Magnetic resonance imaging has become the primary tool for the ante mortem diagnosis of white matter disease in humans due to its high sensitivity for detecting changes in white matter. Decreased myelin and elevated water content is revealed by increased T1 and T2 relaxation times, with a consequent reduction in signal intensity in T1-weighted images and increased signal intensity in T2-weighted images
. Thus, the pattern of MRI changes is very helpful in defining disease because it reveals the distribution of histopathologic changes
At present, there are few case reports describing the use of MRI for the diagnosis of canine and feline neurodegenerative diseases. T2-weighted hyperintensities of white brain matter were evident in cats with GM2 gangliosidosis
[26, 27] and in a West Highland white terrier with globoid cell leukodystrophy
. Increased signal was also evident in T2-weighted images of the spinal cord of Leonberger dogs with leukoencephalomyelopathy
. Dogs with GM2 gangliosidosis displayed T2-weighted hyperintensities in the region of the caudate nucleus and atrophy of the cerebrum and cerebellum
[30, 31]. MRI of Papillon dogs with neuroaxonal dystrophy
 and Scottish Terriers with hereditary cerebellar degeneration demonstrated atrophy only and failed to detect changes in white matter
MRI of the cervical spine may be used to support the clinical diagnosis of LEM in Rottweilers. A similar MRI pattern has been described in Leonberger dogs with LEM
. Interestingly, however, brain lesions were not detected using MRI in the Leonberger dogs or in the case reported here, although histological analyses showed that the optic tracts and particularly the cerebellar medulla were significantly affected in both breeds
[2, 29]. It is possible that the white matter lesions in the brain were less advanced than those in the spinal cord at the time of imaging, and improved imaging protocols may be required for the visualisation of brain lesions. These protocols may include smaller slice thicknesses and the application of sequences other than conventional T1- and T2-weighted imaging, e.g., diffusion tensor imaging, magnetisation transfer imaging or MR spectroscopy
. It is also questionable whether the pathological changes in these regions have sufficiently altered the physics of the tissue to induce changes visible with a 1.5 T clinical scanner.
Many inherited white matter diseases and associated genetic defects have been described in humans
. Leukoencephalopathies may be characterised as lysosomal or peroxisomal disorders, mitochondrial disorders, methylation cycle disorders, organic acidaemias or amino acid disorders or as leukoencephalopathy associated with calcification, hypomyelination, abnormal lipid metabolism, vasculopathy or muscular dystrophy. Many distinct entities, e.g., Alexander’s disease, adult onset autosomal dominant leukoencephalopathy, vanishing white matter disease and adult polyglucosan encephalopathy, have also been recognised. A vast number of genetic defects are currently associated with these conditions, and many more remain to be elucidated; the molecular cause remains unknown in ~50% of affected humans
. The lesion distribution and MRI appearance of LBSL are considered unique and diagnostic in humans; consequently, only a single candidate gene was examined in the present study
Leukoencephalopathy with brain stem and spinal cord involvement is a rare, autosomal recessive disorder that typically manifests in childhood or adolescence. The diagnosis of LBSL in humans is based on clinical presentation and is characterised by a slowly progressive cerebellar ataxia, spasticity, dorsal column dysfunction and a highly characteristic pattern of abnormalities observed using MRI and spectroscopy. Typical MRI findings include a combination of high T2-weighted signal changes in the cerebral white matter accompanied by the selective involvement of the brain stem and spinal cord tracts (the entire length of the pyramidal tracts with the additional involvement of cerebellar connections and the intraparenchymal and mesencephalic parts of the trigeminal nerve)
[13, 37]. MR spectroscopy demonstrates an elevation in lactate levels in the abnormal white matter of almost all of the affected human patients. These findings led researchers to assume that the disease was a mitochondrial disorder, which was subsequently confirmed by the discovery of various mutations in the DARS2 gene, which encodes mitochondrial aspartyl-tRNA synthetase
[14, 38]. As demonstrated by multiple case reports of LBSL in humans, normal CSF and blood lactate concentrations, as were noted in the case reported herein, do not exclude a mitochondrial disorder as the underlying cause of leukoencephalomyelopathy. Thus, further investigations should utilise MR spectroscopy to investigate the possible mitochondrial origin of Rottweiler LEM.
