Abstract
Postmortem examination was performed on 18 Pembroke Welsh Corgi dogs (mean age 12.7 years) with clinical signs and antemortem diagnostic tests compatible with a diagnosis of degenerative myelopathy. Tissue sections from specific spinal cord and brain regions were systematically evaluated in all dogs. Axonal degeneration and loss were graded according to severity and subsequently compared across different spinal cord segments and funiculi. White matter lesions were identified in defined regions of the dorsal, lateral, and ventral funiculi. The dorsolateral portion of the lateral funiculus was the most severely affected region in all cord segments. Spinal cord segment T12 exhibited the most severe axonal loss. Spinal nerve roots, peripheral nerves, and brain sections were within normal limits, with the exception of areas of mild astrogliosis in gray matter of the caudal medulla. Dogs with more severe lesions showed significant progression of axonal degeneration and loss at T12 and at cord segments cranial and caudal to T12. Severity of axonal loss in individual dogs positively correlated with the duration of clinical signs. The distribution of axonal degeneration resembled that reported in German Shepherd Dog degenerative myelopathy but differed with respect to the transverse and longitudinal extent of the lesions within more clearly defined funicular areas. Although these lesion differences might reflect disease longevity, they could also indicate a form of degenerative myelopathy unique to the Pembroke Welsh Corgi dog.
Introduction
Degenerative myelopathy (DM) is a late onset, slowly progressive degeneration of spinal cord white matter that is reported primarily in large breed dogs. The German Shepherd Dog is the most commonly affected breed, but the condition has been described in other purebred dogs. 5, 9, 14, 22 Dogs with DM show an insidiously progressive ataxia and paresis of the pelvic limbs, ultimately leading to loss of the ability to ambulate within 6 to 12 months of the onset of signs. 3, 6, 14 Thoracic limb function and urinary and fecal continence are usually spared until end-stage disease. 9 Genetic, metabolic, nutritional, vascular, and immune-mediated etiologies have been proposed, but conclusive evidence for a specific pathogenic mechanism is lacking.
Definitive diagnosis of DM is confirmed by postmortem histologic examination of the spinal cord. Descriptions of nervous system lesions in DM vary both within and across different breeds. In affected German Shepherd Dogs, degenerative lesions within spinal cord axons and myelin have been described as discontinuous, bilateral, and asymmetric. 6 All white matter funiculi are affected, but a predominance of lesions in the dorsal portion of the lateral funiculus and dorsal funiculus has been reported. 14 Pathologic changes are characteristically most severe in the thoracic segments with mild to moderate degeneration found in cervical and lumbar segments. 3, 6, 14, 17 Brainstem, spinal cord gray matter, and dorsal nerve root lesions have been described in some studies but not in others. 3, 6, 14, 17 Other breeds with DM have similar patterns of axonal and myelin degeneration, but with varying degrees of lesion severity and funicular distribution. 5, 22 The variability in clinical time course before euthanasia and the low numbers of affected dogs examined retrospectively in many of these studies have made pathologic comparisons of DM across different breeds difficult.
In all affected breeds studied to date, degenerative changes in white matter funiculi consist of dilation of the myelin sheath, axonal swelling, and concurrent fragmentation and phagocytosis of axonal and myelin debris. 3, 5, 6, 14, 17, 22 Decreased Luxol fast blue staining confirms concurrent myelin degeneration, and moderate astrogliosis accompanies the degenerative changes. A majority of studies of DM in large breed dogs describe scattered individual fiber degeneration and loss without extensive involvement of groups of axons within distinct funicular regions. 3, 6, 14, 17
In this study, we characterize the pathologic changes in 18 Pembroke Welsh Corgi dogs (PWC) with degenerative myelopathy. Our objective was to establish postmortem diagnostic criteria for this recently described, breed-associated disorder. 10 The study population was unique in that many dogs had a protracted clinical course of up to 2 years before their owners elected euthanasia. The prospective nature of the study allowed a systematic and comprehensive assessment of patterns and severity of white matter degenerative changes at multiple spinal cord segmental levels. Pathologic findings that were consistently observed in the PWC are compared with those previously described in other canine breeds with DM. Potential pathogenic mechanisms underlying the primary morphologic lesions and their distribution in the central nervous system are discussed.
