Mouse Spinal Cord Anatomy

Working Through Rodent Spinal Cord Dissections

Most protocols start with perfusing the animal through the left ventricle using phosphate-buffered saline followed by 4% paraformaldehyde. If you skip the PBS step or rush it, the fixation will be uneven and your histology will look like garbage. I spent three weeks troubleshooting what I thought was a staining failure before realizing I hadn't given the saline enough time to clear the blood from the vascular bed. The rule of thumb is roughly 20 to 30 milliliters of PBS at a steady flow rate before you switch to the fixative. The adult mouse spinal cord runs from the foramen magnum down to approximately the L1 vertebra, though there is individual variation between strains. The cord itself sits inside the vertebral column and is protected by meninges—dura, arachnoid, and pia. Cross-sections reveal the characteristic butterfly-shaped gray matter surrounded by white matter tracts. The dorsal horns handle sensory input, the ventral horns contain motor neurons, and the lateral horns appear only in the thoracolumbar segments where sympathetic outflow originates. Gray matter ratio to total cross-sectional area is roughly 40 to 45 percent in adult C57BL/6 mice. One thing people consistently get wrong is the orientation of the spinal cord during sectioning. The cord is not a straight cylinder. It has natural curvature and the segmental levels shift as you move caudally. When I was first learning this, I would cut what I thought was a mid-thoracic section and end up with lumbar tissue because the cord had migrated slightly within the canal during fixation. The workaround is to pre-label the vertebral column after removal, counting vertebrae from the sacrum upward and marking the corresponding cord level with a fine marker before you embed anything.

The segmental nomenclature is another common tripwire. Standard atlases list 7 cervical, 13 thoracic, 7 lumbar, 5 sacral, and 1 coccygeal segments. But that doesn't mean each vertebra maps to one segment. Cervical segments run roughly parallel to their vertebral counterparts, but thoracic and lumbar spinal segments sit cranial to their vertebrae. By the time you reach the lumbar enlargement, the cord segments are positioned around T11 to L2 vertebral levels. If you're targeting the lumbar enlargement for studies involving hindlimb circuitry, you need to know this offset or you will miss the region entirely.

Sectioning and Processing

After fixation, the tissue needs to be post-fixed for another 12 to 24 hours in 4 percent PFA before you attempt any dissection. Rushing this step leaves the cord too soft to handle and you'll tear the meninges during extraction. Remove the vertebral column by carefully cutting along the midline of the dorsal arch with microsurgery scissors. Strip away the bone and ligaments while preserving the meningeal sheath. The cord should come free in one piece if the fixation is adequate. Cryoprotection is the next mandatory step. Transfer the cord into 30 percent sucrose dissolved in phosphate buffer until it sinks, which typically takes 24 to 48 hours at 4 degrees Celsius. The tissue will swell if you push this longer than necessary and the section quality drops. Once fully, embed the tissue in OCT compound and freeze on dry ice or in an isopentane bath pre-chilled to minus 40 degrees. For transverse sections, aim for 20 to 30 micrometer thickness on a freezing microtome or cryostat. Thinner sections give you better cellular resolution but they tear more easily. Thicker sections retain structural integrity but you lose detail in the deeper laminae of the gray matter. I ran into a persistent problem where my dorsal columns would delaminate from the ventral funiculi during section collection. It turned out the knife angle was too aggressive and the blade was slightly dull. Switching to a fresh blade set at a shallower angle and using adhesion-enhanced slides—charged slides rather than plain glass—solved it completely. The sections laid flat immediately and there was no further delamination. That single change saved me probably 15 to 20 slides that would have been wasted.

Staining and Identification

For basic cytoarchitecture, a simple Nissl stain with cresyl violet is more than adequate. It highlights neuronal cell bodies across all laminae and gives you a clear boundary between gray and white matter. Immunohistochemistry works well for specific marker identification—ChAT for cholinergic motor neurons, SMI-32 for large pyramidal and motor neurons, Pax2 for inhibitory interneurons in the ventral spinal cord. The mounting medium matters more than most protocols acknowledge. Fluoromount-G with DAPI gives decent preservation but the fluorescence fades noticeably after a few months. ProLong Gold or equivalent antifade reagents hold signal intensity much longer for archived sections you plan to revisit. The ventral horn motor neuron pools are where beginners waste the most time trying to identify specific populations. Motor neuron columns are organized somatotopically. Medial motor neuron pools innervate axial muscles while lateral pools target distal limb muscles. The lumbar enlargement contains the lateral motor neuron columns responsible for hindlimb control. If you're doing retrograde tracing, inject the label into the target muscle and allow 48 to 72 hours for transport. Then section the appropriate lumbar segments and count labeled cells in the correct lateral column. Skipping the transport time window is the fastest way to get zero or near-zero labeling and waste an entire animal. One counter-intuitive detail that rarely gets emphasized: the cervical enlargement is proportionally much larger in mice than what you see in rat atlases. The forelimb representation dominates the C5 through C8 segments and the gray matter expansion is significant. If you're comparing across species or pulling reference data from rat studies, the scale difference will throw off your expectations for what a normal cross-section should look like. Also, the central canal in adult mice is often collapsed or occluded, so don't assume you'll see a clear open lumen like you might in developmental stages. A faint ependymal line is normal and expected.

Pitfalls and Limitations

Sagittal sectioning can be tempting for visualizing the full rostrocaudal extent of a tract or nucleus, but the curvature of the cord makes true sagittal cuts nearly impossible without severe distortion. Transverse sections remain the standard for a reason. There are atlases you can reference—the Bregma-based coordinates work reasonably well for stereotactic injections near the spinal cord surface, but those coordinates don't translate directly to histological section identification. You need a dedicated spinal cord atlas like the one from the Mouse Brain Reference Atlas or the Workman et al. dataset for accurate level mapping. Perfusion quality is the single biggest variable affecting downstream results. Over-perfusion causes excessive shrinkage and hard tissue that cracks during sectioning. Under-perfusion leaves antigenic epitopes masked by unFIXED proteins, which means your IHC signals will be weak or absent. The compromise is straightforward but easily missed: perfuse with PBS until the effluent runs clear, then use PFA for a measured volume rather than an indefinite duration. Roughly 20 to 30 milliliters of 4 percent PFA is sufficient for most fixed-time protocols. If you're doing electrophysiology instead of histology, skip the PFA entirely and use cold artificial cerebrospinal fluid for acute slice preparation. One hard limitation worth stating plainly: whole-mount immunolabeling of the intact mouse spinal cord is possible but it takes roughly 5 to 7 days of incubation with shaking, and the penetration depth is limited to about 500 micrometers from the surface. Core regions remain unlabeled regardless of how long you extend the protocol. If you need deep-tissue antibody access, sectioning is still the only reliable approach. Confocal microscopy helps with optical sectioning after the fact but it doesn't solve the penetration problem during the labeling step itself.