Acta Neuropathol (2009) 118:697–709
DOI 10.1007/s00401-009-0548-6
ORIGINAL PAPER
Leigh-like subacute necrotising encephalopathy in Yorkshire
Terriers: neuropathological characterisation, respiratory chain
activities and mitochondrial DNA
Kerstin Baiker Æ Sabine Hofmann Æ Andrea Fischer Æ
Thomas Gödde Æ Susanne Medl Æ Wolfgang Schmahl Æ
Matthias F. Bauer Æ Kaspar Matiasek
Received: 3 May 2009 / Revised: 7 May 2009 / Accepted: 8 May 2009 / Published online: 23 May 2009
Springer-Verlag 2009
Abstract Our knowledge of molecular mechanisms
underlying mitochondrial disorders in humans has
increased considerably during the past two decades.
Mitochondrial encephalomyopathies have sporadically
been reported in dogs. However, molecular and biochemical data that would lend credence to the suspected
mitochondrial origin are largely missing. This study was
aimed to characterise a Leigh-like subacute necrotising
encephalopathy (SNE) in Yorkshire Terriers and to shed
light on its enzymatic and genetic background. The
S. Hofmann and M. F. Bauer contributed equally.
K. Baiker W. Schmahl K. Matiasek
Chair of General Pathology and Neuropathology,
Institute of Veterinary Pathology, Ludwig-Maximilians
University of Munich, Munich, Germany
S. Hofmann
Institute of Microbiology & Laboratory Diagnostics,
Gauting, Germany
A. Fischer
Section of Neurology, Small Animal Medical Clinic,
Ludwig-Maximilians University of Munich, Munich, Germany
T. Gödde
Small Animal Referral Practice, Piding, Germany
S. Medl
Small Animal Referral Clinic, Babenhausen, Germany
M. F. Bauer
Institute of Clinical Chemistry and Molecular Diagnostics,
Ludwigshafen City Hospital, Ludwigshafen, Germany
K. Matiasek (&)
Neuropathology Laboratory, The Animal Health Trust,
Lanwades Park, Kentford, Newmarket, Suffolk CB8 7UU, UK
e-mail: [email protected]
possible resemblance to SNE in Alaskan Huskies and to
human Leigh syndrome (LS) was another focus of interest.
Eleven terriers with imaging and/or gross evidence of
V-shaped, non-contiguous, cyst-like cavitations in the
striatum, thalamus and brain stem were included.
Neuropathological examinations focussed on muscle, brain
pathology and mitochondrial ultrastructure. Further investigations encompassed respiratory-chain activities and the
mitochondrial DNA. In contrast to mild non-specific
muscle findings, brain pathology featured the stereotypic
triad of necrotising grey matter lesions with relative preservation of neurons in the aforementioned regions, multiple
cerebral infarcts, and severe patchy Purkinje-cell degeneration in the cerebellar vermis. Two dogs revealed a
reduced activity of respiratory-chain-complexes I and IV.
Genetic analyses obtained a neutral tRNA-LeuUUR A-Gtransition only. Neuropathologically, SNE in Yorkshire
Terriers is nearly identical to the Alaskan Husky form and
very similar to human LS. This study, for the first time,
demonstrated that canine SNE can be associated with a
combined respiratory chain defect. Mitochondrial tRNA
mutations and large genetic rearrangements were excluded
as underlying aetiology. Further studies, amongst relevant
candidates, should focus on nuclear encoded transcription
and translation factors.
Keywords Leigh syndrome Subacute necrotizing
encephalopathy Mitochondrial Respiratory chain defect Canine Yorkshire Terrier
Introduction
In humans, mitochondrial disorders affecting the central
nervous system comprise a very heterogeneous group of
123
698
neurological and neuromuscular phenotypes that are either
caused by mutations in the nuclear genome or the
mitochondrial DNA (mtDNA). The vast majority of
mitochondrial diseases are related to respiratory chain
dysfunction. This may affect all tissues, but metabolically
active, post-mitotic tissues with high energy demands such
as the central nervous system (CNS), heart and skeletal
muscle are most vulnerable [16, 35].
With a birth prevalence ranging between 1:30,000 and
1:40,000 live births, human Leigh syndrome (LS) represents
one of the most important mitochondrial encephalopathies
that manifests as neurodegenerative disorder [7, 24]. The
fatal infantile form [19] presents with a characteristic brain
pathology consisting of bilaterally symmetrical necroses
confined to the basal ganglia, thalamus, and brain stem,
cerebellar cortical degeneration, and widespread gliosis and
vascular proliferation [8, 10, 29].
Both the large panel of biochemical and genetic
abnormalities, and the variable clinical presentation indicate that LS encompasses a heterogeneous group of
disorders rather than a single disease entity.
Accordingly, numerous genetic defects in nuclear genes
and within the mtDNA have been identified in human LS
patients, but the molecular mechanism linking a specific
gene defect to the observed clinical phenotype is still
poorly understood. The epidemiological situation and
ethical considerations impede systematic studies in human
patients; therefore, an animal model appears essential for
gaining further insights into LS pathobiology.
