概要

Semi-Quantitative Determination of Dopaminergic Neuron Density in the Substantia Nigra of Rodent Models using Automated Image Analysis

Published: February 02, 2021
doi:

概要

Here we present an automated method for semi-quantitative determination of dopaminergic neuron number in the rat substantia nigra pars compacta.

Abstract

Estimation of the number of dopaminergic neurons in the substantia nigra is a key method in pre-clinical Parkinson’s disease research. Currently, unbiased stereological counting is the standard for quantification of these cells, but it remains a laborious and time-consuming process, which may not be feasible for all projects. Here, we describe the use of an image analysis platform, which can accurately estimate the quantity of labeled cells in a pre-defined region of interest. We describe a step-by-step protocol for this method of analysis in rat brain and demonstrate it can identify a significant reduction in tyrosine hydroxylase positive neurons due to expression of mutant α-synuclein in the substantia nigra. We validated this methodology by comparing with results obtained by unbiased stereology. Taken together, this method provides a time-efficient and accurate process for detecting changes in dopaminergic neuron number, and thus is suitable for efficient determination of the effect of interventions on cell survival.

Introduction

Parkinson's disease (PD) is a prevalent neurodegenerative movement disorder characterized by the presence of protein aggregates containing α-synuclein (α-syn) and the preferential loss of dopaminergic neurons in the substantia nigra pars compacta (SNpc)1. Quantification of dopaminergic neuron number is a vital part of PD research as it permits the evaluation of the integrity of the nigrostriatal system, thus, providing an important endpoint to assess the effectiveness of potential disease-modifying therapeutics. Currently, the standard for quantification of cell number is unbiased stereological counting, which utilizes two-dimensional (2D) cross-sections of tissue to estimate volumetric features in three-dimensional (3D) structures2,3,4. Modern design-based stereological methods employ comprehensive random sampling procedures and apply counting protocols (known as probes) to avoid potential artifacts and systematic errors, allowing for reliable detection of differences only slightly greater than inter-animal variation5. While stereology provides a powerful analytical tool for in vivo histological studies, it is time intensive, assumes uniform specimen preparation, and requires validation at several steps, which can impact the efficiency increasingly required for pre-clinical translational investigation.

Recent technological advances in digital science make it possible to adopt novel applications for more efficient assessments of pathology without a stereomicroscope, while filling a need as a surrogate of unbiased stereology. These methods increase speed, reduce human error, and improve the reproducibility of stereological techniques6,7. HALO is one such image analysis platform for quantitative tissue analysis in digital pathology. It comprises a variety of different modules and reports morphological and multiplexed expression data on a cell-by-cell basis across entire tissue sections using pattern recognition algorithms. The cytonuclear FL module measures the immunofluorescent positivity of fluorescent markers in the nucleus or cytoplasm. This allows for reporting of the number of cells positive for each marker, and the intensity score for each cell. The module can be adapted to provide individual cell sizes and intensity measurements, although this feature is not required for quantification of dopaminergic neurons.

The aim of this study is to verify this method with a previously validated viral vector-based α-syn rat model of nigral neurodegeneration8,9,10. In this model, human mutant A53T α-syn is expressed in the SNpc by stereotactic injection of adeno-associated virus hybrid serotype 1/2 (AAV1/2), resulting in significant neurodegeneration over a period of 6 weeks. The contralateral uninjected SNpc may, in some studies, serve as an internal control for the injected side. More commonly, injection of AAV-Empty Vector (AAV-EV) in a control cohort of animals is used as a negative control. We present a step-by-step guide to estimate the density of dopaminergic neurons remaining in the injected SNpc after 6 weeks using an automated image analysis software (Figure 1).

Protocol

All procedures were approved by the University Health Network Animal Care Committee and performed in accordance with guidelines and regulations set by the Canadian Council on Animal Care.

