MRI Findings in a Rottweiler with Leukoencephalomyelopathy

Introduction

Leukoencephalomyelopathy (LEM) in the rottweiler is a rare degenerative disorder that has been recognized since the early 1980s.^1–3 The clinical syndrome manifests as a long-strided gait with the appearance of stiffness and overreaching as the limbs are advanced when walking, consistent with general proprioceptive ataxia and upper motor neuron tetraparesis that begins between 1.5 and 3.5 yr of age.³ Affected dogs initially develop abnormalities in the thoracic limbs, and the thoracic limbs are often more severely affected than the pelvic limbs. Gender predilection has not been reported. Despite a familial relationship among affected dogs, a mode of inheritance has not been determined.^1–3 The etiology of LEM remains unknown. Progression of clinical signs occurs over months to up to a year, with most dogs being euthanized due to increased difficulty ambulating.² Currently, the antemortem diagnosis is based on exclusion of other diseases that result in similar clinical signs, and diagnosis often entails advanced imaging. To the authors’ knowledge, this is the first report to describe the MRI characteristics of LEM in the rottweiler. In the present case, the topography of the lesions observed with MRI exactly paralleled the pathologic changes observed on gross and microscopic examination of the central nervous system (CNS), which consisted of severe myelin loss, decreased axonal numbers, astrogliosis, and astrocytosis. Consequently, MRI of the nervous system may aid clinicians in establishing a presumptive antemortem diagnosis of LEM in rottweilers.

Case Report

A 22 mo old male rottweiler presented to the Veterinary Teaching Hospital, University of Georgia with a 1 mo progressive history of an abnormal gait. The dog was current on vaccinations and was receiving heartworm preventative monthly. There was no history of other medical problems. Physical examination was normal. On neurologic examination, the dog displayed a longer than normal stride that was characterized by stiffness and overreaching, consistent with general proprioceptive ataxia and upper motor neuron paresis in all four limbs. The gait was characterized by hypermetria resulting in marked overreaching in the thoracic limbs. The dog scuffed the nails on all four feet while walking. Deficits in postural reactions (proprioceptive placing and hopping) were observed in all four limbs. Deficits were worse in the thoracic limbs. Additionally, the left thoracic and pelvic limbs were more affected than the right thoracic and pelvic limbs. Spinal reflexes were normal in the thoracic limbs. In the pelvic limbs, the withdrawal reflexes were normal and patella reflexes were increased bilaterally. Muscle mass and muscular tone were both normal in all four limbs. Cranial nerve examination was normal. There was normal range of motion of the neck. The dog did not appear painful with either manipulation or palpation along the entire vertebral column. Neuroanatomic diagnosis was consistent with a lesion affecting the first through fifth cervical spinal cord segments. Differential diagnoses included cervical vertebra(e) malformation/malarticulation, intervertebral disc disease, fibrotic stenosis, subarachnoid diverticula, neoplasia, LEM, and neuroaxonal dystrophy (NAD).

Hematologic and serum biochemical examinations revealed lymphocytosis (3.3 × 10³/µL; reference range, 0 0.4–2.9 × 10³/µL) and eosinophilia (1.53 × 10³/µL; reference range, 0–1.3 × 10³/µL). Urinalysis was normal. Under anesthesia, MRI of the vertebral column from the first cervical vertebra to the third thoracic vertebra and the brain was performed using a 3.0T MR unit^a and a multichannel phase array spine coil. The following pulse sequences were obtained: T1-weighted fluid-attenuated inversion recovery (T1W FLAIR), T2-weighted (T2W), T2-weighted FLAIR (T2W FLAIR), and T2*-weighted gradient echo (T2*W) images. Additionally, axial and sagittal plane T1W FLAIR images of the vertebral column and the brain were obtained after IV administration (0.1 mmol/kg) of contrast agent^b.

In comparison with unaffected white matter of the spinal cord, bilaterally symmetric intra-axial hyperintensities on T2W images were noted in the white matter of the dorsolateral funiculi from the cervicomedullary junction extending contiguously to the level of the sixth to seventh cervical intervertebral disc (Figures 1A, B). In comparison with the unaffected areas of brainstem, bilaterally symmetric intra-axial hyperintensities on T2W images were noted in the pyramids (Figure 1C) and ventral aspect of the crus cerebri of the brain. The lesions also were hyperintense on T2*W and T2W FLAIR images and were isointense on T1W FLAIR images. Abnormal contrast enhancement was not observed in the spinal cord and brain.

