellular neuronal structures are visualized in mouse brain slices optically cleared with ScaleSF. This clearing method is a modified version of ScaleS and is a hydrophilic tissue clearing method for tissue slices that achieves potent clearing capability as well as a high-level of preservation of fluorescence signals and structural integrity. A customizable three dimensional (3D)-printed imaging chamber is designed for reliable mounting of cleared brain tissues. Mouse brains injected with an adeno-associated virus vector carrying enhanced green fluorescent protein gene were fixed with 4% paraformaldehyde and cut into slices of 1-mm thickness with a vibrating tissue slicer. The brain slices were cleared by following the clearing protocol, which include sequential incubations in three solutions, namely, ScaleS0 solution, phosphate buffer saline (&#8211;), and ScaleS4 solution, for a total of 10.5&#8211;14.5 h. The cleared brain slices were mounted on the imaging chamber and embedded in 1.5% agarose gel dissolved in ScaleS4D25(0) solution. The 3D image acquisition of the slices was carried out using a confocal laser scanning microscope equipped with a multi-immersion objective lens of a long working distance. Beginning with mesoscopic neuronal imaging, we succeeded in visualizing fine subcellular neuronal structures, such as dendritic spines and axonal boutons, in the optically cleared brain slices. This protocol would facilitate understanding of neuronal structures from circuit to subcellular component scales.

The protocol provides a detailed method of neuronal imaging in brain slice using a tissue clearing method, ScaleSF. The protocol includes brain tissue preparation, tissue clarification, handling of cleared slices and confocal laser scanning microscopy imaging of neuronal structures from mesoscopic to microscopic levels.

A detailed protocol is provided here to visualize neuronal structures from mesoscopic to microscopic levels in brain tissues. Neuronal structures ranging ...