1. Stereotactic injection

  1. Pair-house adult female Sprague-Dawley rats (250-280 g) in cages with wood bedding and ad lib access to food and water. Maintain the animal colony in a regular 12 h light/dark cycle (lights on 06:30) with constant temperature and humidity.
  2. Perform unilateral stereotactic injection of AAV directly to the SNpc on the right side of the brain (right or left side, according to the preferences of each lab) as previously described8,10. Inject 2 μL of AAV1/2 at a final titer of 3.4 x 1012 genomic particles/mL.

2. Brain sectioning and immunohistochemistry (IHC)

  1. Anaesthetize the rat with 5% isoflurane by placing in an anaesthetizing chamber for 3 min. Other approved methods may be used for this step after appropriate institutional review.
  2. Once the rat has reached a surgical plane of deep anesthesia, transfer it to a nose cone firmly affixed to a necropsy table. Secure the rat's fore-paws using tape and use toe pinch-response method to determine the depth of the anesthesia. The animal must be unresponsive before continuing.
  3. Make a lateral incision below the sternum and cut through the diaphragm along the entire length of the rib cage to expose the pleural cavity. Lift and clamp the sternum with a hemostat and place above the head.
  4. Clamp the heart using forceps and insert a butterfly needle connected to a perfusion pump into the posterior end of the left ventricle. Perfuse rat transcardially with 150 mL of heparinized saline, or until the eyes and skin are clear. Perfusion with 4% paraformaldehyde (PFA), instead of saline, may be preferred to facilitate immunostaining with certain antibodies or thinner brain sectioning.
  5. Once perfusion is complete, decapitate with a guillotine and extract the brain to a brain matrix, ventral surface facing up.
  6. Using a fresh razor blade, make a cut in the coronal plane 2 mm rostral to the optic chiasm. Slide the blade from side to side to avoid warping the brain while slicing.
  7. Immerse the posterior portion of the brain in a pre-labeled vial containing approximately 20 mL of 4% PFA for 48 h of post-fixation at room temperature. The anterior portion of the brain may be flash frozen in 2-methylbutane chilled to -42 °C before storage at -80 °C.
  8. After 48 h, transfer the fixed brains to a labeled vial containing 30% sucrose in phosphate buffered saline (PBS) and store at 4 °C until they sink (48-72 h).
  9. Prepare a microtome by placing dry ice in the trough of the specimen stage, followed by 100% ethanol. Once the stage has cooled, squeeze optimal cutting temperature (OCT) compound onto the stage until it forms a circle 2 cm in diameter and 0.5 cm thick. Once it has partially frozen, carefully lower the brain onto the mound of OCT, ensuring the striatal cutting surface remains parallel with the stage.
  10. Add more dry ice to the stage to help the brain to freeze. Once the brain has turned a cream color, clear the stage of dry ice.
  11. Poke a hole into the right side of the brain with a 25G needle to distinguish the right and left hemispheres. Take care not to pass the needle through anatomical structures of interest.
  12. Serially cut 40 μm sections in the coronal plane beginning at bregma -3.8 and ending at bregma -6.8.
  13. Store six series in labeled tubes with anti-freeze solution (40% PBS, 30% 2-ethoxyethanol, 30% glycerol). Each series should contain 12 brain sections.
  14. Select one set of sections for immunohistochemical staining, and wash off anti-freeze solution with 3 x 10 min washes in 0.2% PBS-T.
  15. Block for 1 h at RT with gentle nutation in blocking solution (10% normal goat serum (NGS), 2% bovine serum albumin (BSA) in 0.2% PBS-T). Follow this with incubation with rabbit anti-tyrosine hydroxylase (TH) antibody (1:500) and mouse anti-α-syn antibody (1:500) in 2% NGS in 0.2% PBS-T overnight at room temperature.
  16. Wash off primary antibody with 3 x 5 min washes in 0.2% PBS-T, followed by 1 h incubation with goat anti-rabbit Alexa Fluor 488 secondary antibody (1:500) and goat anti-mouse Alexa Fluor 555 secondary antibody in 2% NGS in 0.2% PBS-T. Ensure the sections are protected from light and nutating gently.
  17. Wash off secondary antibody with 3 x 5 min washes in 0.2% PBS-T and mount the complete set of sections on slides protected from light and dust using a narrow paintbrush. Coverslip with fluorescence mounting medium and seal with clear nail varnish.

