Evidence of the Primary Afferent Tracts Undergoing Neurodegeneration in Horses With Equine Degenerative Myeloencephalopathy Based on Calretinin Immunohistochemical Localization

Abstract

Equine degenerative myeloencephalopathy (EDM) is characterized by a symmetric general proprioceptive ataxia in young horses, and is likely underdiagnosed for 2 reasons: first, clinical signs overlap those of cervical vertebral compressive myelopathy; second, histologic lesions—including axonal spheroids in specific tracts of the somatosensory and motor systems—may be subtle. The purpose of this study was (1) to utilize immunohistochemical (IHC) markers to trace axons in the spinocuneocerebellar, dorsal column–medial lemniscal, and dorsospinocerebellar tracts in healthy horses and (2) to determine the IHC staining characteristics of the neurons and degenerated axons along the somatosensory tracts in EDM-affected horses. Examination of brain, spinal cord, and nerves was performed on 2 age-matched control horses, 3 EDM-affected horses, and 2 age-matched disease-control horses via IHC for calbindin, vesicular glutamate transporter 2, parvalbumin, calretinin, glutamic acid decarboxylase, and glial fibrillary acidic protein. Primary afferent axons of the spinocuneocerebellar, dorsal column–medial lemniscal, and dorsospinocerebellar tracts were successfully traced with calretinin. Calretinin-positive cell bodies were identified in a subset of neurons in the dorsal root ganglia, suggesting that calretinin IHC could be used to trace axonal projections from these cell bodies. Calretinin-immunoreactive spheroids were present in EDM-affected horses within the nuclei cuneatus medialis, cuneatus lateralis, and thoracicus. Neurons within those nuclei were calretinin negative. Cell bodies of degenerated axons in EDM-affected horses are likely located in the dorsal root ganglia. These findings support the role of sensory axonal degeneration in the pathogenesis of EDM and provide a method to highlight tracts with axonal spheroids to aid in the diagnosis of this neurodegenerative disease.

Equine degenerative myeloencephalopathy (EDM) and cervical vertebral compressive myelopathy (CVCM) share clinical signs of a general proprioceptive symmetric ataxia, abnormal resting basewide stance, and proprioceptive deficits of all limbs. Clinical signs of EDM usually develop by 6 to 12 months of age, and there is no sex predilection.^1,10 There are no ante-mortem diagnostic tests for EDM, and because it is not readily distinguishable from CVCM in the field, a clinical diagnosis often goes unrecognized. Definitive diagnosis of EDM requires postmortem examination shortly after death to avoid the early autolysis that precludes diagnosis of disorders in equine central nervous system tissues. EDM is considered an advanced form of neuroaxonal dystrophy (NAD).¹ Histologically, horses are diagnosed with NAD if dystrophic axons are confined to specific nuclei within the gray matter, whereas a diagnosis of EDM is used when the lesions also include axonal loss and demyelination of specific tracts in the white matter of the spinal cord.¹ At postmortem, EDM was rated the second-most common cause (24% of cases) of equine spinal cord disease at Cornell University in 1978¹¹ and ranked second in the causes of spinal ataxia at the University of Montreal from 1985 to 1988.¹²

The pathophysiology of EDM is not completely understood, but there is evidence for a genetic susceptibility.^1,4 The primary neural tracts that demonstrate axonal loss, demyelination, and astrogliosis in EDM-affected horses are sensory and include the spinocuneocerebellar, dorsal column–medial lemniscal, and dorsal spinocerebellar tracts.¹ In other species, the primary afferent neuronal cell bodies for these tracts reside in the dorsal root ganglia (DRG),⁷ and it has been presumed that this is the same in horses; however, this has not been definitively determined.

Immunohistochemical (IHC) stains can be utilized to identify different types of neuronal populations within the central nervous system and therefore could be of great value in tracing the axons undergoing degeneration to their somata in EDM-affected horses.¹³ The purpose of the present study was to (1) identify the specific IHC staining pattern of spinocuneocerebellar, dorsal column–medial lemniscal, and dorsal spinocerebellar nuclei and tracts and (2) trace the origin of the specific neurons and axons that are degenerating in EDM-affected horses. Calcium-binding proteins (calbindin [also motoric], calretinin, parvalbumin, S-100) are present in neurons of sensory tracts,¹³ and these markers were therefore selected for evaluation. In the rat spinal cord, axonal fibers immunoreactive for calretinin and parvalbumin are found in high densities in the dorsal and dorsolateral funiculi, while calbindin-immunoreactive fibers are present in the highest density in the lateral spinal nucleus.¹⁵ Additional markers were selected on the basis of previous studies characterizing abnormalities of axonal transport (synaptophysin, glial fibrillary acidic protein [GFAP], ubiquitin, and phosphorylated and nonphosphorylated neurofilament)¹⁷ or in an effort to identify potential neurotransmitters within affected tracts (vesicular glutamate transporter 2 [vGLUT2], glutamic acid decarboxylase [GAD]). Through the use of specific IHC stains to highlight the neural tracts affected in EDM, we aimed to gain insight into the disease pathogenesis and provide an additional tool to aid in the histologic diagnosis of EDM.

