7 Tesla MRI Followed by Histological 3D Reconstructions in Whole-Brain Specimens

Abstract

Post mortem magnetic resonance imaging (MRI) studies on the human brain are of great interest for the validation of in vivo MRI. It facilitates a link between functional and anatomical information available from MRI in vivo and neuroanatomical knowledge available from histology/immunocytochemistry. However, linking in vivo and post mortem MRI to microscopy techniques poses substantial challenges. Fixation artifacts and tissue deformation of extracted brains, as well as co registration of 2D histology to 3D MRI volumes complicate direct comparison between modalities. Moreover, post mortem brain tissue does not have the same physical properties as in vivo tissue, and therefore MRI approaches need to be adjusted accordingly. Here, we present a pipeline in which whole-brain human post mortem in situ MRI is combined with subsequent tissue processing of the whole human brain, providing a 3-dimensional reconstruction via blockface imaging. To this end, we adapted tissue processing procedures to allow both post mortem MRI and subsequent histological and immunocytochemical processing. For MRI, tissue was packed in a susceptibility matched solution, tailored to fit the dimensions of the MRI coil. Additionally, MRI sequence parameters were adjusted to accommodate T1 and T2^∗ shortening, and scan time was extended, thereby benefiting the signal-to-noise-ratio that can be achieved using extensive averaging without motion artifacts. After MRI, the brain was extracted from the skull and subsequently cut while performing optimized blockface imaging, thereby allowing three-dimensional reconstructions. Tissues were processed for Nissl and silver staining, and co-registered with the blockface images. The combination of these techniques allows direct comparisons across modalities.

Keywords: formalin fixation; histology; post mortem human brain; ultra-high field MRI; whole brain imaging.

FIGURE 1

Representation of the tissue processing. Tissue (A) is subjected to MRI scanning, after which autopsy is performed (note that the brain is still covered by the meninges in his image) (B). After blockface imaging (C), sections are stained for Nissl or silver (D). MRI contrast (E) and blockface reconstructions (F), and section reconstructions (G) are aligned in the same space.

FIGURE 2

Quantitative maps of MR parameters and MRI images of the basal ganglia of specimen #8 (axial view). Contrasts are proton density (A), longitudinal relaxation rate R₁ (B), MTw (C), T₁w (D), PDw (E), and OLS R${}_2^{*}$ (F). Asterisks in F indicate air bubbles.

FIGURE 3

Surface reconstruction of specimen #8 based on blockface images created in MIPAV. Surface of the (A) left and (B) right hemisphere, (C) dorsal surface, (D) ventral surface, (E) occipital lobe and cerebellum. Note the artifacts caused by incomplete coverage in OCT in the frontal pole. The line along the side of the entire right hemisphere is the reconstruction of a cut resulting from the autopsy procedure.

FIGURE 4

Blockface image of specimen #8 (A) and the corresponding Nissl (B) stained sections.

FIGURE 5

Examples of Nissl staining in specimen #8. Coronal section (A), Subthalamic nucleus (B), Cingulate cortex (C), high power magnification of the cingulate cortex (D). Scale bar represents 2 cm (A), 50 μm (B,D), 250 μm (C).

FIGURE 6

Examples of Silver staining in specimen #8. Coronal sections (A,D), High power magnifications of the ventral putamen (B,C), and the inferior rostral gyrus (E). Scale bar represents 2 cm (A,D), 50 μm (B,C,E).

FIGURE 7

Examples of immunohistochemical staining in specimen #8. Coronal section stained for parvalbumin (A), High power magnification of parvalbumin staining of layer III of the paracentral lobule (B), Coronal section stained for Calretinin (C), High power insert of layer III of the cingulate cortex (D). Calretinin staining of the subthalamic nucleus (E,F). Scale bar represents 2 cm (A,C), 50 μm (B,D,F), 200 μm (E). Squares in (A,C) indicate the area in which (B,D) were sampled.

FIGURE 8

(A) Realignments/reconstructions of the blockface images of specimen #8 (top panel), Nissl stained sections (middle panels), and silver stained sections (lower panels), (B) Additional views of silver staining at different levels. Stained sections have a slice gap of 1.2 mm. Note that the coronal plane represents the cutting plane, axial and sagittal sections were realigned without any registration steps. Nissl and silver sections were registered to the corresponding blockface images. For registration purposes color images were transformed to gray scales of saturation, thereby inverting the appearance of the staining. Note that in areas in which small cortical pieces are detached as a result of cutting in the coronal plane, some tissue loss at the outer surface of the brain is observed.

FIGURE 9

(A) Registration of a Nissl section (lighter contrast) to the blockface image (darker contrast). (B) Registration of a Silver stained section (lighter contrast) to the blockface image (darker contrast).