3. Confocal microscopy and image acquisition

  1. Capture IHC images using software coupled to a confocal microscope at 10x magnification. Open the pinhole to 1.5 AU to capture a wide plane totaling ~1.5 μm and set the focus on the injected side of the brain.
  2. On the Acquisition tab, check the Tile Scan imaging option and set the dimensions to 10 x 4.
  3. Under the Acquisition Mode panel, set the Zoom to 1.1. This helps to avoid any obvious stitching marks between tile scan images.
  4. Set the Frame Size to 1024 x 1024 pixels and the Averaging to 2 to ensure high quality image acquisition.
  5. In the Channels panel, set track 1 to Alexa488 and track 2 to Alexa555.
  6. Load the slide onto the stage and choose a section with strong TH staining. Click on Live on the Acquisition panel.
  7. In the Channels panel, set the Laser Strength and Gain to levels that maximize signal and limit noise from the background. Use the range indicator to ensure that the signals are not overexposed (as indicated by a dark red overlay).
  8. Repeat the above step with multiple slides to ensure staining is consistent between slides as the laser strength/gain cannot be adjusted between slides.
  9. On the Acquisition tab, check the Positions box.
  10. At this point, you are ready to begin imaging. Using the eyepiece, choose the first section showing positive TH staining, set the focus at the point of interest (i.e., SNpc) and then move the stage to the midline of the section. This saves the position in the x, y, and z axes and will image a tile scan capturing the whole section.
  11. Repeat the above step for all sections throughout the SNpc giving a complete set of images of the SNpc. If detailed analysis of the uninjected side is required, steps 3.10 to 3.11 should be repeated by setting the focus on the uninjected side.

4. Image analysis and quantitation

  1. Separate image files using appropriate software and import image files to automated image analysis software.
  2. Define a region of interest by selecting the Pen annotation tool to draw an annotation around the SNpc.
    NOTE: In sections which have a large amount of dopaminergic neuron loss, temporarily increasing the emittance/absorption can help to clearly define the SNpc (Figure 2).
  3. Move to the Analysis tab and from the drop-down Analyze menu, select Real-Time Tuning. This opens a separate window on the section image allowing for real-time modification of analysis parameters (Figure 3).
  4. Under the Analysis Magnification section, select the appropriate image zoom.
  5. Under the Cell Detection section, select nuclear dye as the dye used for TH staining (Alexa Fluor 488).
  6. Adjust the Nuclear Contrast Threshold, Minimum Nuclear Intensity, Nuclear Segmentation Aggressiveness, and Nuclear Size settings while carefully watching the Real-Time Tuning window.
    NOTE: Accurate representation of each individual cell as a single cell in the Real-Time Tuning window is vital for accuracy. These settings are on an arbitrary scale depending on the software used, but correct adjustment is needed to allow the software to accurately differentiate between individual cells, and between cells and the background (Figure 3).
  7. Repeat this process with a minimum of 10 separate samples to ensure a uniform agreement of what constitutes a cell across different sections.
    NOTE: Additional cell markers (such as α-syn or NeuN) can be identified within the same analysis platform using the Marker 1 or Marker 2 sections on the analysis tab.
  8. Once an appropriate number of images have been sampled and Real-Time Tuning has been adjusted accordingly, save the analysis settings in the Settings Actions drop-down menu.
  9. Select all images to be analyzed and click on Analyze.
  10. Choose the analysis setting you have just saved and in the Region of Analysis window, check the Annotation Layer(s) box. Then, check Layer 1 and click on Analyze.
    NOTE: For a single brain, the analysis typically takes about 5 mins. The completed result will clearly show each item that has been counted as a cell (Figure 4).
  11. Once complete, export the summary analysis data for all sections. There is an option to export Object Analysis Data, which will give detailed data, including cell size of each individual cell detected. This dataset could be used to examine changes in cell size in response to a toxin/therapeutic.
  12. Add the Total Cells from each section analyzed per animal and the Total Analyzed Area (mm2). Divide the total number of cells by the total area analyzed to calculate the number of cells/mm2 in the SNpc for each rat