Materials and Methods

Animals

Nervous tissue samples were collected prospectively to obtain high-quality samples with adequate fixation for IHC. In total, 7 horses were evaluated, including 3 EDM-affected horses (a Morgan colt at 90 days of age [EDM No. 1], a shire gelding at 1.5 years of age [EDM No. 2], and a Thoroughbred filly at 1 year of age [EDM No. 3]), 2 control horses (a 5-month-old Percheron filly [C No. 4] euthanized for hermaphroditism and a 3-week-old quarter horse colt [C No. 5]) euthanized for a diffuse refractory skin condition, and 2 diseased positive controls (a 1.5-year-old Tennessee walking horse filly with CVCM [DC No. 6] and a 1-year-old Pony of the Americas colt with syringomyelia [DC No. 7]). All horses underwent a complete neurologic evaluation by one of the authors (C.J.F.) as previously described¹ using a well-established grading scale.⁹ Details regarding these horses and their neurologic evaluation are provided in Supplemental Table 1. A symmetric general proprioceptive ataxia was observed in EDM Nos. 1–3 (EDM No. 2; Supplemental Video 1) and DC No. 6. An inconsistent menace response was noted in EDM Nos. 1 and 3, as previously described.¹ A low serum α-tocopherol concentration was found in EDM Nos. 1 and 2 and DC No. 6 (Supplemental Table 1). Neurologic examination yielded normal findings for C Nos. 4 and 5 and DC No. 7. DC No. 7 was originally donated as a potential control horse, and the neurologic examination was unremarkable. However, a cyst within the spinal cord was found at the level of T8 on necropsy, and the horse was subsequently included in the diseased positive control category.

Horses were euthanized on separate days by intravenous administration of a barbiturate and immediately transported to the necropsy facility. Tissue collection was completed within 2 hours of euthanasia. The study was approved by the University of Minnesota Institutional Animal Care and Use Committee (No. 1206A15941).

Neuropathology

Gross abnormalities of the vertebral column, brain, spinal cord, and peripheral nerves were evaluated, and the entire brain and spinal cord from cervical vertebra 1 to lumbar vertebra 1 were removed and placed in 10% buffered formalin within 2 hours of euthanasia. Nerve roots, ganglia, and peripheral nerves were dissected and preserved intact. Central and peripheral nervous tissues were fixed in formalin for a minimum of 7 days, and all paraffin-embedded sections were stained with hematoxylin and eosin.

Areas of the brain that were examined included telencephalon (coronal sections at 2 levels), basal nuclei (coronal sections at 3 levels), thalamus (coronal sections at 2 levels), cerebellum (sagittal section at 1 level and horizontal section at 2 levels), mesencephalon (coronal sections at 3 levels), and brainstem (coronal sections at 8 levels). Cervical (C1, C2, C5, C6, C7), thoracic (T2, T10, T13), lumbar (L1, L3, L4, L5, L6), and sacral (3 levels) spinal cord segments were evaluated. To complete the neuropathologic assessment, peripheral nerves were evaluated at the level of the trigeminal ganglia (2 sections), sensory and motor nerve roots, and DRG (all spinal cord sections). A diagnosis of EDM was based on the presence of >10 dystrophic axons bilaterally within the nucleus cuneatus lateralis (known in humans as the nucleus cuneatus accessories), as well as the presence of dystrophic axons within the nucleus cuneatus medialis, nucleus gracilis, and nucleus thoracicus, in addition to evidence of axonal loss and demyelination of the dorsal spinocerebellar tracts and ventromedial tracts as previously described.^1,6 A diagnosis of CVCM was defined as gross evidence during necropsy examination of cervical spinal cord compression, vertebral column subluxation, articular process osteophytosis, or vertebral canal stenosis (alone or in combination), with microscopic evidence conforming to typical patterns of focal compression, including secondary, ascending, and descending white matter tract neuronal degeneration, as described by Mayhew et al.¹¹ A diagnosis of syringomyelia was based on gross and histologic evidence of a fluid-filled cavity within the spinal cord.