The diagnosis of mitochondrial disorders faces specific difficulties due to the complex genetics of these conditions. Mitochondrial disorders may occur due to mutations in mitochondrial genes or mutations in nuclear proteins, with mitochondrial tRNA representing a hot spot for mutations. Heteroplasmy, i.e., the simultaneous presence of mutated and normal RNA/DNA in the cell, is a characteristic feature of mitochondrial disorders. The degree of heteroplasmy varies between tissues in the same organism, which is considered a critical factor in the manifestation of mitochondrial disease in specific tissues
[36, 39]. Finally, we investigated the coding region of the canine DARS2 gene as a candidate causative gene for LEM, and no mutation was found. At this time, we cannot rule out the possibility of variants in the promoter or intronic regions that could affect DARS2 expression. More comprehensive DNA sequencing approaches, such as the use of next-generation technologies for whole-exome or whole-genome resequencing, may enable the identification of the causative mutation of Rottweiler LEM. A recent study identified the causative mutation of canine neonatal cerebellar cortical degeneration in SPTBN2 (genome-wide mRNA sequencing) using only a single case of this neurodegenerative disease
Further limitations of our case report include the lack of pedigree analysis and brain lactate MR spectroscopy measurements and the failure of MR to demonstrate the involvement of the cerebrum despite the pathology observed in histological sections. Another limitation is that the comparison of the pathologies of these diseases in dogs and humans is limited by the paucity of case data from both. To date, there is only one short description of the pathology of LBSL in humans which has shown spongy white matter degeneration, rarefaction of the neuropil, macrophage infiltration and an increased number of astrocytes in the white matter of the brain and axonal degeneration of the peripheral nerves
. Spinal cord changes have not been noted, but it is unknown whether this part of the CNS was sampled and investigated. In dogs with LBSL-like changes, the neuroanatomical mapping of CNS lesions is more precise
[2, 29]. Clinical and pathological findings emphasise cerebellar and spinal changes, although the microscopic white matter damage is far more widespread and extends from the lower brain stem to the subcortical white matter. Consistent with the fibres affected, Gamble et al. discovered secondary grey matter changes in connected brain stem nuclei, such as the accessory cuneate nucleus, nucleus gracilis, nucleus cuneatus and nucleus of the dorsal spinocerebellar tract
. However, the involvement of multiple independent centres and tracts is compatible with multisystemic degeneration, as has been shown in Leonbergers and Rottweilers (discussed above)
[2, 29]. We also discovered macrovacuolar nuclear degeneration in the Rottweiler, which has not previously been described in dogs. Vacuole formation in LBSL patients was predominantly perineuronal and was therefore dissimilar to the neuronal vacuolation and spinocerebellar degeneration observed in Rottweiler dogs . This degeneration merits further examination to investigate the relationship between neurons and astrocytes in the subcortical grey matter. It also remains to be established whether this manifestation causes the white matter pathology or whether it is an additional, co-existing disorder that is distinct from the breed-specific neurodegenerative disorders described above.
In summary, magnetic resonance imaging revealed leukodystrophic lesions in the lateral funiculi of the cervical spinal cord; these findings will assist in the ante mortem diagnosis of future cases of suspected LEM in Rottweilers. Further investigations should utilise MR spectroscopy to investigate the possible mitochondrial origin of Rottweiler LEM. Although LEM is similar to LBSL based on its clinical features and imaging results, we were unable to identify a coding or splice site mutation in the canine DARS2 gene in our case, suggesting that other genes may be involved in Rottweiler LEM and potentially also in human LBSL.