Micro-CT Imaging and Morphometric Analysis of Mouse Neonatal Brains

Summary

This study describes the steps for obtaining high-resolution images of neonatal mouse brains by combining micro-computed tomography (micro-CT) and a contrast agent in ex vivo samples. We describe basic morphometric analyses to quantify brain size and shape in these images.

Abstract

Neuroimages are a valuable tool for studying brain morphology in experiments using animal models. Magnetic resonance imaging (MRI) has become the standard method for soft tissues, although its low spatial resolution poses some limits for small animals. Here, we describe a protocol for obtaining high-resolution three-dimensional (3D) information on mouse neonate brains and skulls using micro-computed tomography (micro-CT). The protocol includes those steps needed to dissect the samples, stain and scan the brain, and obtain morphometric measurements of the whole organ and regions of interest (ROIs). Image analysis includes the segmentation of structures and the digitization of point coordinates. In sum, this work shows that the combination of micro-CT and Lugol's solution as a contrast agent is a suitable alternative for imaging the perinatal brains of small animals. This imaging workflow has applications in developmental biology, biomedicine, and other sciences interested in assessing the effect of diverse genetic and environmental factors on brain development.

Introduction

Micro-computed tomography (micro-CT) imaging is a valuable tool for different fields of research. In biology, it is especially suitable for bone research because of X-ray absorption in mineralized tissues. Due to this feature, diverse questions regarding bone development¹, metabolism², and evolution^3,4, among other topics, have been approached with the assistance of micro-CT. In 2008, de Crespigny et al.⁵ showed that micro-CT images of adult mouse and rabbit brains could be obtained using iodine as a contrast agent. This work opened a new application for this imaging technique, since iodine allowed the acquisition of images from soft tissues which otherwise would be insensitive to X-rays. Thus, the general goal of combining micro-CT and an iodine-based contrast agent is to obtain high-resolution images, in which soft tissues can be distinguished and identified at a meso or macro anatomical level.

This technique has notable potential for studies that require detailed ex vivo phenotypic characterization of small specimens, such as mouse embryos, which are widely used in experimental designs⁶. Iodine contrast in combination with micro-CT imaging has been used to obtain volumetric quantifications of organs⁷ and landmark three-dimensional (3D) structures^8,9. In recent years, micro-CT scanning of stained samples has been applied to describe brain phenotypic features of rodents¹⁰, and different improvements to the technique have been proposed. For adult brains, a protocol of 48 h of immersion in iodine, with a previous step of perfusion with a hydrogel, was found to produce images of high quality¹¹. Gignac et al.¹² expanded the limits of this technique by showing that rat brains stained with iodine could be processed to perform routine histological techniques. Similarly, these procedures demonstrate promising results for embryonic and pre-weaning rodent brains^8,13,14,15.

Although neuroscience has largely applied microscope-based techniques to assess different structural and functional aspects of brain development, such studies are more suitable for characterizing specific cell populations or spatially limited structures. Conversely, micro-CT imaging allows the description of whole structures and the acquisition of 3D models that preserve relevant spatial information, which is complementary to microscopic techniques. Magnetic resonance imaging (MRI) is also a standard technique applied to explore the structural features of small animals^16,17,18. However, micro-CT, with the use of a contrast agent, has two main advantages for ex vivo fixed samples: micro-CT scanners are largely less expensive and easy to operate, and allow a higher spatial resolution than MRI¹².

This work aims to describe the procedure to obtain high-resolution images from neonatal mouse brains using micro-CT scanning after staining with Lugol's solution, an iodine-based contrast agent. A comprehensive protocol is presented, which starts with preliminary stages such as sample collection and fixation of tissues, and goes through staining, micro-CT image acquisition, and standard processing. Image processing includes the segmentation of a 3D volume of the complete head, as well as of the brain, and the selection of specific anatomical planes to digitize point coordinates that could be then used in morphometric analyses. Although the focus here is the neonatal mouse brain, similar strategies can be applied to other soft tissues. Thus, the protocol presented here is flexible enough to be applied, with subtle modifications, to other types of samples.