Materials and Methods
Cases were prospectively recruited between August 2001 and July 2004. Inclusion criteria for all dogs included a history of progressive caudal paresis, neurologic examination findings consistent with a nonpainful T3–L3 spinal cord localization, and absence of significant spinal cord compression on myelography, magnetic resonance imaging, or postmortem examination. Clinical laboratory data collected in a majority of dogs included a complete blood count, chemistry profile, urinalysis, and cerebrospinal fluid analysis. Thoracic radiographs were performed in all dogs. All pet owners of candidate dogs signed an approved informed consent for study participation.
All dogs were euthanatized with an overdose of barbiturate, and an immediate gross postmortem examination was performed. The brain, spinal cord, and when available, dorsal and ventral root fibers and sciatic nerve were immersion-fixed in neutral buffered 10% formalin for light microscopic evaluation. All tissues were subsequently processed and analyzed at The Ohio State University College of Veterinary Medicine. Dogs with evidence of lesions suggestive of a focally compressive process were excluded from the study. Tissue sections from 4 regions of the spinal cord (C6, T3, T12, and L4) in each dog were paraffin-embedded and sectioned at 4 µm. Blocks were oriented so that both transverse sections and longitudinal sections of cord funiculi could be obtained in the horizontal plane. Brain tissue of representative areas of the cerebellum, midbrain, pons, and medulla was also embedded and sectioned from all dogs. Histologic stains employed included hematoxylin and eosin (HE), Luxol fast blue with a periodic acid–Schiff (PAS) counterstain, and Bielschowsky silver stains. Immunostaining was performed on paraffin-embedded tissues with primary antibodies directed against neuron-specific enolase, glial fibrillary acidic protein (GFAP), vimentin, synaptophysin, CD3, and CD79a (DAKO, Carpinteria, CA); CD18 (courtesy of Dr. Peter Moore, Davis, CA); and complement component C3 (ICN Biomedicals, Inc., Aurora, OH). For all immunohistochemical procedures, a streptavidin–biotin method was used. Thoracic sections from a 13-year-old control dog were processed with the above test reagents to determine the relationship of age-related changes to changes found in DM-affected PWC.
Axonal degeneration (characterized by distended myelin sheaths with either absent, fragmented, or swollen axonal profiles in HE sections) and axonal loss (characterized by complete loss of recognizable axons and myelin sheaths in HE and Luxol fast blue sections) were graded according to severity. Grading was performed independently by 2 of the authors (PAM, RJA), and any discrepant scores were re-evaluated by both investigators. A scoring system from 0 to 6 was used to grade numbers of degenerating axons per microscopic field, and a scoring system of 0 to 4 was used to grade the degree of axonal loss per microscopic field. Entire white matter regions within each funiculus could be incorporated into each microscopic field. All white matter regions on both sides of the spinal cord were scored, and mean scores from homologous regions were calculated. Degeneration scores of 0, 1, 2, 3, 4, 5, and 6 represented 0, 1–10, 11–20, 21–40, 41–60, 61–80, and >80 lesions per 100× field, respectively. Axonal loss scores of 0, 1, 2, 3, and 4 represented 0–5%, 6–10%, 11–20%, 21–40%, and 41–60% of axonal loss per 100× field, respectively. Loss of intensity of Luxol fast blue staining and increases in intensity of GFAP immunostaining were subjectively assessed as mild, moderate, or severe.
The predominant distribution of lesions within axons in each funiculus was determined to develop an understanding of progression of axonal degeneration and disease pathogenesis. Degenerating fibers and axonal loss in each segment and funiculus were compared across all dogs. Means and standard deviations of these parameters were compared by a 1-way analysis of variance. When indicated by a significant F-statistic, specific means were compared by Tukey's post hoc test for multiple comparisons. In addition, subgroups of dogs classified as having mild to moderate (subgroup A) or severe (subgroup B) axonal loss were compared. In these subgroups, the degree of axonal degeneration and loss were compared in different segments and funiculi by Student's t-test. For individual dogs, a grand mean axonal loss score for all funiculi within all 4 segmental regions examined was calculated. A linear regression analysis of individual mean axonal loss scores versus duration of clinical signs before euthanasia was performed. For all statistical analyses, P < .05 was considered statistically significant.