View Figure

• Download as Powerpoint
• Download Figure

Cytology and protein analysis of cerebrospinal fluid obtained from the cerebellomedullary cistern were normal. Based on the signalment, history, the neurologic examination, and abnormalities observed on MRI, a presumptive diagnosis of LEM was made; however, infiltrative disease, such as neoplasia or infectious/noninfectious inflammatory myelitis, could not be excluded. Consequently, the dog was administered prednisone^c (0.5 mg/kg per os q 24 hr). After 10 days, no improvement was noted. Fifty days after initial presentation, the dog was presented for humane euthanasia. The owner stated that the dog could no longer walk without falling down; however, if allowed to either run or walk fast, the dog would not fall as often.

Prior to euthanasia and with owner consent, MRI of the cervical vertebral column and brain was performed to determine if the lesions had progressed. Imaging was performed as previously described with the same unit, with the exception that the brain was imaged with the dog in sternal recumbency using an extremity coil to provide improved image quality. The previously identified lesions were present, and new lesions were not identified. Subjectively, the lesions in the white matter of the dorsal aspect of the lateral funiculi were larger. The lesions in the brain were unchanged in size. Following MRI, the dog was euthanized and immediately necropsied.

At necropsy, gross lesions were restricted to the cervical spinal cord and brainstem. Opaque white, well-demarcated, bilaterally symmetric foci were noted in the white matter in the dorsal aspect of the lateral funiculi of the cervical spinal cord (Figures 1D, 2A) and the pyramidal tracts of the medulla oblongata. The crus cerebri were grossly normal. Brain, spinal cord, and representative tissue samples of internal organs were fixed in 10% neutral buffered formalin, processed, embedded in paraffin, sectioned at 5µm, and stained with hematoxylin and eosin. Selected sections of the brain and spinal cord were stained with luxol fast blue.

Immunohistochemistry was performed with monoclonal antibodies against neurofilament^d (1:8,0000), glial fibrillary acidic protein^e ([GFAP], 1:8,000), myelin basic protein^f ([MBP], 1:2,000), and canine distemper virus^g. Also samples from the affected dorsolateral region of the lateral funiculus of the first cervical spinal cord segment were removed at necropsy and fixed in 2% (para)formaldehyde and 2% glutaraldehyde in 0.1 M phosphate buffer for transmission electron microscopy. Following dehydration in graded alcohols, the tissues were embedded in epon-araldite. Thin sections were made and stained with lead citrate and uranyl acetate.

On low power magnification, the white matter of the affected areas observed on MRI showed a marked pallor consistent with loss of myelin (Figures 2B, C). Microscopically, lesions were confined to the white matter of the brain and spinal cord. The most severe lesions were located in the cervical spinal cord, but lesions extended caudally into the thoracic spinal cord and rostrally into the brainstem. Within the cervical spinal cord, the lesion affected the white matter in the dorsal portion of the lateral funiculi in the area of the dorsal spinocerebellar, lateral corticospinal, reticulospinal, and rubrospinal tracts. A subpial rim of normal white matter was always preserved except at the first cervical spinal cord segment where the lesion extended to the pia. In the brain, the pyramidal tracts, crus cerebri, area of medial lemniscus, caudal cerebellar peduncle, trapezoid body, area of the spinal tract of the trigeminal nerve, and area of the optic tracts were most severely affected. The white matter of the cerebellar folia was also multifocally affected. The parenchymal portions of the oculomotor nerves in the mesencephalon were mildly affected.

View Figure

• Download as Powerpoint
• Download Figure

Within the affected white matter, myelin and axon loss was observed. The myelin and axonal loss was replaced by prominent astrogliosis and astrocytosis with numerous GFAP-positive astrocytic processes and gemistocytic astrocytes in the affected regions. Many normal appearing axons were present within the area of severe demyelination. The degree of myelin loss exceeded the loss of axons. Degenerative axonal changes were relatively mild, digestion chambers were rare, and few axonal spheroids were seen. In affected areas, vessels were mildly thickened with prominent endothelial cells and increased cellularity of the perivascular space (Figure 2D).

To better assess myelin content and axonal changes, banked tissue samples of the cervical spinal cord from an age matched control dog free of neurologic disease were used for comparison. In the affected dog, immunohistochemistry for MBP confirmed partial to total loss of myelin in the white matter of affected areas with patchy staining of the remaining white matter and naked axons compared to the control dog. (Figures 3A, B) Immunohistochemistry for neurofilament demonstrated decreased numbers of axons. (Figures 3C, D) Multiple axons had irregular profiles and many were smaller. Only a few axons were larger than in the control dog. In the affected dog, immunohistochemistry for GFAP revealed the abnormal neuroparenchyma to consist mostly of gemistocytic astrocytes as well as astrocytic processes (Figures 3E, F) Immunohistochemistry for canine distemper virus was negative.

