A Micro-CT-based Method for Characterizing Lesions and Locating Electrodes in Small Animal Brains

Summary

This article describes a straightforward method to prepare small animal brains for micro-CT imaging, in which lesions can be quantified and electrodes located with high precision in the context of the whole brain.

Abstract

Lesion and electrode location verification are traditionally done via histological examination of stained brain slices, a time-consuming procedure that requires manual estimation. Here, we describe a simple, straightforward method for quantifying lesions and locating electrodes in the brain that is less laborious and yields more detailed results. Whole brains are stained with osmium tetroxide, embedded in resin, and imaged with a micro-CT scanner. The scans result in 3D digital volumes of the brains with resolutions and virtual section thicknesses dependent on the sample size (12–15 and 5–6 µm per voxel for rat and zebra finch brains, respectively). Surface and deep lesions can be characterized, and single tetrodes, tetrode arrays, electrolytic lesions, and silicon probes can also be localized. Free and proprietary software allows experimenters to examine the sample volume from any plane and segment the volume manually or automatically. Because this method generates whole brain volume, lesions and electrodes can be quantified to a much higher degree than in current methods, which will help standardize comparisons within and across studies.

Introduction

Neuroscientists have relied on lesions for a long time in order to understand the relationship between function and location in the brain. For example, our understanding of the hippocampus as being indispensable for learning and memory and of the prefrontal cortex as being key for impulse control were both products of serendipitous lesions in humans^1,2. The use of animal models, however, has allowed neuroscientists to harness the power of lesions by going beyond serendipity, and the function of countless brain areas has been elucidated through systematic studies of structure-function relationships through lesions^3,4.

To correctly assign function to a structure, however, lesion studies require precise quantification procedures, which is an area that has been lacking. The current gold standard for quantifying lesions is to section, mount, and image brains with a light microscope. The imaged slices are then matched to the closest sections on an atlas, and the approximate coordinates of the lesions across subjects are indirectly reported, often through the use of camera lucida images or example histological slices^{3,4,5,6,7,8,9,10}.

Beyond the imprecision of current lesion quantification procedures, these techniques are time-consuming and prone to failure. Small changes in brain stiffness, blade sharpness, and temperature can lead to botched, warped, or torn sections. Sections can also stain unevenly and be improperly imaged because of bubbles in the mounting medium. Importantly, upon sectioning, the three-dimensional context of the lesion's location in the brain is lost, making precise 3D reconstruction of the lesion in the brain challenging.

Another common application for lesions has been to determine the location of single and multiple electrode recordings in the brain. At the end of the final recording session, researchers induce small electrolytic lesions at the electrode tip and process the brain histologically as done in a conventional lesion experiment¹¹. This technique suffers from the same drawbacks described above, with additional problems being that the electrolytic lesions are usually larger than the electrodes used to make them but are usually small enough that they are challenging to find histologically. When multiple electrodes are inserted, as in the case of a tetrode array, verification through electrolytic lesions is even more challenging. An alternative to electrolytic lesions is the use of a dye on the electrode to later verify histologically¹², but this technique suffers from the same drawbacks that come with conventional histology.

Here, we describe in-depth a recently described method¹³ based on staining techniques in electron microscopy (EM) and X-ray computed tomography (micro-CT) that quantifies lesions and locates electrodes in small animal brains better than current methods. Micro-CT is an imaging technique in which X-rays are shot at a sample that is rotated 360° while a scintillator collects the X-rays not deflected by the sample. The result is a high-resolution digital 3D reconstruction of the sample that can be visualized in any orientation and quantified precisely. Many academic institutions have micro-CT scanners, which are also available commercially.

Protocol

All care and experimental manipulation of animals were reviewed and approved by the Harvard Institutional Animal Care and Use Committee. The perfusion described here is specific for rats, but the procedure is applicable to any animals with smaller or similarly sized brains.