Results
Eighteen dogs met inclusion criteria for this study. An additional 3 dogs met clinical and antemortem diagnostic testing inclusion criteria but were not included because of a lack of available tissues necessary to perform a complete histologic analysis. Mean age at onset of clinical signs was 11.4 years (range, 9–14.8 years) and mean age at euthanasia was 12.7 years (range, 10–16 years). Twelve dogs were spayed females, 2 dogs were intact females, 3 dogs were castrated males, and 1 dog was an intact male. Body weight ranged from 8.6 to 18.0 kg with a mean of 11.6 kg. The mean duration of clinical signs before a presumptive diagnosis of DM was 10 months (range, 1–24 months). At the time of presumptive diagnosis, 15 dogs were either nonambulatory or exhibited marked paraparesis and ataxia with intermittent falling in the pelvic limbs. All 18 dogs were nonambulatory, and 15 dogs had signs of thoracic limb weakness at the time of euthanasia. Nine dogs exhibited either fecal or urinary incontinence. The total duration of clinical signs before euthanasia ranged from 10 to 31 months (mean, 17 months). Nine of the affected dogs had a known relative also affected with DM.
Neural tissues were removed within 1–2 hours of euthanasia. Gross extradural annular disc protrusions were present in the spinal canal of 5 dogs. These disc protrusions were mild and were not causing spinal cord compression. No gross lesions were detected in the brain, spinal cord, and peripheral nerves of affected dogs. The primary microscopic lesions were axonal degeneration and axonal loss. Degeneration was characterized by distended axons and myelin sheaths along white matter tracts. On transverse sections stained with Luxol fast blue, areas of fiber degeneration and loss were characterized by severe loss of staining (pallor) (Figs. 1–3) compared with sections from an age-matched control dog. Extensive astrogliosis of these regions was confirmed by intense GFAP immunoreactivity (Fig. 4). A diffuse increase in Bielschowsky staining in these areas was observed (data not shown). Axonal loss was characterized by complete loss of recognizable axon and myelin profiles (Fig. 5), with replacement by zones of homogeneous eosinophilic matrix compatible with astrocytosis-associated astroglial processes (Fig. 6). Degenerating fibers were devoid of central axonal profiles in most instances, but an occasional axonal spheroid was present. Macrophages were also observed sporadically within swollen empty myelin sheaths (Fig. 6). On longitudinal sections, Bielschowsky staining also readily illustrated areas of axonal degeneration characterized by swollen, irregular, fragmented axon material within dilated tubular spaces (Fig. 7). T and B lymphocytes were absent in most microscopic fields examined. Immunoreactivity for the complement component C3 was also minimal to absent.
Spinal cord segment T12; dog No. 1. White matter degeneration is depicted by defined regions of pallor within the dorsal, dorsolateral, ventrolateral, and ventral funiculi. Luxol fast blue with PAS counterstain. Bar = 1 mm.
Spinal cord segment T12; dog No. 2. Patterns of white matter degeneration were similar between dogs (compare with Fig. 1); however, some variation in lesion severity was found between different funiculi. Luxol fast blue with PAS counterstain. Bar = 1 mm.
Spinal cord segment T12; dog No. 3. Moderate to severe white matter degeneration is present in all funiculi but is most evident in the fasciculus gracilis and the dorsolateral funiculus. Luxol fast blue with PAS counterstain. Bar = 1 mm.
Spinal cord segment T12; dog No. 2. Large areas of moderate to severe astrogliosis were identified in areas of moderate to severe axonal degeneration and loss (arrow). Compare this section with an adjacent section from the same dog stained with Luxol fast blue and PAS (Fig. 2). GFAP immunohistochemistry counterstained with hematoxylin. Bar = 1 mm.
Spinal cord segment T3, dorsolateral white matter; dog No. 1. Regions of pallor were most severe in the dorsolateral funiculus and were characterized by extensive loss of recognizable axonal and myelin profiles (asterisk). Blood vessel walls are PAS positive. Luxol fast blue with PAS counterstain. Bar = 50 µm.