Representative Results

By applying the above methods to brain tissue collected 6 weeks after AAV injections, we demonstrated that stereotactic injection of AAV expressing mutant A53T α-syn (AAV-A53T) in the SNpc of rat brain results in a significant reduction in the density of dopaminergic neurons compared to injection of empty vector AAV (AAV-EV) as a control (Figure 5A,B). The mean number of TH-positive neurons/mm2 in the SNpc of rats injected with AAV-EV was 276.2 ± 34.7, and in the SNpc of rats injected with AAV-A53T was 41.2 ± 17 (P = 0.0003). Quantification of the number of dopaminergic neurons/mm2 in the SNpc is similar to previously published reports10, 11. For the methods described here, 4 sequential sections per animal were analyzed. Previous studies have shown significant differences with as little as 3 sections, but analysis can be further increased up to 12 sections to encompass the whole SNpc depending on the model and intervention being studied by the investigator.

Unbiased stereology was also performed as previously described12 on another set of brain sections from the same animals. Using this method, we also demonstrated that stereotactic injection of mutant A53T α-syn in the SNpc of rat brain results in a significant reduction in the estimated total number of TH-positive neurons in the SNpc, as compared to injection of EV-AAV (Figure 5C). Importantly, there was a strong correlation between the dopaminergic neuron density estimated using automated image analysis software and dopaminergic neuron number estimated using unbiased stereology (r = 0.8819, P=0.0007) (Figure 5D).

We also applied our methods using automated image analysis software to determine the number of TH-positive neurons/mm2 on the uninjected side of rats injected with AAV-A53T or AAV-EV. The mean number of TH-positive neurons/mm2 in the uninjected SNpc of rats injected with AAV-A53T was 123.2 ± 26.4, which was significantly greater than in the injected SNpc, which was 44.0 ± 15.8 (P = 0.0331) (Supplementary Figure 1A). The mean number of TH-positive neurons/mm2 in the uninjected SNpc of rats injected with AAV-EV (215.6 ± 35.5) was not significantly different from the injected SNpc (276.2 ± 34.7), confirming there was no degeneration due to injection with AAV-EV (Supplementary Figure 1B). We calculated these results as a percentage of injected/uninjected and found that animals injected with AAV-A53T had a 69% reduction compared to the AAV-EV animals (Supplementary Figure 1C).

Figure 1
Figure 1: Workflow of the method. Workflow demonstrating the steps required to inject AAVs, section and stain tissue, define a region of interest and optimize the software for counting of cells. Representative images of confocal tile scan, ROI definition, and quantitation of cells. Scale bar = 100 μm. Please click here to view a larger version of this figure.

Figure 2
Figure 2: Defining the region of interest. (A) Coronal brain section, including the SNpc immunostained for TH (green) from a rat injected with AAV-A53T α-syn. In rats with severe neurodegeneration (such as shown here), it can be difficult to identify the SNpc. (B) Temporarily increasing the absorption of the image can identify the structure and allow an accurate identification of the region of interest. Scale bar = 1 mm. Please click here to view a larger version of this figure.

Figure 3
Figure 3: Optimizing cell detection using the Cytonuclear method in HALO. Real-time tuning of the cytonuclear module allows the user to see changes in cell detection in real-time by altering Nuclear Contrast Threshold, Minimum Nuclear Intensity, Nuclear Segmentation Aggressiveness, and Nuclear Size. (A) Representative image with region of interest displayed. (B) Real-time tuning showing under-sampling in which the software does not detect all cells in the tuning window. (C) Over-sampling in which the software detects more cells than are evident in the tuning window. (D) Optimized tuning in which the correct number of cells are counted. Scale bar = 500 μm. Please click here to view a larger version of this figure.

Figure 4
Figure 4: HALO optimized settings within a defined region of interest. Representative images of completed analysis using optimized settings for cytonuclear detection in HALO. Scale bar = 500 μm. Please click here to view a larger version of this figure.