IHC Staining

Preliminary Evaluation

To identify neuronal populations and trace the tracts containing axonal degeneration, sequential regions of the central nervous system were evaluated in the youngest control horse (C No. 5) with a panel of IHC stains. Successful neuroanatomic tracing identified somata and associated axonal fibers with the same IHC staining properties along multiple sections of the particular tract within the central nervous system. Focus was placed on 3 main somatosensory tracts that are potentially affected in EDM; the spinocuneocerebellar, dorsal column–medial lemniscal, and dorsal spinocerebellar tracts (Fig. 1). The IHC panels used monoclonal and polyclonal antibodies against several neuronal and astrocytic proteins—including GFAP, phosphorylated neurofilament, nonphosphorylated neurofilament, synaptophysin, calbindin K28, calretinin, vGLUT2, GAD65 and GAD67, ubiquitin, S-100, and parvalbumin (Supplemental Table 2). Immunohistochemistry was completed with an automated slide stainer (Dako, Carpenteria, CA) or manual benchtop staining, and a peroxidase-labeled polymer conjugate system (Dako) was used as a secondary antibody. In brief, 4-μm-thick sections were deparaffinized and rehydrated in a decreasing graded alcohol series. Antigens were unmasked by the heat-induced epitope retrieval method via a Biocare Decloaking Chamber (Biocare Medical, Concord, CA) and a retrieval buffer of pH 6.0 or 9.0. Endogenous peroxidase was blocked with 3% H₂O₂ for 15 minutes. Nonspecific binding sites were blocked with normal goat serum, 1:10 in tris buffered saline, for 15 minutes. Slides were incubated with the primary antibody (Supplemental Table 2). Thereafter, sections were incubated with the horseradish peroxidase–conjugated secondary antibody. Immunoreactivity was detected with 3-amino-9-ethylcarbazole+ for 5 to 15 minutes. Slides were lightly counterstained with Mayer’s hematoxylin for 5 minutes.

Figure 1.

Diagram of spinocuneocerebellar tract (pink), dorsal column–medial lemniscal tract (blue; forelimbs only depicted), and dorsal spinocerebellar tract (green). The locations of degenerative axons associated with equine degenerative myeloencephalopathy are denoted in red. The neuronal cell body of the primary afferent of the spinocuneocerebellar tract resides in the dorsal root ganglia (DRG), and the axon traverses through the dorsal funiculus at the level of C1 to C7 to synapse on the nucleus cuneatus lateralis. The secondary afferent synapses in the ipsilateral cerebellar cortex. Within the forelimbs, the neuronal cell body of the primary afferent of the dorsal column–medial lemniscal tract resides in the DRG, and the axon traverses through the fasciculus cuneatus at the level of C1 to C7 to synapse on the nucleus cuneatus medialis. The secondary afferent synapses on contralateral thalamic nuclei. The neuronal cell body of the primary afferent of the dorsal spinocerebellar tract resides in the DRG, and the axon traverses through the dorsal horn to synapse on the nucleus thoracicus at the level of T1-L1. The secondary afferent synapses in the ipsilateral cerebellar cortex. CF, fasciculus cuneatus; GF, fasciculus gracilis; GN, gracilis nucleus; LCN, nucleus cuneatus lateralis; MCN, nucleus cuneatus medialis; TN, nucleus thoracicus. Figures 2–6. Calretinin-positive pathways, control horse No. 5, immunohistochemistry for calretinin. Figure 2. Dorsal root ganglia at C7. Both calretinin-negative and calretinin-positive somata are present. The calretinin-positive somata are likely associated with the spinocuneocerebellar or dorsal column–medial lemniscal tracts. Figure 3. Fasciculus cuneatus at C7. Both calretinin-negative and a subpopulation of calretinin-positive axons (arrow) are present. Figure 4. Nucleus cuneatus lateralis. Calretinin-positive axons synapse on calretinin-negative soma (thin arrow). A few spheroids (thick arrows) representing mild axonal degeneration are present. Figure 5. Dorsal root ganglia at T11. Both calretinin-positive (thin arrow) and calretinin-negative (arrowhead) myelinated axons are present, as well as calretinin-positive and calretinin-negative somata. Figure 6. Thoracic nucleus at T11. Calretinin-positive axons synapse on calretinin-negative soma (thin arrow). A few axonal spheroids are present (thick arrows).