Spinal cord segment T12, dorsolateral white matter; dog No. 1. Axonal degeneration was characterized by swollen myelin sheaths with either absent axon profiles or degenerating axon debris. Occasional macrophages were identified engulfing axon and myelin debris (arrow). Regions of axonal loss were identified by zones of complete loss of recognizable fiber cylinders and replacement by homogeneous eosinophilic matrix (asterisk). HE. Bar = 50 µm.
Spinal cord segment T12, dorsolateral white matter; dog No. 3. Longitudinal section of dorsolateral white matter depicting axonal degeneration in multiple fibers. Swollen, elliptical digestion chambers containing fragmented axon and myelin (arrowheads) and occasional macrophages are demonstrated. Bielschowsky silver stain. Bar = 50 µm.
CD18-positive macrophages were moderately increased in number in affected dogs, especially in the dorsal portion of the lateral funiculi (referred to as the dorsolateral funiculus in this manuscript) (Fig. 8). Brain sections were within normal limits, with the exception of areas of mild astrogliosis in gray matter of the caudal medulla.
Spinal cord segment T12, dorsolateral white matter; dog No. 5. Activated macrophages immunoreactive for CD18 were more numerous in the dorsolateral white matter areas of affected dogs compared with an age-matched unaffected dog. CD18 immunohistochemistry. Bar = 50 µm.
Degeneration of white matter tracts followed a similar overall pattern between dogs, with lesions present in all funiculi (Fig. 9). Axonal loss was significantly greater in the dorsolateral funiculus (Figs. 10, 11). The areas of dorsolateral axonal loss appeared well demarcated and involved both peripheral and deeper white matter tracts. This lesion was best visualized on Luxol fast blue staining (Figs. 1–3, 5). This pattern of greater pathologic change in the dorsolateral funiculus was characteristic of all cord segments examined. Cord segment T12 exhibited greater axonal loss in the dorsolateral funiculus compared with segments C6, T3, and L4 (Fig. 11). The dorsal funiculus, ventral funiculus, and ventral and lateral parts of the lateral funiculus (referred to as the ventrolateral funiculus below) showed similar degrees of axonal loss (Fig. 10). Lesions in the dorsal funiculus were localized to the most medial and peripheral portions of this funiculus in all segments and were frequently asymmetric (Figs. 1–3, 9). Patterns of axonal loss in the ventral funiculus varied depending on the cord segment level examined. Generally, the more ventromedial and peripheral white matter tracts were affected, although involvement of deeper ventral tracts was observed at T3 and T12 in some dogs (Fig. 9). Axonal loss in the ventrolateral funiculus caudal to T3 was consistently concentrated in the outer 20% of this funicular zone (Figs. 1, 3, 9).
Diagrammatic representation of lesion distribution in different spinal cord segments in PWC DM. Segmental level is denoted to the right of each transverse section. Small filled circles (black dots) represent areas of axonal loss (complete loss of recognizable fiber cylinders). Larger open circles represent areas of axonal degeneration (distended myelin cylinders with empty or degenerating axon profiles). Lesion severity was greatest at spinal cord segment T12, but similar patterns of axonal loss and degeneration were identified at all levels.
Bar graph showing mean ± standard error of the mean (SEM) scores for axonal loss in different funiculi of spinal cord white matter in 18 PWC dogs with DM. Axonal loss in the dorsolateral funiculus was significantly greater than in all other funiculi when scores were pooled from all spinal cord segments examined in each dog. Bar graph means with different letters are significantly different (P < .05). DF = dorsal funiculus, DLF = dorsolateral funiculus, VLF = ventrolateral funiculus, VF = ventral funiculus.
Bar graphs showing mean ± SEM scores for axonal loss within the dorsolateral funiculus (DLF) in different spinal cord segments. Axonal loss in the DLF was greater at T12 than in other segments; however, this difference was significant for the T12/C6 comparison only. Bar graph means with different letters are significantly different (P < .05).