1. Perfusion

1. Prepare 1x phosphate-buffered saline (PBS). For a rat (age: 0.5–1.5 years old, weight: 250–600 g), 800–1,000 mL should be sufficient. Use 400 mL to perfuse the animal and an additional 400 mL to dilute the fixative.
2. Prepare the fixative consisting of 2% (w/v) paraformaldehyde (PFA) and 2.5% (w/v) glutaraldehyde (GA) in 1x PBS. For a rat, 400 mL is sufficient. Save 50 mL in a 50 mL conical tube for post-fixation of the brain after perfusion.
3. Anesthetize the rat with 4–5% isoflurane gas (at 0.8 L/min O₂ at 1.0 bar at 21 °C) for 15 min.
4. Inject a lethal dose of sodium pentobarbital intraperitoneally (180 mg/kg).
5. Test for reflex loss in the animal by conducting a toe-pinch reflex exam. Wait to begin the perfusion until the animal has lost its reflex response.
6. Follow the previously described intracardial perfusion and brain extraction protocol¹⁴ with the following solutions: perfuse the animal with 400 mL of 1x PBS at 125 mm Hg to remove the blood. Once all blood has been removed and replaced with 1x PBS, begin perfusing with 400 mL of a solution of 2% PFA and 2.5% GA dissolved in 1x PBS at 125 mm Hg.
   NOTE: If the procedure is being conducted in an animal that is not a rat, the only important components of the perfusion procedure are the use of 1x PBS and 2% PFA, 2.5% GA in 1x PBS. The solution volumes and perfusion pressure may be adjusted according to the species.

2. Post-fixation

1. Place the extracted brain in 2% PFA, 2.5% GA in 1x PBS solution (same solution used to perfuse the animal previously). Ensure that the solution volume is at least 10x the volume of the brain. For rats, place the brain in a 50 mL conical tube with 50 mL of solution. Store the sample in post-fixation for 2–3 days, shaking lightly at 4 °C.
   NOTE: If the sample is in a 50 mL conical tube, placing it horizontally on an orbital shaker will ensure the best results.
2. After the sample has been post-fixed for long enough, wash the sample in double-distilled water (ddH₂O), de-ionized water (diH₂O), or ultrapure water (see Table of Materials) four times for the following durations: 1, 1, 1, and 15 min.
   NOTE: For this protocol, ddH₂O, diH₂O, or ultrapure water should be interchangeable. For simplicity, ddH₂O will be used to refer to purified water henceforth.

3. Staining

CAUTION: For this step, conduct all solution preparations under a hood while using gloves.

1. Prepare at least 10x the brain volume of staining solution. For rat brains, prepare 50 mL of 2% (w/v) osmium tetroxide (OsO₄) in ddH₂O by combining 25 mL of stock 4% OsO₄ solution and 25 mL of ddH₂O.
   CAUTION: Osmium tetroxide is volatile and may cause temporary blindness and respiratory problems if not handled appropriately. Discard all materials that contact the osmium tetroxide in an appropriate chemical hazard container within a secondary container.
2. Place the brain in a new 50 mL conical tube and add the OsO₄ solution. The brain should begin to turn brown as OsO₄ reacts with lipids in the tissue.
3. Close the tube and seal it thoroughly with paraffin film (see Table of Materials) to ensure that it does not leak during incubation.
   NOTE: The osmium is volatile and will react mildly with the plastic of the tube, so sealing the tube properly is very important. The tube may be wrapped with aluminum foil for added protection.
4. Store the sealed tube at 4 °C, shaking lightly on an orbital shaker at 50 rpm for 2 weeks. Place the tube horizontally to ensure the best mixing. Ensure that the sample is fully submerged in the solution while shaking.
   NOTE: If the osmium is not allowed to circulate continuously, it may not fully penetrate the sample, so horizontal placement on the shaker is very important.