Protocol

All experimental procedures followed the guidelines of the Canadian Council on Animal Care.

1. Sample collection and preparation

1. Prepare 500 mL of 4% paraformaldehyde (PFA).
   1. Under an extraction flux in a cabinet, add 20 g of PFA powder to 250 mL of 1x phosphate-buffered saline (PBS) in a 1 L glass beaker. Place the beaker with a magnet on a magnetic stirring plate.
   2. Stir while heating. With a thermometer, constantly check the temperature of the solution to keep it under 60 °C.
   3. With a plastic Pasteur pipette, add drops of 1 M NaOH to the solution to dissolve the PFA. Once the PFA powder has completely dissolved, cool the solution at room temperature (RT).
   4. Adjust the volume to 500 mL with 1x PBS. Adjust the pH to 7.2-7.3 by adding drops of 1 M HCl with a plastic Pasteur pipette.
   5. Using a paper filter and a glass funnel, filter the solution in a 1 L glass bottle.
   6. Close the glass bottle, remove it from the flux cabinet, and store in a refrigerator at -4 °C.
2. Collect brain samples.
   1. To obtain samples, sacrifice mouse neonates the morning following their birth, at 0.5 days of age (P 0.5), using surgical scissors to decapitate them.
   2. Place the scissors in the neck of the neonate while holding the body, and perform a single and clean cut.
3. Fixate the samples in PFA.
   1. Fill a 50 mL conical tube with 40 mL of 4% PFA.
   2. Immerse each head in the tube containing PFA as a fixation agent. For this action, use a Paton spatula or a spoon.
   3. Store the conical tubes in the refrigerator at -4 °C for 48 h, then change to 40 mL of 1x PBS with 0.1% sodium azide as a preservative. Maintain refrigerated at -4 °C.

2. Sample staining

1. Prepare Lugol's solution (3.75% w/v).
   1. Under an extraction flux in a cabinet, in a 1 L glass beaker, dissolve 10 g of Kl and 5 g of I₂ in 400 mL of distilled water, with constant mixing using a magnetic stirrer.
2. Immerse each head in a 50 mL conical tube containing 40 mL of Lugol's solution.
   1. For neonates of an approximate body size of 1.3 g, leave the samples immersed for 15-20 h with constant mixing using an orbital shaker.
   2. Place the samples in a 15 mL conical tube containing 10 mL of 1% agarose for 30 min before scanning.

3. Micro-CT scanning

NOTE: Micro-CT scanning requires specific equipment. There are different options for this kind of scanner, and details on acquisition depend on the characteristics of the used equipment. Dr. Hallgrímsson's lab counts on some alternatives of micro-CT scanners. Here, the protocol is based on the use of a basic desktop scanner, which is among the more accessible small animal imaging machines used by labs around the world. If the scanning objective is the skull, skip the steps in section 2 of the protocol and proceed to section 3. After scanning the skull, the staining process and the scanning of the brain could be performed to obtain images from the same specimens of both the brain and the skull.