Within 1 control horse (C No. 5), for each particular region evaluated, IHC stains were selected that would differentiate specific axons and neurons or identify astrocytic proteins within that section. Sections of the cerebellar vermis were stained with calbindin and GFAP; sections of the inferior cerebellar peduncle (including the nuclei vestibularis lateralis and cochlearis) and the caudal medulla at the level of the obex were stained with calbindin, vGLUT2, parvalbumin, calretinin, GAD, GFAP, ubiquitin, phosphorylated neurofilament, nonphosphorylated neurofilament, S-100, and synaptophysin. Three sections of the spinal cord—including DRG (1 trimmed at C7, 1 between T10 and T15, and 1 between L4 and L6)—were stained with parvalbumin, calretinin, GFAP, vGLUT2, and GAD. The IHC staining characteristics from this 1 control horse were used to trace the neural tracts implicated in EDM to identify IHC stains for evaluation of all other horses.

All Horses

Based on the preliminary IHC evaluation, sections of the cerebellar vermis were stained with calbindin and GFAP; sections of the inferior cerebellar peduncle were stained with calbindin, vGLUT2, parvalbumin, calretinin, GAD, and GFAP; and spinal cord sections at C7, T10–T15, and L4–6 were stained with parvalbumin, calretinin, and GFAP in the remaining 6 horses (EDM No. 3, C No. 4, and DC Nos. 6 and 7).

Results

Gross and Histopathology

There were no apparent gross abnormalities of the vertebral column, brain, spinal cord, and peripheral nerves of the EDM-affected horses or 2 negative control horses. The CVCM case (DC No. 6) demonstrated a narrowing of the spinal canal at the level of C1–2, with gross evidence of spinal cord compression at that site. A fluid-filled cavity within the spinal cord was found at the level of T8 in DC No. 7.

In EDM Nos. 1–3, an increased number (>30 bilaterally) of axonal spheroids were identified in the nucleus cuneatus lateralis with associated vacuolation and gliosis. Axonal spheroids were also identified in the nucleus thoracicus of the thoracic spinal cord, with axonal loss and demyelination along the dorsal spinocerebellar and ventromedial tracts of the thoracic spinal cord, most prominent at the level of T8–T15. Axonal spheroids (10–20 bilaterally) were also found within the nucleus vestibularis lateralis, nucleus cuneatus medialis, and nucleus gracilis of these horses. A few axonal spheroids (<10 bilaterally) were present in the nucleus cuneatus lateralis of C Nos. 4 and 5; however, no axonal spheroids were observed in any other nuclei in these horses.

IHC Tracing

Preliminary Evaluation

As the primary neuroanatomic tracts that contain degenerate axons in EDM-affected horses include the spinocuneocerebellar, dorsal column–medial lemniscal, and dorsal spinocerebellar tracts,¹ these particular tracts were traced in a 3-week-old control horse (C No. 5) via IHC antibodies. An initial evaluation of ubiquitin, phosphorylated neurofilament, nonphosphorylated neurofilament, S-100, and synaptophysin revealed nonspecific staining of the targeted regions. Sections of the spinal cord at C7, T12, and L6 in C No. 5 stained with vGLUT2 and GAD did not demonstrate adequate differentiation of specific axons within the spinal cord and were therefore not applied across spinal cord sections from the other horses. Preliminary evaluation of calbindin, vGLUT2, parvalbumin, calretinin, GAD, and GFAP revealed staining of specific axons or nuclei of interest in equine nervous tissue and were therefore evaluated in the 6 remaining horses (EDM Nos. 1–3, C No. 4, and DC Nos. 6 and 7; Table 1). Evaluation of the other control horse (C No. 4) was similar to that of C No. 5.

Table 1.

Immunohistochemical Characteristics of Selected Regions of the Nervous System in the 2 Normal Control Horses (Nos. C4 and C5).^a