View Figure

• Download as Powerpoint
• Download Figure

Ultrastructurally, affected spinal cord white matter contained oligodendroglia, a small number of small naked axons, small numbers of myelinated axons, and hypertrophied astrocytic processes (Figure 4A). Myelinated axons were some of the larger axons present and were typically irregularly shaped. The myelin sheaths had extensive splitting of the lamellae and lack of compaction. Schwann cells with similarly myelinated axons were present near vessels indicating Schwann cell remyelination (Figure 4B).

View Figure

• Download as Powerpoint
• Download Figure

Discussion

The clinical and pathologic features of the dog described in this report are consistent with a diagnosis of LEM as previously described in rottweilers.^1,3 A definitive diagnosis of LEM requires histopathology. In this case, MRI precisely identified the affected areas, allowing representative transverse sections to be obtained for macroscopic, histopathologic, and ultrastructural examination. In the case reported here, bilaterally symmetric, continuous hyperintensities on T2W images were identified in the same topography as reported for the most severe lesions in LEM in rottweilers.^1–3 Moreover, the lesions observed on MRI in the present case correlated with the degenerative changes in the white matter observed histologically, which included severe loss of myelin and astrogliosis. Those findings were similar to findings previously described in rottweilers.³ Further, microscopic investigation in the present case using immunohistochemistry (i.e., antibodies against MBP, neurofilament, and GFAP) and electron microscopy corroborated severe myelin loss, astrogliosis/astrocytosis, as well as demonstrating decreased numbers and degeneration of axons. In addition, minimal abortive attempts of oligodendroglial and Schwann cell remyelination were present ultrastructurally. Interestingly, Schwann cells have not be observed in the spinal cord in previously reported cases of LEM.¹ Schwann cell invasion into the spinal cord can occur in a variety of conditions, including primary myelin disorders as well as focal compressive and concussive processes.⁴ Although typically excluded from the CNS, Schwann cell invasion occurs at transition zones where the peripheral nervous system interfaces with the CNS, such as the dorsal and ventral root entry zones and near blood vessels.⁵ Consistent with this, Schwann cells in the present case were observed near blood vessels.

Given the MRI characteristics and topography of the lesions observed in the dog reported here, many of the differential diagnoses for clinical signs referable to the first through fifth cervical spinal cord segments can be eliminated from consideration. Other important degenerative nervous system disorders may affect rottweilers. Those degenerative diseases must be considered in the differential diagnoses. NAD is a neurodegenerative disease that presents in young rottweilers as a chronic, progressive, general proprioceptive ataxia and upper motor neuron paresis of all four limbs.^6,7 Pathologically, NAD results in axonal swellings (spheroids) throughout the CNS gray matter, except the cerebral cortex.⁶ The spheroids are primarily localized in the distal regions and axon terminals of afferent fibers entering sensory nuclei in the spinal cord, brainstem, and diencephalon.⁶ Cerebellar atrophy may also be appreciated in more chronic cases.^6,7 Although not described in rottweilers with NAD, the MRI findings have been documented in one papillon with NAD.⁸ At 3 mo of age, no abnormalities were detected; however, at 6 mo of age, diffuse atrophy of the cerebrum, cerebellum, and brainstem were seen in the affected dog.⁸ In humans with NAD, the most consistent MRI abnormality observed is a hyperintense cerebellum on T2W and T2 FLAIR images.⁹ To the authors’ knowledge, reports detailing imaging studies involving encephalomyelopathy and polyneuropathy in the rottweiler have not been published. Of the available MRI data for LEM and NAD in the dog, no similarities are seen. Encephalomyelopathy and polyneuropathy has also been reported in young rottweilers.¹⁰ Signs consist of ataxia and paresis involving all four limbs as well as laryngeal and pharyngeal dysfunction.¹⁰ Although initial signs share similarities with LEM, dogs with encephalomyelopathy and polyneuropathy are typically younger and, with disease progression, signs reflect neuromuscular dysfunction.¹⁰