1. Place the sample in the micro-CT holder.
   1. Place the sample in a 50 mL conical tube. Fix the sample to avoid movements during the scanner session. For this purpose, some polyurethane material can be used to fill the empty parts of the tube.
   2. Open the equipment and place the tube with the sample in the micro-CT holder.
   3. Log into the micro-CT control panel and register the sample to be scanned. This will automatically generate a sample number. In addition to the number, fill out the "Name" field. Both the sample name and number should be recorded elsewhere (e.g., in a spreadsheet) in case the raw data needs to be retrieved again.
   4. Start the scanning program. Enter the sample name or number and select a Control file, which should contain the scanning parameters.
2. Scan the sample.
   1. In the screen of the micro-CT software, set the following parameters: isotropic voxel size of 0.012 mm, 45 kVp, 177 lA, 800,000 ms integration time, and 500 projections per 180°.
   2. Set the scanning region by selecting Scout View. Once the reference image appears, press Reference Line to set the region. Move the solid green line to the edge of the sample, then hold Shift while dragging the cursor to extend the scan region to the other edge of the sample.
   3. Select OK to begin the scan.
3. Evaluate and export the scan.
   1. Open up the micro-CT scanner Evaluation Program and click on Select Sample and Measurement. In the Filter field, type the sample name, select the sample, and select the measurement file.
   2. The file will load up as slices, the number of which depends on the initial scanning window. Click on the Load All button, which will load every slice and facilitate movement between slices later on.
   3. Select Tasks > Evaluation 3D to initialize a cropping box around the slices. A 3D-Evaluation prompt will appear with a number of fields, including Task/Evaluation, VOI (Volume of Interest), Start X, Y, Z, and Dimension X, Y, Z.
   4. Set the Task > Evaluation 3D field to the appropriate script, as a variety of evaluation tasks can be performed, such as reconstruction, segmentation, morphometry (e.g., bone mineral density), and file conversions or transfers. The Default Evaluation script can be used as a guide.
   5. After selecting the script, adjust the VOI box on the main image. Start on the first slice. Ensure that the box contains the anatomy to be evaluated. Move the VOI by right-clicking on one of the small white squares toward the box edge and resizing the VOI using the middle mouse button. Alternatively, adjust the Dimension fields numerically.
   6. Pass through the sample using the scroll bar underneath the image to ensure that all of the anatomies have been captured in the VOI box. Click on XY, XZ, or YZ near the top left of the software to change the orientation.
   7. Select Start Evaluation. Depending on the Task Evaluation script, this will generate a 3D .aim image file with a corresponding .txt header file. Additionally, a path can be specified in the script to File Transfer Protocol (FTP) the image to a particular location.