Figure 5
Figure 5: Expression of human mutant A53T α-syn in SNpc results in severe neurodegeneration at 6 weeks as quantified by HALO and unbiased stereology. (A) Representative images showing degeneration of TH-positive neurons in the SNpc 6 weeks after stereotactic injection of AAV-A53T α-syn at a titer of 3.4 x 1012 viral particles/mL. Immunofluorescent staining with anti-TH (green) and anti-α-syn (red) antibodies. Scale bar = 200 µm. (B) Quantification of the number of TH-positive neurons in the SNpc of rats injected with AAV-A53T α-syn or AAV-EV demonstrates that expression of mutant α-syn results in significant dopaminergic neuron loss. Unpaired t-test; n = 5 rats/group. Graph shows mean ± SEM, ***P < 0.001. (C) Representative images of colorimetric staining of dopaminergic neurons in the SNpc of AAV-A53T (left) or AAV-EV (right) injected rats used to perform unbiased stereology. Scale bar = 200 µm. (D) A significant correlation between HALO counting of TH-positive neurons/mm2 (y-axis) and unbiased stereology cell numbers (x-axis). Pearson correlation (r = 0.8819, P = 0.0007). Please click here to view a larger version of this figure.

Supplementary Figure 1: Significant unilateral neurodegeneration is observed in the SNpc of rats who received AAV-A53T injection. (A) Quantification of the number of TH-positive neurons in the injected or uninjected SNpc of rats that received a unilateral AAV-A53T stereotactic injection shows a significant decrease on the injected side. (B) Quantification of the number of TH-positive neurons in the injected or uninjected SNpc of rats that received unilateral AAV-EV stereotactic injection shows no significant changes. (C) Normalization to the uninjected contralateral side demonstrate a >50% decrease upon injection with AAV-A53T α-syn compared to AAV-EV. Unpaired t-test; n = 5 rats/group. Graphs show mean ± SEM, *P < 0.05, ***P < 0.001. Please click here to download this Supplementary Figure.

Discussion

The reliable assessment of the integrity of the dopaminergic system in pre-clinical models of PD is critical to determine the effectiveness of potential disease-modifying therapeutics. Therefore, it is important to control and minimize potential confounds that may reduce the reliability and reproducibility of histopathological data. Careful quantitative outcomes can provide more information than qualitative or semi-quantitative descriptions alone. At the same time, we must recognize that constraints in time and resources can make it difficult to perform unbiased stereological counting to quantitate pathological changes or loss of cells. However, with recent advancements, many of these criteria can be fulfilled using computerized and automated imaging platforms6.

This protocol describes a number of important steps in determining the stereological estimation of dopaminergic neurons/mm2 in rodent brain. It should be noted that we used stereotactic injection to deliver AAVs to the SNpc, which places an importance on accurate delivery of these viruses in order to determine the effect of any treatment. The co-ordinates used for our study are bregma -5.2 mm (anterior-posterior), -2 mm (medial-lateral to the right) and -7.5 mm (dorsal ventral from the skull) on adult female rats weighing ~275 g. Correct delivery of the virus using these co-ordinates will ensure delivery of the AAVs to the SNpc.

In addition to this, care must be taken when preparing tissues for staining. Foremost, care should be taken to mark which side is the right/left. In our hands, the easiest and most
reproducible way to do this was to use a 25G needle to poke a hole periodically in the left dorsal midbrain while identify the uninjected side. An in-depth knowledge of the neuroanatomical regions of the rodent brain is important to differentiate when to begin cutting, and later to identify the SNpc when drawing a region of interest. It is important to note that TH staining does not exclusively stain neurons of the SNpc, and one should be able to differentiate between TH-positive neurons in the ventral tegmental area and retrorubal field. A rat brain atlas is a useful guide for those lacking experience in rodent neuroanatomy. Antibody incubation times are uniform throughout the protocol and the focus is set in the minimum time possible to avoid photobleaching.