Region	Calbindin	vGLUT2	Parvalbumin	Calretinin	GAD	GFAP
Cerebellar vermis: anterior frontal	+ Purkinje cell bodies, axons	N/A	N/A	N/A	N/A	(+)
Cerebellar peduncle
Lateral vestibular nucleus	+ axons	(+) cell bodies	+ axons	(+) axons	+ axons	++
Cochlear nucleus	−	−	+ axons	(+) axons	−	+
Medulla oblongata at obex
Lateral cuneate nucleus	−	+ cell bodies	+ axons, medial	+ axons, lateral	+ axons	(+)
Medial cuneate nucleus	−	+ cell bodies	−	+ axons	+ axons	−
Gracilis nucleus	−	+ cell bodies	−	+ axons	+ axons	−
Inferior olivary nucleus	−	+ cell bodies	−	+ cell bodies	+ axons	−
Reticular formation	−	+ cell bodies	−	+ cell bodies	+ axons	+
Spinal cord at C7
Dorsal funiculus	N/A	−	−	+ axons	−	−
Dorsolateral funiculus	N/A	−	−	(+) axons	−	−
Ventrolateral funiculus	N/A	−	−	−	−	−
Ventromedial funiculus	N/A	−	−	−	−	−
Dorsal horn	N/A	+ cell bodies	−	−	−	−
Intermediate column	N/A	−	+ cell bodies	+ cell bodies	−	−
Ventral horn	N/A	+ axons	−	−	−	(+)
Dorsal root ganglion	N/A	+ cell bodies	−	+ cell bodies	+ cell bodies	−
Spinal cord at T10–T15
Dorsal funiculus	N/A	−	−	+ axons	−	−
Dorsolateral funiculus	N/A	−	−	(+) axons	−	−
Ventrolateral funiculus	N/A	−	−	−	−	−
Ventromedial funiculus	N/A	−	−	−	−	−
Dorsal horn	N/A	+ cell bodies	−	−	+ axons	−
Nucleus thoracicus	N/A	−	+ axons	+ axons	(+) axons	−
Intermediate column	N/A	−	+ cell bodies	+ cell bodies	(+) axons	−
Ventral horn	N/A	+ axons	−	−	(+) axons	(+)
Dorsal root ganglion	N/A	+ cell bodies	−	+ cell bodies	+ cell bodies	−
Spinal cord at L4–L6
Dorsal funiculus	N/A	−	−	+ axons	−	−
Dorsolateral funiculus	N/A	−	−	(+) axons	−	−
Ventrolateral funiculus	N/A	−	−	−	−	−
Ventromedial funiculus	N/A	−	−	−	−	−
Dorsal horn	N/A	−	−	−	+ axons	−
Intermediate column	N/A	−	(+) cell bodies	+ cell bodies	−	−
Ventral horn	N/A	+ axons	−	−	−	(+)
Dorsal root ganglion	N/A	+ cell bodies	−	+ cell bodies	−	−

Abbreviations: (+), rarely found; +, mild density; −, not found; N/A, not examined.

^aImmunohistochemical markers included calretinin, calbindin, glutamic acid decarboxylase (GAD), vesicular glutamate transporter 2 (vGLUT2), parvalbumin, and glial fibrillary acidic protein (GFAP).

Spinocuneocerebellar Tract

A subpopulation of calretinin-positive cell bodies and axonal fibers in the DRG of the cervical spinal cord (Fig. 2) were examined through the cuneate funiculus at C7 (Fig. 3) to their synapse in the nucleus cuneatus lateralis. The neuronal cell bodies of the nucleus cuneatus lateralis were calretinin negative (Fig. 4).

Dorsal Column–Medial Lemniscal Tract

A subpopulation of calretinin-positive cell bodies and axons in the DRG of the cervical spinal cord were traced through the fasciculus cuneatus to their synapse in the nucleus cuneatus medialis. In addition, a subpopulation of calretinin-positive cell bodies and axonal fibers in the DRG of the lumbar spinal cord were traced through the dorsal fasciculus to their synapse in the nucleus gracilis.

Dorsal Spinocerebellar Tract

The primary afferent neurons (Fig. 5) could also be traced with calretinin. In the DRG of the thoracic and spinal cord, a subpopulation of calretinin-positive cell bodies and axons were traced to synapse within the nucleus thoracicus (Clarke column; Fig. 6). The neurons within the nucleus thoracicus were calretinin negative, and only a few axonal fibers within the dorsal spinocerebellar tract were immunoreactive for calretinin. Therefore, it appeared that only the primary afferent neuron for the dorsal spinocerebellar tract was immunoreactive for calretinin.

IHC Staining

EDM-Affected Horses

Calbindin, vGLUT2, parvalbumin, calretinin, GAD, and GFAP revealed successful staining of specific tracts of interest and were therefore evaluated in EDM cases (Table 2). Within the medulla oblongata, a high density of axonal spheroids within the nucleus cuneatus lateralis were diffusely immunoreactive for calretinin (Figs. 7, 8); a moderate density were immunoreactive for parvalbumin in the lateral, but not medial, portion of the nucleus; and a low density were immunoreactive for calbindin, GAD, and GFAP. A high density of axonal spheroids found in the nucleus thoracicus were immunoreactive for calretinin (Fig. 9), and a moderate density were immunoreactive for parvalbumin. Additionally, a high density of axonal spheroids in the lateral vestibular nuclei were immunoreactive for calbindin (Fig. 10) and GAD (Fig. 11), and a moderate density were immunoreactive for parvalbumin (Fig. 12).