4. Embedding

1. After the sample has been incubated in OsO₄ for 2 weeks, wash it with ddH₂O 5 times at room temperature (RT) for the following durations: 1, 1, 1, 15, and 60 min to remove all the unbound OsO₄ in the sample.
   NOTE: The multiple exchanges, including the last 60 min exchange, are necessary to allow all the osmium in the circulatory system to diffuse out.
2. Wash the sample with ddH₂O for 30 min at 4 °C.
   ​NOTE: Osmium may continue to diffuse out of the sample, but its quantity should be greatly reduced from the previous step.
3. Dehydrate the sample with ethanol to eventually infiltrate it with resin. To dehydrate, replace the ddH₂O with 10–20 mL of the following ethanol dilutions (for 30 min each at 4 °C): 20% (v/v) ethanol and 80% (v/v) ddH₂O; 50% (v/v) ethanol and 50% (v/v) ddH₂O; 70% (v/v) ethanol and 30% (v/v) ddH₂O; 90% (v/v) ethanol and 10% (v/v) ddH₂O; 100% ethanol.
4. Prepare the acetone/resin dilutions as follows.
   1. Prepare 100 mL of the resin for embedding (see Table of Materials) as per the manufacturer's instructions.
   2. To make the 33% (v/v) resin-67% acetone solution, pour 15 mL of resin into a 50 mL conical tube and add 30 mL of 100% glass-distilled acetone.
   3. To make the 50% (v/v) resin-50% acetone, pour 22.5 mL of resin into a 50 mL conical tube and add 22.5 mL of 100% glass-distilled acetone.
   4. To make the 67% (v/v) resin-33% acetone solution, pour 30 mL of resin into a 50 mL conical tube and add 15 mL of 100% glass-distilled acetone.
   5. Use the remaining 32.5 mL of resin as the first solution of 100% resin below.
5. Begin the resin infiltration process by moving the sample through 10-20 mL of the following acetone and acetone/resin dilutions: 100% acetone for 30 min at 4 °C; 100% acetone for 30 min at 4 °C; and 100% acetone for 30 min at room temperature (RT).
   NOTE: The rest of the infiltration process will take place at RT.
6. Immerse the sample in 33% (v/v) resin 67% acetone for 3 h at RT, then 50% (v/v) resin-50% acetone for 3 h at RT, 67% (v/v) resin 33% acetone for 3 h at RT, and 100% resin for 12 h at RT.
7. Make a fresh 50 mL batch of resin following the instructions on the bottle. Transfer the sample to the container in which it will be cured (e.g., the disposable molds described in Table of Materials). Infuse the sample with fresh 100% resin for 4 hours at RT.
   NOTE: If fresh resin is not used, the resin made the previous day will begin to harden prematurely, and the sample will be hard to manipulate.
8. Degas the sample in a vacuum oven for 15 min at 45 °C.
   NOTE: This step will help remove any trapped air bubbles within the sample, but it is non-essential and will not affect quality of the data.
9. Finally, cure the sample in an oven for 48 h at 60 °C.

5. micro-CT

1. Once the sample has been cured, peel off the disposable mold and scan it with the micro-CT machine.
   NOTE: Depending on the machine used, the settings will be different. For the scanner used by the authors listed in the Table of Materials, the recommended settings are 130 kV, 135 µA with a 0.1 mm copper filter and a molybdenum source, a 1-second exposure, and an average of 4 frames per projection. However, experimenters should calibrate the scanner individually, as many factors will affect optimal settings during a particular session.
2. Once the scan is completed, reconstruct the sample with the recommended parameters for the experimenter's scanner/software combination on a computer with the scanner software.
3. Finally, visualize and analyze the reconstructed digital volume using the experimenter's software of choice (see Table of Materials for examples).

Results

Traditionally, brains are sectioned and stained in order to quantify lesions and locate electrodes, but this method is error-prone, labor-intensive, and typically requires estimation of the results. By preparing whole brains for micro-CT imaging, the probability of damaging the samples is greatly reduced, features of interest may be analyzed in the context of the entire brain, and the method lends itself to parallel processing of many samples, considerably speeding up sample preparation.<...

Discussion

The following are critical steps to the protocol: first, the use of a combination of PFA and GA to perfuse the animal and subsequently post-fix the brain was paramount to achieving consistent full osmium penetration of the tissue. Although we did not test this explicitly, a plausible explanation is that PFA fixation is reversible¹⁵, whereas GA fixation is not reversible^16,17. Because a two-week incubation in osmium tetroxide is required fo...

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
Paraformaldehyde (PFA)	Electron Microscopy Sciences (EMS)	15710	2% (w/v/) in 1X PBS
Glutaraldehyde (GA)	EMS	16220	2.5% (w/v) GA in 1x PBS
OsO4	EMS	19190	Work in fume hood
Ethanol	Decon Labs	Koptec	140, 190, 200 proof
Acetone	EMS	10015	Glass-distilled
Durcupan ACM resin	Sigma-Aldrich	44610	A, B, C and D components, resin for embedding
Disposable molds	Ted Pella	27114	Suggested
milliQ water (ultrapure water)	Millipore Sigma	QGARD00R1 (or related purifier)	Suggested
Parafilm (paraffin film)	Millipore Sigma	P7793	Suggested paraffin film
Micro-CT scanner	Nikon Metrology Ltd., Tring, UK	X-Tek HMS ST 225	Used by authors
Software for visualizing and analyzing micro-CT scans:
	Volume Graphics	VG Studio Max	Used by authors
	FEI / Thermo Scientific	Avizo	Used by authors
	FEI / Thermo Scientific	Amira	Similar to Avizo
	Mark Sutton & Russell Garwood	Spiers	Free, http://spiers-software.org/
	Pixmeo Sarl	Osirix Lite	Free, https://www.osirix-viewer.com/
	Open Source	FIJI	Free, https://fiji.sc/
	Adobe	Photoshop	Good for analyzing one slice at a time