Update on the Genetic Basis of Canine Degenerative Myelopathy
ACVIM 2008
Joan R. Coates, DVM, MS, DACVIM (Neurology); Claire Wade, PhD
Columbia, MO, USA; Cambridge, MA, USA


Dogs with degenerative myelopathy (DM) show an insidious, progressive ataxia and paresis of the pelvic limbs beginning in adulthood, ultimately leading to paraplegia and euthanasia.1 Degenerative myelopathy was first described by Averill as a spinal cord disorder that predominates in German Shepherd dogs (GSDs).1 If pelvic limb hyporeflexia is observed, nerve root involvement is presumed and the disease termed chronic degenerative radiculomyelopathy.2 Initially thought to be specific to the GSD, it also was designated German Shepherd Dog myelopathy.3 Clinical progression and histopathology of degenerative myelopathy have been described in few breeds and anecdotally reported in a number of purebred dogs. Many reports of DM in the veterinary literature lack thorough descriptions for the spectrum of clinical signs or lack histopathologic confirmation for cases in study.

Breed Predilection

Although the German Shepherd Dog is the most commonly affected breed, DM has been reported in other breeds4,5 and most recently in the Pembroke Welsh Corgi (PWC).6 Higher prevalence has been determined in a number of other purebred dogs.6 (Table 1)

Table 1. Canine breeds commonly affected with degenerative myelopathy.6


Total number

Number of
DM affected

Prevalence of
DM (%)

All dogs




German Shepherd Dog

19, 053



Welsh Corgi, Cardigan




Chesapeake Bay Retriever




Rhodesian Ridgeback




Irish Setter








Welsh Corgi, Pembroke




Fox Terrier, Wire








Old English Sheepdog




Mixed Breed




aBreeds that contain less than 5 cases of DM were excluded. DM, Degenerative Myelopathy
(Extracted from the aVeterinary Medical DataBases (VMDB) 2002--Courtesy of Roy D. Berghaus)

Clinical Spectrum

Detailed clinical descriptions of DM in dogs have been sporadic in the veterinary literature. There appears to be no sex predilection. Age of onset of neurologic signs is usually 5 years or older with a mean age of 9 years in large breed dogs with DM.1,2,3,7,8 A study in Pembroke Welsh Corgis reported a mean age of onset of 11 years.6 The clinical course of DM can vary up 3 years after the suspected diagnosis with a mean time for disease duration to be 6 months in large breed dogs.1 Breeds of smaller size allow the pet owner to give the appropriate care for their pet over a longer time.5,6 The median time of disease duration in the PWC was 19 months.6 Longer survival times in DM affected dogs may also be attributed to appropriate access to physical rehabilitation.8

Progressive, asymmetric upper motor neuron paraparesis and lack of paraspinal hyperesthesia are key clinical features of DM. Physical examination findings include weight loss of caudal trunk muscles, pelvic limb weakness, and worn nails. Descriptions for severity of loss of muscle mass have differed. Most reports attribute loss of muscle mass to disuse1,2,5,6 but flaccidity has been noted in the later stage of disease.4,5,6 Asymmetric weakness at disease onset is frequently reported.1,2,5,6,8 Gait deficits at time of onset show ataxia and mild paresis. Most large breed dogs progress to nonambulatory paresis or paraplegia within 4 to 8 months from time of diagnosis. If the disease progresses over a longer duration, clinical signs will ascend to affect the thoracic limbs.1,5,6 Due to its smaller size and longer disease duration, this was a more common scenario in the PWC. At disease onset, descriptions of spinal reflexes correspond with UMN paresis.1 Hyporeflexia noted with the patellar and withdrawal reflexes occurs in the latter disease stage.2,5,6,8 Urinary and fecal continence usually are spared also until the latter disease stage.1,5,6,8

Diagnosis and Prognosis

Tentative antemortem diagnosis presently is based upon ruling out other diseases causing progressive myelopathy.9,10 Common differentials include intervertebral disc disease, inflammatory disease, and spinal cord neoplasia. Hip dysplasia and degenerative lumbosacral stenosis often can be confused with DM but the neurologic findings are different if a careful examination is performed.10 Neurodiagnostic techniques for evaluation of spinal cord disease include CSF analysis, electrodiagnostic testing, myelography, computed tomography and magnetic resonance imaging. Electrodiagnostic testing has been rarely reported with abnormalities noted but significance still remains to be determined.2,6 Studies specific to the peripheral nervous system in DM affected dogs still need to be pursued. A presumptive diagnosis of DM often is made based on lack of clinically relevant compressive myelopathy as determined by myelography or MRI. A study using CT/myelography in confirmed and presumptively diagnosed DM dogs documented other spinal disorders: stenosis, spinal cord atrophy, and focal attenuation of subarachnoid space.11