Table 2.

Immunohistochemical Characteristics of Spheroids in Various Regions of the Nervous System.^a

Region	EDM No. 1	EDM No. 2	EDM No. 3	DC No. 6
Cerebellar vermis: anterior frontal	N	N	N	N
Cerebellar peduncle
Lateral vestibular nucleus	++ calbindin, GAD; + parvalbumin, GFAP	++ calbindin, GAD; + parvalbumin, GFAP	++ calbindin, GAD; + parvalbumin, GFAP	++ calbindin, GAD, GFAP; + parvalbumin
Cochlear nucleus	N	N	N	N
Medulla oblongata at obex
Lateral accessory cuneate nucleus	+++ calretinin (diffuse); + parvalbumin (lateral); (+) calbindin, GAD, GFAP	+++ calretinin (diffuse); + parvalbumin (lateral); (+) GAD	+++ calretinin (diffuse); + parvalbumin (lateral); (+) calbindin, GAD, GFAP	N
Medial cuneate nucleus	++ calretinin; + parvalbumin	+ calretinin	++ calretinin; + parvalbumin	N
Gracilis nucleus	N	N	+ calbindin	N
Inferior olivary nucleus	N	N	N	N
Reticular formation	N	N	N	N
Spinal cord at C7
Dorsal funiculus	N	N	N	N
Dorsolateral funiculus	(+) calretinin	N	N	N
Ventrolateral funiculus	N	N	N	N
Ventromedial funiculus	N	N	N	N
Dorsal horn	N	N	N	N
Intermediate column	N	N	N	N
Ventral horn	N	N	N	N
Ganglia	N	N	N	N
Spinal cord at T10–T15
Dorsal funiculus	+ calretinin	N	N	N
Dorsolateral funiculus	+++ GFAP	N	+++ GFAP	N
Ventrolateral funiculus	N	N	N	N
Ventromedial funiculus	N	N	N	N
Dorsal horn	N	N	N	N
Nucleus thoracicus	++ calretinin; + parvalbumin	++ calretinin	++ calretinin; + parvalbumin	N
Intermediate column	N	N	N	N
Ventral horn	N	N	N	N
Ganglia	N	N	N	N
Spinal cord at L4–L6
Dorsal funiculus	(+) parvalbumin	N	N	N
Dorsolateral funiculus	N	N	N	N
Ventrolateral funiculus	N	N	N	N
Ventromedial funiculus	N	N	N	N
Dorsal horn	N	N	N	N
Intermediate column	+ calretinin	N	N	N
Ventral horn	N	N	N	N
Ganglia	N	N	N	N

Abbreviations: (+), rarely found; +, low density; ++, medium density +++, high density; N, density similar to negative control horses (see Table 1).

^aIn the 3 horses affected with equine degenerative myeloencephalopathy (EDM) and 2 diseased positive control (DC) horses (cervical vertebral compressive myelopathy [No. 6] and syringomyelia [No. 7]). Note that for every region, DC No. 7 had density similar to that of negative control horses (see Table 1). Immunohistochemical markers included calretinin, calbindin, glutamic acid decarboxylase (GAD), vesicular glutamate transporter 2 (vGLUT2), parvalbumin, and glial fibrillary acidic protein (GFAP).

Figures 7–9.

Calretinin-positive pathways, equine degenerative myeloencephalopathy–affected horse No. 1, immunohistochemistry for calretinin. Figure 7. Nucleus cuneatus lateralis. Axonal degeneration represented by large numbers of calretinin-positive spheroids is evident within the nucleus. Figure 8. Nucleus cuneatus lateralis. Varying sizes of calretinin-positive spheroids are present within the nucleus. Figure 9. Nucleus thoracicus at T15. Calretinin-positive spheroids are evident within the nucleus thoracicus. Figures 10–13. Equine degenerative myeloencephalopathy–affected horse No. 3, nucleus vestibularis lateralis. Figure 10. Calbindin-positive axons (presumed to be Purkinje cell axons) contain calbindin-positive spheroids (arrow) at end-terminal synapses with calbindin-negative soma. Immunohistochemistry for calbindin. Figure 11. Presumptive glutamic acid decarboxylase (GAD)–positive Purkinje cell axons contain GAD-positive spheroids (arrow) at end-terminal synapses with GAD-negative soma. Immunohistochemistry for GAD. Figure 12. Parvalbumin-positive spheroids (arrow) are present in axons that synapse on parvalbumin-negative soma. Immunohistochemistry for parvalbumin. Figure 13. Increased glial fibrillary acidic protein staining representing astrogliosis. Immunohistochemistry for glial fibrillary acidic protein.