There are several reports on dogs with clinical and
neuropathological phenotypes resembling human mt-disorders [1]. For example, a mitochondriopathy with
similarities to human mitochondrial encephalomyopathy
with lactic acidosis and stroke-like episodes (MELAS
syndrome) was described in a Jack Russell Terrier some
years ago [13].
These anecdotic cases are of limited value for mitochondrial science but larger groups of affected individuals
could be useful for systematic investigations. In that regard,
it may be of interest that a severe subacute necrotising
encephalopathy (SNE), featuring major neuropathological
aspects of human Leigh syndrome, has been reported in
Alaskan Huskies [2, 3, 34] and sporadically diagnosed ever
since. In contrast to the comprehensive neuropathological
analyses, the underlying functional and molecular genetic
defects have not been further investigated in these or other
dogs with suspected mitochondrial encephalomyopathy.
In 2005, the first pathogenic canine mtDNA mutation
was identified by Li and colleagues [20]. This gene defect
is associated with an inherited spongiform leukoencephalomyelopathy in Australian Cattle Dogs and Shetland
Sheepdogs [20]. The nucleotide exchange affects the gene
for the mitochondrially encoded cytochrome b, a subunit of
123
Acta Neuropathol (2009) 118:697–709
the complex III, and causes decreased levels of the components of complexes III and IV of the mitochondrial
respiratory chain and, thus, less ATP production.
Adding similar biochemical and molecular tools to
classic neuropathological investigations, we reinvestigated
a series of SNE cases in Yorkshire Terriers and assessed
the resemblance of Yorkshire Terrier SNE to the Alaskan
Husky form and human LS.
The availability of animals with a spontaneous LS-like
disease may provide a feasible approach for investigation
into the mechanisms of mitochondrial disease development, in general, since establishment of experimental
animals affected by phenotypes that resemble human disease is still hampered by considerable difficulties. Hence,
naturally occurring mitochondrial disorders may serve as
preferable models for comparative and translational
research.
Materials and methods
Animals
This retrospective investigation enrolled 11 unrelated
Yorkshire Terriers—euthanised at an age of 4 months to
5 years (13.5 ± 16.21 months; median: 7.0 months)—
with a tentative neuroimaging and/or gross postmortem
diagnosis of SNE [2, 3, 34]. Relevant animal data are
summarised in Table 1.
Inclusion criteria
The following imaging features were considered indicative
for SNE: well circumscribed, non-contiguous bilateral,
oblique hypodense (computed tomography, CT) areas
within the basal nuclei, mid-thalamus and brain stem that
appear hyperintense on T2-weighted magnetic resonance
(MR) images and hypointense in T1-weighted spin echos.
These changes may be accompanied by T2-hyperintense/
T1-hypointense lesions within the frontal, parietal and/or
temporal neocortex [3, 34].
On necropsy, a bilateral gross discolouration, softening
and/or rarefaction of the brain tissue confined to the basal
nuclei and/or mid-thalamus was accepted as being suggestive of SNE [3, 34].
Morphological investigations
CNS pathology
On necropsy, the brains and first cervical spinal cord segments were carefully harvested and underwent an external
whole-brain examination before being immersed in 10%
Acta Neuropathol (2009) 118:697–709
699
Table 1 Signalment and clinical history
Yorkshire terrier
Gender
Age of onset
Duration
Clinical signs
I
F
18 mo
\6 mo
Disorientation, gait abnormalities
II
F
4 mo
2 mo
Gait abnormalities, central visual deficits
III
M
7.5 mo
1 wk
Gait abnormalities, central visual deficits, dysphagia
IV
F
18 mo
1 wk
Seizures, gait abnormalities
V
F
5 mo
6 mo
Gait abnormalities, reduced mental state
VI
F
6 mo
1 wk
Gait abnormalities, stupor, central visual deficits
VII
M
6 mo
\6 mo
Gait abnormalities, central visual deficits
VIII
M
5 mo
2 mo
Gait abnormalities, reduced mental state
IX
M
12 mo
1 wk
Gait abnormalities, postural deficits, dysphagia,
X
M
5 yrs
?
Seizures, gait abnormalities
XI
M
7 mo
2 mo
Gait abnormalities, muscle fasciculations, central visual deficits
M male, F female; wk weak(s), mo months, yrs years
neutral-buffered formalin for 24 h. Trimming of the brains
in transverse sections was followed by processing of all
macroscopically conspicuous areas plus seven standard
cross sections at the level of the precallosal frontal lobe, the
rostral commissure, the optic chiasm, the rostral and caudal
colliculi, the trigeminal nerve emergence and the facial
nerve motor nuclei. From the upper cervical spinal cord
several cross sections were taken caudal to the gracile
nucleus and the pyramidal decussation.
The CNS tissue was processed in an automatic tissue
processor, embedded in paraffin, sectioned at 8 lm and
stained with haematoxylin & eosin (H&E), luxol fast bluecresyl echt violet (LFB), Bodian0 s technique (Bod) and
Woelcke-Schroeder-Spielmeyer stain (WSS).