Dogs with severe overall pathologic change (group B; n = 7) and those with mild to moderate overall pathologic change (group A; n = 11) were analyzed separately to determine the predominant pattern or funicular distribution of lesions with disease progression. Patterns of axonal loss were similar between groups A and B. However, severely affected dogs in this study showed significant progression of axonal loss at each spinal cord segment area examined compared with mild to moderately affected dogs. When individual funiculi were examined, axonal loss was significantly greater in group B for the dorsal, ventral, and ventrolateral funiculi at T3 and for the dorsal and dorsolateral funiculi at T12. Axonal degeneration scores were significantly greater in group B for the dorsolateral, ventral, and ventrolateral funiculi at C6 and T12 and for the dorsolateral and ventrolateral funiculi at T3.
Mean duration of clinical signs in group B dogs before euthanasia was 25 months (range, 17–31 months). Mean duration of clinical signs in group A dogs was 13 months (range, 6–20 months). A significant positive correlation (P < .01) was found between individual mean axonal loss score and duration of clinical signs before euthanasia (r 2 = .523) (Fig. 12).
Plot of correlation between mean axonal loss and duration of clinical signs before euthanasia in PWC DM. Linear regression analysis yielded a significantly positive correlation (P < .01). Each plotted data point represents a grand mean axonal loss score calculated by pooling all scores across all funiculi and segments in an individual dog.
Axonal loss and degeneration were sporadic or absent in sections of sciatic nerve, cauda equina, brainstem, and cerebellum. Dorsal and ventral nerve roots that were available for study exhibited occasional mild degeneration but were considered within normal limits compared with changes observed in an age-matched control dog. Neither neuronal loss nor astrogliosis were evident in Clarke's column, dorsal column nuclei, or brainstem nuclei. Axonal degeneration and axonal loss scores in spinal cord funiculi of the age-matched control dog were consistently less than 2 and 1, respectively.
Discussion
The presence of primary axonal degeneration and loss restricted to spinal cord white matter in the dogs of this study is compatible with a diagnosis of degenerative myelopathy. Despite the regionally defined areas of white matter lesions, a large number of ascending and descending spinal cord tracts were affected, with no apparent predilection for a specific proprioceptive or upper motor neuron system. Propriospinal and other non- or thinly myelinated tracts were spared. Obvious brainstem and cerebellar lesions were absent, distinguishing this disorder from other central axonopathies and myelinopathies with concurrent spinal cord and brain pathology. 29 Mild histologic changes in nerve roots and peripheral nerves appeared to be compatible with age-related changes and did not support an obvious radiculopathic or peripheral nervous system component of this disorder. Indicators of inflammatory change were not evident in any of the spinal cords examined. Progressive, multisystem central axonal degeneration with secondary myelin loss appeared to be the primary pathologic process occurring independent of the duration of clinical signs before euthanasia.
The overall pattern of multifocal involvement of different white matter tracts, widespread segmental distribution of lesions with a predilection for the thoracic spinal cord, and temporal progression of clinical signs parallel findings in German Shepherd Dog DM. The severity and more extensive segmental distribution of lesions along defined white matter pathways are histopathologic features in PWC DM that could help distinguish this disorder from DM in large breed dogs. Even in affected German Shepherd Dogs with DM, however, lesion severity and distribution can be variable. 3, 6, 14, 17 Lesion variability between and within breeds affected by DM raises questions concerning whether there are different pathogenic variants of DM. Dogs in this study generally had a more protracted clinical course before euthanasia (mean of 17 months) than that reported in German Shepherd Dog DM. 3, 6 This observation suggests that duration of disease is related to lesion severity. However, additional factors might be responsible for these different disease phenotypes.