The region of the lateral vestibular nucleus of EDM Nos. 1–3 (Fig. 13) and DC No. 6 was found to be immunoreactive for GFAP as compared with negative control horses. This immunoreactivity corresponded to the regions containing a high abundance of axonal spheroids. Additionally, a high density of axons in the dorsolateral funiculus of EDM Nos. 1 and 3 were immunoreactive for GFAP as compared with negative control horses.

Disease Controls

A high density of axonal spheroids observed in the lateral vestibular nucleus of DC No. 6, the CVCM-affected horse (Table 2), were strongly immunoreactive for calbindin and GAD, and a moderate density were immunoreactive for parvalbumin. A small number of axonal spheroids (<10 bilaterally) were observed in the nucleus cuneatus lateralis of DC Nos. 6 and 7 in hematoxylin and eosin stains. These spheroids were not immunoreactive with any IHC markers assessed.

Discussion

In this study, we prospectively recruited horses with a high suspicion of NAD/EDM based on clinical and diagnostic evaluations. Three cases had histologic lesions classified as the more severe variant of the disease, EDM. EDM is more readily diagnosed in routine hematoxylin and eosin–stained sections than equine NAD because of the more extensive lesions.¹ Axonal spheroids within the nuclei cuneatus medialis and gracilis also may be an incidental finding in older horses without evidence of neurologic disease and have been considered a consequence of aging.⁶ These are typically single spheroids and often not associated with gliosis or vacuoles.^2,6 Despite these distinctions, it still may be difficult to distinguish this possible age-related axonal degeneration from pathologic NAD in older animals. To avoid this complication, we selected age-matched control horses and EDM-affected rather than NAD-affected cases for the present study. Although only a small number of horses were evaluated, we employed strict phenotypic criteria, and tissue samples were collected immediately following euthanasia, thereby providing high-quality fixed tissue for evaluation. Additionally, the evaluation of diseased positive control horses, as a means of determining the specificity of the IHC findings for EDM, was not performed in the previous 2 studies evaluating IHC markers in EDM.^17,19

Horses affected with NAD/EDM display a general proprioceptive ataxia and often abnormal posture.^1,5 Previous studies demonstrated that the primary neuroanatomic tracts affected in horses with EDM are the sensory tracts involved in conscious proprioception (dorsal column–medial lemniscal tract) and unconscious proprioception (spinocuneocerebellar [fore-limb] and dorsal spinocerebellar [hind limb]).^1,5,7 This is the first study to examine neural tracts affected with EDM and to confirm homologous neuroanatomic location of the neurons affected by this disease. The spinocerebellar tracts transmit information about the activity of all skeletal muscles from muscle proprioceptors (muscle spindles and Golgi tendon organs) to the cerebellar cortex and therefore are responsible for unconscious proprioception, affecting basic posture and simple gait activities. The dorsal spinocerebellar tract provides input from the hind limbs and trunk, while the spinocuneocerebellar tract provides input from the forelimbs. In other species,¹⁸ and as traced in the 3-week-old control horse in this study, the cell body of the primary afferent neuron resides in the DRG and its axonal process projects in the cuneate fascicle (spinocuneocerebellar tract) or dorsolateral horn (dorsal spinocerebellar tract). Within the spinocuneocerebellar tract, the axon ascends to the nucleus cuneatus lateralis, where the axon terminates in the ipsilateral caudal cerebellar peduncle. Within the dorsal spinocerebellar tract, the primary afferent axon synapses on the nucleus thoracicus at the base of the dorsal horn. The secondary afferent then ascends in the dorsolateral region of the lateral funiculus and enters the cerebellum by the caudal cerebellar peduncle. It is of particular interest then that the nucleus cuneatus lateralis of the spinocuneocerebellar tract across species is regarded as homologous to the thoracic nucleus of the dorsal spinocerebellar tract,⁷ as the primary afferent neurons of both these tracts are the most severely affected with EDM.¹ Deficits of these tracts in other species may manifest as abnormal posture and ataxia,¹⁸ clinical findings consistent with those observed in EDM-affected horses.¹