Definitive diagnosis of DM is determined postmortem by histopathologic examination of the spinal cord. Spinal cord pathology of DM is consistent with a noninflammatory axonal degeneration. Histopathologic studies of DM have been more frequently described in the GSD.1,2,3,7 Affected dogs have characteristic patterns of axon cylinder vacuolization and drop out in this form of degenerative myelopathy. Lesion distribution involves the spinal cord myelin and axons in all funiculi,1,2 but affects the mid- to caudal thoracic region most extensively. A recent description of DM in the Pembroke Welsh Corgi characterized the overall distribution of white matter lesions to be similar to that observed in DM of large breed dogs.1,2,3,7 Descriptions of specific tract involvement in DM-affected GSDs have differed with respect to the degree of spinocerebellar tract involvement. 1,2,3,7 Previous studies, however, have consistently reported longitudinally discontinuous lesions in affected tracts with a tendency for increased lesion severity within the dorsal portion of the lateral funiculus and dorsal funiculus. 1,2,3,7 An important histopathological difference in PWC dogs is the longitudinally continuous nature of lesions within more clearly defined funicular areas.6 Ascending spinocerebellar and descending rubrospinal and lateral corticospinal tract involvement was especially evident in all spinal cord segments examined. The more severe and well-demarcated lesions in PWC dogs simply may reflect disease longevity.

Until a cause of DM is known, it is difficult to recommend an appropriate treatment regimen. Aminocaproic acid, an antiprotease agent, has been advocated for long-term management of DM;12 however, there have been no published clinical data to support drug efficacy. While vitamin deficiencies can cause spinal cord degeneration in some species,14-17 therapy with parenteral cobalamin or oral tocopherol did not affect neurologic progression in a study of DM affected dogs.18 (DA Williams unpublished data) Moreover, serum concentrations of alpha-tocopheral in DM affected GSDs did not yielded significant differences when compared to unaffected DM dogs.18 Combination therapies with an exercise regimen have been advocated for treatment of DM.8 However the long term prognosis remains poor.


The pathogenesis of canine degenerative myelopathy remains unknown. Griffiths and Duncan hypothesized DM to be a "dying-back disease" confined to the CNS suggesting a toxic etiology.2 Recently, studies of the brain of DM affected dogs showed neuronal degeneration and loss in the red nucleus, lateral vestibular nucleus and the dendate nucleus.7 Johnston et al. suggested that a defect in the neuronal perikarya may lead to abnormal axonal transport and degeneration in the distal axon. 7 However, we did not detect light microscopic neuronal perikarya pathology in the brainstem of DM affected PWC dogs.6

An immunologic role in the pathogenesis of DM was proposed based upon observations of depressed responses to thymus-dependent mitogens19 and increased concentrations of circulating immune complexes20. Although immune-related degenerative disease is a plausible theory, immunosuppressive therapies have shown no long-term benefits in halting the progression of DM.21 A study showed immunohistochemical evidence for immunoglobulin and complement deposition in the spinal cord of DM dogs but control dogs were lacking.22 However, lymphocyte and macrophage immunohistochemistry in other studies fail to reveal any areas of significant inflammation which further suggests that this form of degenerative myelopathy is non-inflammatory.

Predilection for severe lesions in the thoracic spinal cord may be a result of lower contribution of radicular arteries to perfuse and smaller diameter vessels, when compared to other spinal cord regions.9 Paucity of vessels in the thoracic spinal cord region may predispose neural tissue to damage from oxidative and metabolic disturbances. Oxidative stress, that is cellular damage occurring as a result of exposure to reactive oxygen radicals, prostanoids and excitatory amino acids, is believed to play an important role in degenerative spinal cord conditions.23,24 Elevation of glutamate concentrations have also been documented in dogs with chronic compressive spinal cord disease.25 Recent studies of a vasoactive prostanoid, 8-iso prostaglandin F2α (8-isoprostane) in CSF did not show significant concentration differences between control and DM affected dogs.6

Early DM studies suggested a genetic cause;3 however, the late onset of disease has made it difficult to collect data from parents and siblings to substantiate this theory. The first report to suggest familial disease of DM was in the Siberian Husky.4 Familial disease also is suspected in the PWC6, Rhodesian Ridgeback (Coates--unpublished data), Boxer (Coates--unpublished data), and Chesapeake Bay Retriever (personal communication--Dr. Sam Long, University of Pennsylvania). The lack of pedigree data and histopathologic confirmation in presumed affected dogs of early generations from affected families make it impossible, as of yet to evaluate the inheritance pattern of DM. Obtaining DNA samples from parents of DM affected dogs is difficult due to the late-onset of DM and parents being deceased. This makes it very important to collect samples from off-spring and normal siblings of DM affected dogs. Based on frequency of affected dogs having affected relatives, there is clear familial aggregation.