All affected dogs had a characteristic pattern of axonal degeneration and loss within multiple spinal cord funiculi. Defined areas of axonal degeneration and loss affected consistent funicular compartments in a longitudinally continuous manner. Dogs with a longer clinical course had greater degrees of axonal loss and degeneration within funicular tracts across a longitudinally greater number of spinal cord segments. Regional axonal loss was severe in many dogs, as evidenced by complete loss of recognizable axonal and myelin profiles and replacement by large areas of astrogliosis. The severity and longitudinal extent of axonal loss and degeneration far exceeded that reported in German Shepherd Dog DM. 3, 6, 14 Spinal cord lesions in the German Shepherd Dog have been described as being multifocal and discontinuous, with sporadically affected fibers demonstrating axonal degeneration and myelin loss. 6 Pathologic studies of German Shepherd Dog DM report loss of axons and myelin sheaths in nearly equal proportions but rarely describe extensive areas of complete fiber loss in large regions of spinal cord white matter. Severe and continuous regions of axonal and myelin degeneration similar to those found in PWC DM were described in an adult Miniature Poodle with degenerative myelopathy. 22 As in the PWCs of this study, this dog had a protracted clinical course that might have influenced the degree of spinal cord pathology.
Degeneration and loss of large-diameter myelinated fibers were consistently greater in thoracic segments than in cervical and lumbar segments irrespective of duration of clinical signs. A selective vulnerability of fibers in the thoracic spinal cord was first proposed by Averill 3 in his study of DM in the German Shepherd Dog. Relatively fewer degenerative changes in the cervical spinal cord can be partially explained by a “dilutional effect” of unaffected ascending fibers from thoracic limbs, cervical musculature, and other cranial structures and of unaffected proximal descending fibers from upper motor neuron centers. 27, 30, 31, 35 The predominance of lesions in the thoracic spinal cord, however, cannot be fully explained by normal anatomic changes in tract sizes at different segmental levels.
Location of white matter lesions in this study coincided with the anatomic distribution of descending and ascending tracts in the spinal cord. Descending and ascending tracts in the dorsolateral funiculus were the most severely affected tracts. Other proprioceptive and motor tracts, however, were also moderately to severely affected throughout all spinal cord segments. At C6, T3, and T12, peripherally located lesions in the dorsolateral funiculus were in the area of the dorsal spinocerebellar tract. 11, 15, 19, 27, 30, 31, 35 The location of lesions in the ventral parts of the lateral funiculus at T3, T12, and L4 and in the lateral funiculus at C6 corresponded to regions occupied by the ventral spinocerebellar tracts. 11, 15 Pathologic changes in the ventral parts of the lateral funiculus at C6 were significantly less than those in the same funicular region at T3, T12, and L4. This finding can be explained by the shift of the ventral spinocerebellar tract to a more dorsolateral location as it ascends the cervical spinal cord. 27, 31, 35 White matter lesions in deeper portions of the dorsolateral funiculus and medial areas of the ventral funiculus were concentrated in areas occupied by descending upper motor neuron tracts. 11, 15 Lesions within this region of the dorsolateral funiculus were most likely involving the lateral corticospinal, rubrospinal, and medullary reticulospinal tracts. 14, 17 This was further supported by the finding of extensive lesion in the dorsolateral white matter at L4. The fourth lumbar segment is caudal to the origin of dorsal spinocerebellar fibers, and so the bulk of the dorsolateral white matter contains the descending upper motor neuron tracts described above. 8, 24– 26 Within the dorsal funiculus, the dorsomedial location of the lesions coincided with the location of the fasciculus gracilis. Extensive dorsal funiculi lesions caudal to L3 suggest involvement of not only the fasciculus gracilis but also unconscious proprioceptive afferents ascending to Clarke's column located between cord segments T3 and L3. 11, 13, 15
Lesions in these ascending and descending white matter pathways explain the clinical neurologic signs of proprioceptive loss and paresis in affected dogs. Predominance of pelvic limb signs reflects the relative severity of lesions in the thoracic spinal cord. The transverse and longitudinal extent of spinal cord lesions in PWC DM parallel the more severe neurologic deficits found in this breed compared with other breeds with DM. Cervical lesions in upper motor neuron pathways are a likely explanation for the clinical progression from paraparesis to tetraparesis. Lumbar and thoracic lesions in these same pathways explain the progression from paraparesis to paraplegia. Furthermore, the clinical finding of fecal and urinary incontinence in 9 dogs also suggests involvement of ascending visceral sensory pathways in the dorsal funiculi that signal colon, rectal, and urinary bladder distension. These sensory pathways have been identified in the dorsal funiculus of the lumbosacral and thoracolumbar spinal cord of dogs. 1, 2, 34 Other more noxious visceral sensations are transmitted in the spinothalamic tract. Lesions found in the dorsal funiculus of dogs in this study could be contributing to a lack of visceral sensory feedback to brain centers. The mild and sporadic lesions in the caudal equina nerve roots would not be expected to explain signs of incontinence in dogs of this study.