MRI has been used in other white matter disorders. In spongy degeneration of the CNS in Labrador retrievers, MRI disclosed symmetrical, hyperintense lesions on T2W images, which correlated with degenerative white matter lesions detected during gross and histologic examination.¹¹ Recently, a novel LEM in two leonbergers was described.¹² As in the present case, the bilaterally symmetric hyperintensities on T2W images were observed in the dorsolateral funiculi of the spinal cord; however, lesions were restricted to the second cervical spinal cord segment in one dog and to the second to fourth cervical spinal cord segments in the other dog. Additionally, lesions were not noted on MRI of the brain in affected leonbergers. MRI of globoid cell leukodystrophy of the West Highland white terrier revealed bilaterally symmetrical, hyperintense lesions on T2W images, which included the corpus callosum, centrum semiovale, internal capsule, corona radiate, and cerebellar white matter. Symmetrical contrast enhancement was observed in the corpus callosum, internal capsule, and corona radiata on T1W postcontrast images.¹³ Demyelinating lesions have been found to correlate well with areas of hyperintensity on T2W images in canine distemper.¹⁴ Dysmyelinogenesis in the English springer spaniel and Portuguese water dog secondary to GM₁ gangliosidosis was correlated with a mild hyperintensity of the corona radiata on T2W images.¹⁵

The etiology of LEM remains obscure, and the pathogenesis still remains to be determined. Bilaterally symmetric degenerative lesions in the CNS are usually due to nutritional, metabolic, or toxic causes.^16,17 It is possible that LEM is a result of an inborn error of metabolism.³ In humans, numerous primary disorders of myelin occur secondary to inborn errors of metabolism.¹⁸ In two dogs with LEM, biochemical analysis of peripheral blood leukocytes for evidence of lysosomal storage defects was evaluated for activity of β-galactosidase, β-hexosaminidase, β-hexosaminidase A, aryl sulphatase A, acid phosphatase, β-glucuronidase, α-mannosidase, α-fucosidase, β-glucocerebrosidase, β-galactocerebrosidase, and sphingomyelinase). All evaluations were normal.²

Based on immunohistochemical and ultrastructural studies in those cases, the lesions are clearly demyelinating with simultaneous, although inadequate and abnormal, remyelination. The question remains whether the defect in LEM involves either a primary disorder of oligodendrocytes (primary demyelination) or loss of myelin secondary to primary changes in the axons (axonopathy). Oevermann et al. (2008) hypothesized that LEM in the leonberger was a consequence of an alteration in the intimate relationship between the oligodendrocyte and the neuron so myelin was produced, but not stable, and the topography of the lesion reflected the population of neurons affected.¹² In humans, there is a close relation between myelin and axon, where axonal pathology may precede demyelination.^19–21 Interestingly, in this case, not only was myelin loss observed, but also decreased axonal diameters and irregularity in axonal shape. That finding could be related to either loss or disorganization of the neurofilaments, and thus, impaired cytoskeletal organization and axonal transport in dogs with LEM. Minimal attempts at remyelination were seen in this case, some of which was due to Schwann cells migrating into the spinal cord, suggesting disruption of the glia limitans formed by astroglia, which, as a consequence of disruption, allowed Schwann cells to enter the CNS and promote remyelination.²² That could indicate that the Schwann cells were compensating for some form of oligodendrocyte/neuronal dysfunction that caused demyelination; however, there was also evidence of remyelination by oligodendroglia, albeit abnormal. Although speculative, given the more severe pathologic involvement of myelin than axons observed histologically and ultrastructurally in the case reported herein, the underlying disease process likely represents a leukodystrophy rather than an axonal disorder with secondary demyelination.

In the present case, MRI was helpful in establishing a presumptive antemortem diagnosis of LEM. Although unknown, it is possible that either early in the course of the disease or if clinical signs are mild, lesions may not be detected with MRI. The factors that contribute to lesion conspicuity likely include the degree of myelin loss, number of affected axons (and their diameter), and number of glial cells in the white matter. Additionally, the sophistication of the MRI unit also may play a role. In the present case, the lesions in the white matter of the dorsal aspect of the lateral funiculi were larger after 50 days. Therefore, a repeat MRI should be recommended in cases where there is a high index of suspicion yet lesions are not observed.

Conclusion

Understanding the correlation between lesion characteristics and topography on MRI and histologic findings can refine the diagnostic approach categorizing different morphologic changes observed in the CNS. The ability to establish an accurate presumptive antemortem diagnosis of LEM has many implications. Although not established, given the occurrence of LEM in a specific breed, the rottweiler, a hereditary basis is suspected. The ability to identify affected individuals and remove them from the breeding pool is imperative. Also, owners of affected dogs could be provided an accurate prognosis as rottweilers with LEM are typically euthanized within 1 yr of diagnosis.