In the dorsal column–medial lemniscal tract, touch, pressure, and joint proprioception through low-threshold mechanoreceptors provide information necessary to perform complex motor activities. As defined in other species (cat, monkey, rat)¹⁸ and as traced in the control horse, the cell location of the primary afferent neuron in the dorsal column–medial lemniscal tract resides in the DRG; its axon enters the dorsal horn; and then a long collateral branch ascends in the dorsal funiculus and synapses in the nucleus gracilis (hind limbs) or nucleus cuneatus medialis (forelimbs). The nuclei gracilis or cuneatus medialis contains the cell body of the second afferent neuron of the dorsal column–medial lemniscal tract, and its axons continue in the medial lemniscus to synapse on a ventral group of thalamic nuclei, which project to the primary somatic sensory area of the cerebral cortex. Deficits of this tract in other species manifest as an inability to make temporal discriminations; clinical signs of stumbling or knuckling may be evident,¹⁸ similar to neurologic deficits observed in EDM-affected horses.⁵

In this study, antibodies to calretinin served as a marker to highlight the primary, but not secondary, afferent neurons of the dorsal column–medial lemniscal and dorsal spinocerebellar tracts in the horse. Calretinin staining of the spheroids identified with EDM is likely a reflection of the particular neural tracts affected by the disease and not necessarily associated with the underlying pathologic mechanism. Calretinin rapidly buffers calcium in neurons in addition to affecting intracellular calcium signals pre- and postsynaptically.³ It could be hypothesized that the oxidative injury associated with EDM¹⁹ could lead to calcium release from the endoplasmic reticulum, leading to increased cytosolic and mitochondrial calcium concentrations and subsequent neuronal death, as previously suggested for many neurodegenerative diseases in humans.⁸ However, in the rat, axonal fibers within the dorsal and dorsolateral funiculi and those originating from the nucleus thoracicus are often immunoreactive for calretinin.¹⁴ Within the medulla oblongata, the region of the lateral cuneate nucleus was not calretinin immunoreactive in chicks,¹⁶ yet we have demonstrated immunoreactive axons in this region in our healthy control horse. Thus, although it is possible that calretinin is involved in the underlying disease process of EDM, this is not a conclusion that can be reached from the results of the present study.

The IHC marker GFAP highlighted the affected dorsolateral funiculi in EDM-affected horses in 2 of 3 of our cases. A degree of GFAP immunoreactivity in 2 EDM-affected horses was observed in the medulla oblongata, including the region of the nucleus cuneatus medialis and nucleus gracilis as previously observed.¹⁷ However, strong GFAP immunoreactivity within the medulla oblongata of EDM-affected horses was not identified, as subjectively compared with negative and diseased positive control horses. Increased GFAP expression occurs in astrocytes secondary to many neurodegenerative conditions. In addition to Luxol-fast blue, GFAP may be used to highlight glial scarring due to the degeneration and loss of axons of the dorsal spinocerebellar tracts in EDM-affected horses.

Axonal spheroids strongly immunoreactive for calbindin and GAD and moderately immunoreactive for parvalbumin were noted in the lateral vestibular nucleus of the 3 EDM-affected horses and 1 CVCM-affected horse. Additionally, this region was strongly immunoreactive for GFAP in these horses. Based on the neuroanatomic location and IHC staining patterns, these spheroids most likely arise from the Purkinje cells of the cerebellum. This study demonstrated that these axonal spheroids in the lateral vestibular nucleus may be found in cases of EDM and CVCM; however, additional cases of CVCM are required to validate this finding.

Limitations of this study include the small sample size of EDM and diseased positive control horses. To formulate conclusions regarding IHC immunoreactivity patterns in horses with EDM versus CVCM, additional CVCM-affected horses need to be evaluated. Additionally, our focus was on the sensory tracts affected with EDM; therefore, we did not systematically evaluate the motor tracts that are often involved in EDM and CVCM. As our primary interest was the evaluation of sensory tracts, we did not actively recruit or evaluate horses with equine motor neuron disease for this study.

In conclusion, calretinin was a valuable anatomic marker to trace the spinocuneocerebellar, dorsal column–medial lemniscal, and dorsal spinocerebellar tracts in horses and to identify axonal spheroids within the nucleus cuneatus lateralis and medialis and nucleus thoracicus of EDM-affected equine cases. The primary afferent neuronal cell bodies for these tracts reside in the DRG in the horse, similar to other species. Degenerate axons immunoreactive for calbindin, GAD, and parvalbumin may be found in the lateral vestibular nucleus of EDM- and CVCM-affected horses.