The pathogenesis and cause of canine degenerative myelopathy remain unknown. Both immunologic and metabolic mechanisms have been proposed. 9, 16, 32, 33 Immune-mediated inflammatory events were hypothesized to lead to a chronic demyelinating disease similar to multiple sclerosis or experimental allergic encephalomyelitis. 33 A more recent study indicated that immunoglobulin and complement deposition might be greater in spinal cord tissues of DM-affected dogs. 4 Both routine histologic and more specific immunohistochemical methods in our study failed to reveal any areas of significant inflammation. Complement deposition was not found, and neither T nor B lymphocyte infiltrations were identified in any of the spinal cord tissues examined. Both vitamin E and vitamin B-12 deficiencies have been associated with a primary myelin defect with secondary axonal degeneration in several species. 12, 16, 23, 28, 29 White matter degeneration affects cervical, thoracic, and cranial lumbar spinal cord regions in these nutritional disorders. In this study, patterns of concurrent axonal and myelin degeneration were similar to that found in Wallerian degeneration. Fragmentation and phagocytosis of axonal debris appeared to precede or coincide with myelin disintegration and phagocytosis. Myelin degeneration with concurrent axon preservation was not observed. Large areas of axonal loss in specific white matter tracts suggest a primary axonal/neural disorder instead of a more generalized demyelinating disorder secondary to an immune-mediated or metabolic cause.
A distal dying-back process because of defective axonal transport has a predilection for very long myelinated axons and is an attractive hypothetical mechanism in large breed dog DM because of the relative length of ascending and descending spinal cord fibers. Recently, neuronal degeneration and loss were identified in the red nucleus, the lateral vestibular nucleus, the fastigeal nucleus, and the dentate nucleus of German Shepherd Dogs with DM. 17 Extensive astrogliosis was found in the dorsal horn regions that coincide with the location of cell bodies giving rise to spinocerebellar tracts. A defect in the neuronal perikarya leading to abnormal axonal transport and degeneration in the distal axon was hypothesized. 17 We did not detect neuronal perikarya pathology in the dorsal horn, brainstem, or cerebellum of DM-affected PWC dogs; however, stereologic counts were not performed to assess neuronal loss in these gray matter nuclei. Although this form of distal-to-proximal axonal degeneration might explain some of the pathologic changes found in the upper motor neuron tracts in this study, it does not adequately explain the presence of lesions found in the proximal to intermediate portions of axons within proprioceptive pathways in thoracic segments and the relative paucity of lesions in distal proprioceptive fiber termination sites in brainstem and cerebellum. 7, 11, 15, 18– 21, 31
Involvement of distal, proximal, and middle axonal segments of the various tracts described would be consistent with either a defect in cells supporting axon maintenance (astrocytes, oligodendrocytes, or both) or defects in both anterograde and retrograde axoplasmic transport. An inborn error of neural or supporting cell metabolism with a delayed onset of expression could theoretically lead to late-onset disruption of axonal integrity in distinct neuronal systems. A hereditary basis for PWC DM is likely, given the historic evidence of DM occurrence in relatives of 9 PWCs in this study. In this study, we provide histologic criteria necessary to establish a diagnosis of PWC DM. Knowledge of consistent morphologic changes in this disease will benefit future investigations into disease pathogenesis, diagnosis, and therapy.
Footnotes
Acknowledgements
The authors acknowledge support provided by a grant funded from the American Kennel Club Canine Health Foundation (grant 2220) and The Pembroke Welsh Corgi Club of America. We thank the pet owners, ACVIM Diplomates, other specialists, and general practitioners for assistance with case recruitment. The authors also thank Dr. Alexander de Lahunta for his helpful advice and Beth Mellor for her assistance with the images. This manuscript has been prepared according to the Uniform Requirements for Manuscripts Submitted to Biomedical Journals.