Review of Gene Association Mapping

There are several approaches that can be used to identify disease related genes including candidate gene sequencing, linkage mapping26,27 and genome-wide association28-30. Genome-wide association using single nucleotide polymorphism (SNP) arrays is a cost effective, highly automated and accurate approach. Association mapping uses cases and controls, removing the need of identifying multi-generation families typically used in linkage. The genome structure in dogs is particularly amenable to genome-wide association mapping given the long linkage disequilibrium (LD) within breeds31,32, yet the associated peaks are discrete in size (roughly ~1 Mb). Thus regions containing a gene predisposing to the disease are identified much more precisely than in linkage mapping. Finally, given the strong predisposition of DM in certain breeds lower number of cases and controls are needed to map a genetic risk factor than are required for human gene mapping.32

The canine genome sequencing project has produced a high-quality draft sequence of a Boxer covering ~98% of the genome and identified ~19,000 canine genes. In addition, the sequencing project has identified ~2.5 million SNP markers common to many breeds. With these resources, we can rapidly and easily move from markers associated with disease to candidate genes. This is the single most valuable resource for studies of canine genetic disease and is publicly available.32 As part of the genome project, the extent and number of haplotypes in individual dog breeds and across the dog population as a whole were studied. Haplotypes, and the related measure, LD, extend for long distances (up to megabases) within breeds.31,32 Across the whole dog population, haplotypes and LD are much shorter (of the order of 10Kb) , and the many haplotypes are shared between breeds.

Whereas human association studies require >300,000 evenly-spaced SNPs 28-30, because LD extends over approximately 50-fold greater distances in dog it is suggested that canine association studies would require perhaps only 10,000-15,000 evenly-spaced SNPs. Association mapping canine SNP arrays (Affymetrix and Illumina) has proven that 27,000 SNPs are sufficient to map mendelian traits with high accuracy.33,34

On-Going Study

We hypothesize that a major gene containing a genetic mutation may predispose dogs to develop DM and that the same mutation may be present in multiple breeds. To test this hypothesis we characterized the phenotype in specific breeds to see that it corresponds to the phenotype in Welsh Corgis. We performed genome-wide association mapping in a selected breed to identify a genomic location associated with disease. We then fine-mapped that locus and examined the genes at that chromosomal location in multiple breeds to identify the DM mutation.


1.  Averill DR. J Am Vet Med Assoc 1978;162:1045.

2.  Griffiths IR, et al. J Small Anim Pract 1975;16:461.

3.  Braund KG, et al. Am J Vet Res 39:1309.

4.  Bichsel P, et al. J Am Vet Med Assoc 1983;183:998.

5.  Matthews NS, et al. J Am Vet Med Assoc 1985;186:1213.

6.  Coates JR, et al. J Vet Int Med 2007;21:1323.

7.  Johnston PEJ, et al. Vet Rec 2000;146:629.

8.  Kathmann I, et al. J Vet Int Med 2006;20:927.

9.  Kneller SK, et al. J Am Anim Hosp Assoc 1975;11:414.

10. Braund KG. Vet Med 1987;782.

11. Jones JC, et al. J Vet Sci 2005;6:341.

12. Clemmons RM. Proceed 9th ACVIM Forum 1991;773.

13. Shevell MI, et al. The Molecular and Genetic Basis of Neurological Disease, 2nd ed. 1997;1301.

14. Mayhew IG, et al. J Vet Int Med 1987;1:45.

15. Sokol RJ. Free Radical Biol Med 1989;6:189.

16. Scott JM, et al. Lancet 1981;2:334.

17. Dinn JJ, et al. Lancet 1978;2:1154.

18. Johnston PEJ, et al. Vet Rec 2001;148:403.

19. Waxman FJ, et al. J Immunol 1980;124:1209.

20. Waxman FJ, et al. J Immunol 1980;124:1216.

21. Clemmons RM. Current Veterinary Therapy X Small Animal Practice 1989;830.

22. Barclay KB, et al. Can J Vet Res 1994;58:20.

23. Ilic TV, et al. Funct Neurol 1999;14:141.

24. Beal MF. Ann Neurol 1995;38:357.

25. Olby NJ, et al. J Neurotrauma 1999;16:1215.

26. Werner P, et al. Genomics 1997;42:74.

27. Todhunter RJ, et al. Am J Vet Res 2003;64:418.

28. Gabriel SB, et al. Science 2002;296:2225.

29. Cardon LR, et al. Trends Genet 2003;19:135.

30. Tsui C, et al. Nucleic Acids Res 2003;31:4910.

31. Sutter NB, et al. Genome Res 2004;12:2388.

32. Lindblad-Toh K, et al. Nature 2005;438:803.

33. Karlsson EK, et al. Nature Genetics 2007;39(11):1321.

34. Salmon-Hibbertz NH, et al. Nature Genetics 2007;39:1318.

Speaker Information
(click the speaker's name to view other papers and abstracts submitted by this speaker)

Joan Coates, DVM, MS, DACVIM (Neurology)
University of Missouri
Columbia, MO

Claire Wade, PhD
Broad Institute of MIT and Harvard
Cambridge, MA

MAIN : ACVIM Neurology Specialist : Degenerative Myelopathy
Powered By VIN