There is a low variability between samples once consistent neuroanatomical features are used to distinguish between sections. Having an experienced, blinded observer draw the region of interest is important to maintain consistency across the sections to be analyzed. Analysis of the uninjected side as shown in Supplementary Figure 1 demonstrates there is a consistency to the method. However, care should be taken when interpreting data in this manner. Slicing brains consistently relies on comparing the presence of known neuroanatomical features on each side of the brain. Subtle differences may mean that the uninjected side on a given section may be more posterior/anterior than the injected side, and thus not an accurate direct comparison. Further analysis of the presented data highlights a limitation of this method. There is a large degree of variability in the dopaminergic neuron counts in the uninjected side (123.2 in the AAV-A53T group vs 215.6 in the AAV-EV), which one would not expect between similarly aged animals. There are a number of possible reasons for this discrepancy, including the small sample size used in this study and the above-mentioned anatomical differences from side to side. In addition to this, focus for fluorescent signal is set on the injected side, meaning a slight shift in the z-plane can leave the uninjected side out of focus, and thus some cells will not be detectable by the cell counting software. For this reason, it is recommended to compare the number of dopaminergic neurons/mm2 on one side using known neuroanatomical features, and where focus has been set on the microscope specifically for this region. If a comparison to the uninjected side is needed, imaging should be repeated with the focus set on this side.

We present data obtained and analyzed with one of the several software packages that provide quantitative image analysis with minimal user training13. Other programs are available and are increasingly being adopted to study neuropathology, including the quantification of dopaminergic neuron loss in experimental models of PD6. While some offer built-in algorithms to perform specific functions (e.g., amyloid plaque counting), many of the software packages share several common features that make them particularly useful for laboratories that perform repetitive analysis. In addition to region of interest identification, preprocessing steps that interfere with the signal (tissue folds, edge artifacts, ink marks, stains, etc.) can be removed from region of interest by manual and pattern recognition tools. Furthermore, the readouts provided can be used in conjunction with quantitative values obtained from protein, cellular, or behavioral studies to explore relationships and potential correlations between variables.

The method described above provides an accurate, semi-quantitative, and relatively inexpensive procedure for estimating neuron numbers in immunohistochemically stained brain sections. The real-time tuning feature allows for adaption of the method for all sections and ensures consistency even with slight variances in slicing methods. Having a blinded observer identify regions of interest in the SNpc mitigates confirmation or selection biases and the software provides a simple readout allowing calculation of neurons/mm2. This method could be easily adapted to provide accurate counts of neurons in different brain regions while maintaining efficiency and cost advantages over more established stereological protocols. Advancement of slide-scanning technology for thicker sections would also allow for more efficient processing time for the above-described method.

Limitations
The method described has a number of distinct advantages in its speed and ability to detect large changes in neuron density, but also poses its own challenges. Unlike stereology, the method cannot provide an estimation of absolute cell numbers but instead calculates cell density within the SNpc. The estimates of cell density are highly, but not perfectly, correlated with cell numbers obtained using unbiased stereology. In addition, the method relies on tuning of cell size and shape by the user prior to beginning the analysis. This cannot guarantee that cells will not be missed due to being partially in the focal plane, or that cells of an unusual size and/or shape will be missed by the software. In the authors' experience, the method is particularly useful for examining the efficacy of potential disease-modifying therapeutics in pre-clinical rodent models. In conclusion, while stereological methods to quantify cell numbers remain widely used in neuroscience, the rapid acceleration of digitalization suggests that automated image analysis platforms will increasingly be adopted to study neuropathology, particularly as they continue to improve. It is important that the investigator understands the technical limitations of this approach and applies this methodology after careful consideration.

開示

The authors have nothing to disclose.

Acknowledgements

The authors would like to thank all the staff at the Advanced Optical Microscopy Facility (AOMF) at University Health Network for their time and assistance in developing this protocol.