4. Image processing

1. Open the images in the image processing software.
   NOTE: Image processing can be carried out using various software. This protocol includes the basic steps in a commercial software, which is widely accepted in the area and has versatile tools for basic and advanced processing. In addition, there are different acceptable formats for reconstructed micro-CT images. This aspect usually depends on the equipment used for acquisition and its software. Here, image processing is carried out with .bmp images. The same procedure can be applied for .tiff images and other types. Since the acquisition is carried out in a scanner with .aim as the default format, a conversion from .aim to .bmp is first presented.
   1. To open .aim files, select File > Open Data and choose the file. After that, a window with the title File Format Selection will open. Select Raw Data.
   2. A window with the title Raw Data Parameters will open. On this window, fill in the parameters following the information on the header .txt file that is generated for each .aim file. For Data Type, choose 16-bit for header complete with the number after Image data starts at byte offset. Also, set the dimensions and voxel size using the information on the header .txt file. Once the parameters are complete, press OK.
   3. If conversion from .aim to .bmp is needed, in Main Panel > Project View, right-click on the .aim object and select the option Convert > Convert Image Type. On Properties, select 8-bit as the Output Type. Select Apply.
   4. Right-click on the new 8-bit object. Select Save as. Choose .bmp and save.
   5. To open the .bmp file, select File > Open Data and choose the file. After that, a window with the parameters of the image opens. Confirm the parameters and accept if correct.
   6. In Main Panel > Project View, a new object is created with the name of the .bmp files. Right-click on this object and select the option Ortho Slice. In Properties, select the slice number and the orientation plane to see.
2. Obtain 3D volumes of the complete head.
   1. In Main Panel > Project View, right-click on the image object. Select Interactive Thresholding option > Create. In Properties, change the values for Minimum RGB and Maximum RGB until the part of the image corresponding to the head of the animal is selected. Select Apply.
   2. In Main Panel > Project View, a thresholded object is created. Right-click on this object and, in Properties, select Isosurface. In Properties, choose a threshold to visualize the surface and click Apply. This threshold needs to be chosen empirically, then different values might need to be examined.
3. Digitizing point coordinates
   1. 3D point coordinates can be digitized both based on slices or extracted surfaces. For the first option, right-click in the image option and select Slice. In Translate, choose the desired slice in the axial plane. If the plane is not correctly placed, click Options > Rotate in Properties to establish the right orientation.
   2. To digitize landmarks, right-click in Main Panel > Project View and select Create Object > Points and Lines > Landmarks. Right-click the new Object Landmarks and select Landmark View. In Landmark View, modify the size of the point in Size. To add landmarks in Properties, use the option Add in Edit Mode.
      NOTE: For neonates, a set of landmarks and semilandmarks around the axial and sagittal curves of the brain was previously published¹⁴.
   3. First, choose the axial plane that crosses the most rostral point of the olfactory bulbs and the most caudal point of the cortex.
   4. Select the Object Landmarks and, in Properties, select Landmark Editor. Using the arrow, click on the place the point is to be digitized.
   5. In the proposed set of landmarks, 24 points are digitized in the selected axial plane. Digitize the first in the most caudal point of the brain at the midline.
   6. Digitize five points equally distributed around the curve (semilandmarks) until the next landmark at the intersection between the midbrain and the cortex.
   7. Then, digitize eight points (semilandmarks) around the cortex.
   8. Digitize a new landmark at the intersection between the cortex and the olfactory bulbs. Then, digitize four points (semilandmarks) around the olfactory bulbs.
   9. Digitize a new landmark at the most rostral point of the olfactory bulbs at the midline. Then, digitize two points (semilandmarks) in between both olfactory bulbs.
   10. Digitize a new landmark at the most caudal point of the olfactory bulbs at the midline.
   11. For sagittal points, establish a slice in the right side of the brain at a parasagittal plane where the brain is most caudally prominent.
   12. In the proposed set of landmarks, 33 points are digitized in the selected parasagittal plane. Digitize the first in the intersection between the diencephalon and the hindbrain. Then, digitize nine points (semilandmarks) in the ventral limit of the brain.
   13. Digitize a new landmark at the most rostral point of the olfactory bulb. Then, digitize three points (semilandmarks) around the olfactory bulbs.
   14. Digitize a new landmark in the intersection between the olfactory bulb and the cortex. Then, digitize three points (semilandmarks) around the olfactory bulbs.
   15. Digitize a new landmark in the intersection between the olfactory bulb and the cortex. Then, digitize seven points (semilandmarks) around the cortex.
   16. Digitize a new landmark in the intersection between the cortex and the midbrain. Then, digitize seven points (semilandmarks) around the cortex.
   17. Digitize a new landmark in the intersection between the cortex and the midbrain. Then, digitize six points (semilandmarks) around the midbrain.
   18. Digitize a new landmark in the intersection between the midbrain and the cerebellum. Then, digitize two points (semilandmarks) around the cerebellum.
   19. Digitize a new landmark in the intersection between the cerebellum and the hindbrain.
   20. Save the digitized points. Right-click on the Object Landmarks and choose the option Save Data.
4. Analyze the point coordinates.
   NOTE: Once the point coordinates are digitized for a set of specimens, basic analyses can be performed using geometric morphometrics tools to obtain size (centroid size) and shape variables (shape or Procrustes coordinates). Analyses are carried out in a free open software environment, which is especially suitable for statistical analyses.
   1. Using a notepad, build a file containing the coordinates of all the digitized specimens. For this purpose, follow the TPS format, as described in https://morphometrics.uk/MorphoJ_guide/frameset.htm?index.htm.
   2. Open the statistical software. Select File > Change Dir to choose the directory where the .tps file is saved.
   3. Select Packages > Install packages. Choose one Cran Mirror. Look for geomorph and install it.
   4. Load packages by typing in the console: library(geomorph) and library(Morpho).
   5. Load the dataset by typing in the console: dataset <- readland.tps(file="NAME_OF_FILE.tps", specID="ID").
   6. Perform generalized Procrustes analysis by typing in the console: GPA<- gpagen(dataset).
   7. Obtain the centroid size by typing in the console: CS<-GPA$Csize.
   8. Obtain Procrustes coordinates by typing in the console: ProcCoord<- GPA$coords.
   9. Plot the Procrustes coordinates and the mean shape by typing in the console: plotAllSpecimens(ProcCoord).
5. Segmentation of ROIs
   NOTE: To segment the brain and obtain a 3D reconstruction, manually identify the brain tissues in each slice. The same procedure can be applied to specific ROIs, such as the olfactory bulbs, cortex, etc.
   1. Right-click on Main Panel > Project View and select Create Object > Label Field.
   2. Selecting the Label Field object, press the Segmentation Editor option in Properties.
   3. In Materials, add a new material and, if desired, rename it as Brain.
   4. In Selection, choose the option Current Slice to segment the tissue corresponding to the brain in each slice.
   5. Using the options present in Tools, select the brain tissue in each slice. For this case, the Magic Wand and the Brush are the most suitable.
   6. Once the selection is made in each slice, press Add in Selection.
   7. When selection in all slices is finished, return to Project View in Main Panel and right-click on Label Field to select Generate Surface, then click Apply.
   8. Once the surface is obtained, export it by applying Extract Surface.
   9. To obtain the volume of the segmented structures, in Main Panel > Project View, select the object that contains the label, right-click, and select Measure and Analyze > Volume Fraction. Then, click Apply.

Representative Results

Here, a basic protocol to obtain high-resolution images of neonatal mouse brains is presented. Heads were scanned after immersion in Lugol's solution. Despite their small size, the main brain anatomical structures, such as the olfactory bulbs, cortex, midbrain, cerebellum, and hindbrain, can be distinguished (Figure 1).

Different analyses can be carried out using these images as inputs. A set of landmarks and semilandmarks in two different anatomical planes were digitized. As can be observed in Figure 2, points are placed all along the border of the brain at the chosen planes. Once digitization is finished, the coordinates of the points can be read in a .txt file (Figure 3). The purpose of point digitization was to obtain raw coordinates that can be further used in geometric morphometric analyses. The presented protocol is focused on a single specimen, and morphometric analyses are suitable for larger samples, but the results obtained using the same set of points are presented elsewhere¹⁴.

Finally, the result of manual segmentation of the entire brain is a 3D model of the brain (Figure 4). In these volumes, diverse approaches can be taken, from simply estimating their area and volume to extracting surface points, to performing analyses that are similar to the one presented here for curve points.

Figure 1: Representative slices of a micro-CT image of a neonate's head. (A-C) Axial, (D) coronal, and (E) parasagittal planes are presented. Regions and structures that can be distinguished in the parasagittal plane: (a) olfactory bulbs, (b) cortex, (c) midbrain, (d) cerebellum, (e) hindbrain. Please click here to view a larger version of this figure.

Figure 2: Landmarks and semilandmarks digitized in the proposed protocol. (A) Axial and (B) parasagittal planes are presented. Numbers indicate landmarks. Please click here to view a larger version of this figure.

Figure 3: Example of a file containing point coordinates as exported from the software. The main items are indicated. Please click here to view a larger version of this figure.

Figure 4: 3D reconstruction of a neonate's brain after manual segmentation. An example of an entire reconstructed brain is presented in two views. Please click here to view a larger version of this figure.

Discussion

In this work, a concise protocol to scan neonatal brain tissues of mice using micro-CT with a contrast agent is introduced. In addition, it includes simple procedures to obtain quantitative and qualitative outputs. Building on these methods, further alternative or complementary analyses can be performed.

As shown in the protocol, micro-CT images can be analyzed in different ways. In previous studies, our group estimated the size and shape variation in perinatal brains of mice by digitizing coordinates of points and applying geometric morphometric techniques^8,14. The digitization of coordinates is also useful for obtaining linear measurements, using a simple Euclidean distance estimation. This, together with the volumes derived from brain and ROI segmentations, are valuable inputs for traditional morphometric assessment.