Materials

A-Syn Antibody ThermoFisher Scientific 32-8100
ABC Elite Vector Labs PK-6102
Alexa Fluor 488 secondary antibody ThermoFisher Scientific A-11008
Alexa Fluor 555 secondary antibody ThermoFisher Scientific A-28180
Alkaline phosphatase-conjugated anti-rabbit igG Jackson Immuno 111-055-144
Biotinylated anti-mouse IgG Vector Labs BA-9200
Bovine Serum Albumin Sigma A2153
DAKO fluorescent mouting medium Agilent S3023
HALO™ Indica Labs
Histo-Clear II Diamed HS202
ImmPACT DAB Peroxidase substrate Vector Labs SK-4105
LSM880 Confocal Microscope Zeiss
NeuN Antibody Millipore MAB377
Normal Goat Serum Vector Labs S-1000-20
OCT Tissue-Tek
Paraformaldehyde BioShop PAR070.1
Sliding microtome Leica SM2010 R
Stereo Investigator MBF Bioscience
Sucrose BioShop SUC700
TH Antibody ThermoFisher Scientific P21962
VectaMount mounting medium Vector Labs H-5000
Vector Blue Alkaline Phosphatase substrate Vector Labs SK-5300
Zen Black Software Zeiss
Zen Blue Software Zeiss

参考文献

  1. Kalia, L. V., Lang, A. E. Parkinson’s disease. Lancet. 386 (9996), 896-912 (2015).
  2. West, M. J., Slomianka, L., Gundersen, H. J. Unbiased stereological estimation of the total number of neurons in thesubdivisions of the rat hippocampus using the optical fractionator. The Anatomical Record. 231 (4), 482-497 (1991).
  3. Nair-Roberts, R. G., et al. Stereological estimates of dopaminergic, GABAergic and glutamatergic neurons in the ventral tegmental area, substantia nigra and retrorubral field in the rat. 神経科学. 152 (4), 1024-1031 (2008).
  4. Golub, V. M., et al. Neurostereology protocol for unbiased quantification of neuronal injury and neurodegeneration. Frontiers in Aging Neuroscience. 7, 196 (2015).
  5. Schmitz, C., Hof, P. R. Design-based stereology in neuroscience. 神経科学. 130 (4), 813-831 (2005).
  6. Penttinen, A. M., et al. Implementation of deep neural networks to count dopamine neurons in substantia nigra. European Journal of Neuroscience. 48 (6), 2354-2361 (2018).
  7. Yousef, A., et al. Neuron loss and degeneration in the progression of TDP-43 in frontotemporal lobar degeneration. Acta Neuropathologica Communications. 5 (1), 68 (2017).
  8. Koprich, J. B., et al. Expression of human A53T alpha-synuclein in the rat substantia nigra using a novel AAV1/2 vector produces a rapidly evolving pathology with protein aggregation, dystrophic neurite architecture and nigrostriatal degeneration with potential to model the pathology of Parkinson’s disease. Molecular Neurodegeneration. 5, 43 (2010).
  9. Koprich, J. B., et al. Progressive neurodegeneration or endogenous compensation in an animal model of Parkinson’s disease produced by decreasing doses of alpha-synuclein. PLoS One. 6 (3), 17698 (2011).
  10. McKinnon, C., et al. Early-onset impairment of the ubiquitin-proteasome system in dopaminergic neurons caused by alpha-synuclein. Acta Neuropathologica Communications. 8 (1), 17 (2020).
  11. Henderson, M. X., et al. Spread of alpha-synuclein pathology through the brain connectome is modulated by selective vulnerability and predicted by network analysis. Nature Neuroscience. 22 (8), 1248-1257 (2019).
  12. Ip, C. W., et al. AAV1/2-induced overexpression of A53T-alpha-synuclein in the substantia nigra results in degeneration of the nigrostriatal system with Lewy-like pathology and motor impairment: a new mouse model for Parkinson’s disease. Acta Neuropathologica Communications. 5 (1), 11 (2017).
  13. Webster, J. D., Dunstan, R. W. Whole-slide imaging and automated image analysis: considerations and opportunities in the practice of pathology. Veterinary Pathology. 51 (1), 211-223 (2014).

Play Video

記事を引用
O’Hara, D. M., Kapadia, M., Ping, S., Kalia, S. K., Kalia, L. V. Semi-Quantitative Determination of Dopaminergic Neuron Density in the Substantia Nigra of Rodent Models using Automated Image Analysis. J. Vis. Exp. (168), e62062, doi:10.3791/62062 (2021).

View Video