Although the protocol presented here is easily applied to ex vivo small samples, such as the ones we used, it is not applicable for in vivo studies, as fixation, and especially Lugol's solution as a contrast agent, is not viable. Micro-CT imaging is used in vivo, usually for hard tissue examination¹⁹, but the administration of contrast agents that reach the brain crossing the hematoencephalic barrier in living neonates is not simple. Thus, the technique presented here is still a promising tool to answer different questions of morphological variation.

Since pioneering studies showed that iodine is a good choice for vertebrate embryos²⁰, its application has expanded rapidly, and it became evident that one of the main drawbacks of this technique is the shrinking produced by the contrast agent¹². It is a ubiquitous problem in studies using this approach, and users should consider it when reproducing the protocol presented here, since some amount of shrinking is expected. Diverse strategies have been proposed to reduce this effect; Vickerton et al.²¹ examined different concentrations of iodine and concluded that shrinking depends directly on concentration, and reducing it largely helps to restrain deformation due to this effect. Another point to adjust is the proper time of immersion, which depends on the size and the hardness of the structures. After some tests, we found that 15-20 h is a sufficient period to stain the brain tissues of mouse neonates and preserve the main morphological properties. Some authors have suggested the use of some solutions before staining, such as agarose⁷ and hydrogel^11,13, or the application of a buffered Lugol's solution that prevents acidification and, consequently, shrinkage²². All these procedures could be added to the protocol presented here, which is a general and basic approach to micro-CT imaging with iodine contrast.

The use of micro-CT scanners for imaging soft tissues in larger samples, such as adult mice, has some drawbacks. An immersion procedure for staining is not recommendable because the distribution of the contrast agent will probably be heterogeneous. It will be important to explore some alternative options, like administering Lugol's solution via cardiac perfusion.

In conclusion, micro-CT with the addition of Lugol's solution is a suitable alternative for imaging perinatal brains of small animals, in particular mice. It is a simple way to obtain high-resolution images that, complemented with histology, can provide a consistent characterization of brain morphological variation.

Materials

Name	Company	Catalog Number	Comments
µCT 35	Scanco Medical AG		Note that Scanco does not offer the  µCT 35 anymore. Their smallest scanner is now the  µCT 45
Avizo	Visualization Sciences Group, VSG
C57BL/6 Mice	Bioterio Facultad de Ciencias Veterinarias Universidad Nacional de La Plata
Conical tubes	Daigger	CH-CI4610-1856
Flux cabinet	Esco	AC2-458
Glass beaker	Glassco	GL-229.202.10
Glass bottle	Simax	CFB017
Glass funnel	HDA	VI1108
HCl	Carlo Erba	403872	Manipulate under a flux cabinet and use personal protective equipment (mask, glass and gloves)
I2	Cicarelli	804211	When preparing I2KI, manipulate under a flux cabinet and use personal protective equipment (mask, glass and gloves)
KI	Cicarelli	PA131542.1210	When preparing I2KI, manipulate under a flux cabinet and use personal protective equipment (mask, glass and gloves)
Magnetic stirring	Arcano	4925
NaOH	Cicarelli	1580110	Manipulate under a flux cabinet and use personal protective equipment (mask, glass and gloves)
Orbital shaker	Biomint	BM021
Paraformaldehyde	Biopack	2000959400	Manipulate under a flux cabinet and use personal protective equipment (mask, glass and gloves)
Paton spatula	Glassco	GL-377.303.01
PBS	Biopack	2000988800
Plastic Pasteur pipette	Daigger	9153
R	R Project		The package geomorph for R was used in the protocol (https://cran.r-project.org/web/packages/geomorph/index.html)
Scissors	Belmed
Sodium azide	Biopack	2000163500
Thermometer	Daigger	7650