Immunohistochemical investigations (IHC) were performed on further deparaffinised sections using an avidinbiotin-peroxidase technique with diaminobenzidine as
chromogen (Vector, Burlingame, USA). Primary antibodies
were directed against von Willebrand factor (1:2000; Dako,
Hamburg, Germany), glial fibrillary acidic protein (GFAP;
1:400, Dako, Hamburg, Germany), pan-neurofilament
(pan-NF; 1:50, Zymed, San Francisco, USA) and myelin
basic protein (MBP; 1:400, Dako, Hamburg, Germany).
According to the respective species in which the primary
antibodies were generated, secondary, biotinylated goatanti-rabbit/mouse antibodies (DakoCytomation, dilution
1:100) were used.
(Tissue Tec, Sakura Finetek Europa, Zeoterwonde,
Netherlands). The panel of applied stains included H&E,
Giemsa stain, periodic acid Schiff, Engel0 s modified
Gomori trichrome and oil red O.
For all other dogs, skeletal muscle of various anatomical
origin (triceps, biceps femoris, gastrocnemius, tibialis
cranialis muscle) were evaluated after formalin fixation,
paraffin embedding, sectioning at 5 lm and staining with
H&E and Goldner0 s trichrome stain.
Muscle pathology
The activity of the citrate synthase, complex I
(NADH:oxidoreductase), complex II ? III (succinate:
cytochrome c oxidoreductase) and complex IV (cytochrome c oxidase) were determined in frozen muscle
homogenate using standard spectrophotometric methods
established previously for enzymatic respiratory chain
measurements in human muscle probes [11].
Fresh gastrocnemius muscle samples were available from
two dogs (III and VI). Without delay, the probes were
snap-frozen in isopentane cooled in liquid nitrogen and
stored at -80C until further processing. Cryosections of
8 lm thickness were prepared after embedding in OCT
Mitochondrial morphology
Transmission electron microscopy was carried out on brain
tissue and muscle samples retrieved from paraffin blocks or
per primam subjected to resin embedding. For the latter,
standard protocols were employed. In short, the samples
were fixed in 2.5% glutaraldehyde in Soerensen0 s buffered
saline (pH 7.4) and underwent postfixation in 1% OsO4,
and a graded alcohol series before embedding in epoxy
resin. Ultrathin sections were cut at 80 nm using a diamond
knife on a Reichert-Jung Ultracut. They were mounted on
copper grids and contrasted with uranyl acetate and lead
citrate. Ultrastructural analysis was performed and documented photographically with a Zeiss EM 10 at 80 kV
and a magnification from 1,7009 to 80,0009.
Enzymology
123
700
Genetic analyses
Using DNA isolated from muscle (animals III and VI) or
extracted from paraffin-embedded tissue sections (all other
animals) (GENIAL, First-DNA all tissue kit) all 22
mitochondrial tRNA genes and the ATP6 gene were
amplified using primers designed on the basis of the published canine mtDNA (Acc. No. AY729880) (numbers
represent nucleotide positions): f1, 16619–16628, r1, 375–
356 (tRNA-Phe); f2, 910–928, r2, 1211–1191 (tRNA-Val);
f3, 2552–2571, r3, 2897–2879 (tRNA-Leu(UUR)); f4, 3658–
3678, r4, 3990–3970 (tRNA-Ile, tRNA-Glu, tRNA-Met);
f5, 4810–4830, r5, 5460–5442 (tRNA-Trp, tRNA-Ala,
tRNA-Asn, tRNA-Cys, tRNA-Tyr); f6, 6823-6844, r6,
7069-7052 (tRNA-Ser(UCN), tRNA-Asp); f7, 7676-7698,
r7, 7921-7902 (tRNA-Lys); f8, 9331–9350, r8, 10082–
10064 (tRNA-Gly, tRNA-Arg); f9, 11524–11543, r9,
11846–11824 (tRNA-His, tRNA-Ser(AGY), tRNA-Leu
(CUN)
); f10, 14016–14036, r10, 14651–14632 (tRNA-Glu);
f11, 15014–15033, r11, 15729–15711 (tRNA-Thr, tRNAPro); f12, 7923–7943, r12, 8687-8668 (ATP6); f13, 15401–
15422, r13, 15681–15662 (non-coding D-Loop region).
PCR fragments were purified on agarose gel using standard
column extraction (Amersham, Freiburg, Germany) and
directly sequenced using the BigDye terminator cycle
sequencing protocol on an automatic ABI310 capillary
sequencer (Applied Biosystem Inc.). RFLP-analysis was
performed to screen for the A2691G mutation within the
tRNA-Leu(UUR) gene and determination of a possible
heteroplasmy. The A2691G mutation creates an ApaI
restriction recognition site (GAGCCC ? GGGCCC) and
clips the amplified PCR product into two smaller fragments
of 111 and a 122 bp upon ApaI digestion.
For detection of mtDNA rearrangements, a Southern blot
analysis was carried out on skeletal muscle of affected dogs
III and VI and on further tissue from healthy individuals.
Total muscle DNA (4 lg) was digested with EcoRI, HindIII
or ApaI. Southern blots were performed as previously
described [28] with total canine mtDNA resembling the
hybridisation probe. Approximately 0.05–0.1 lg of the
genomic DNA was used to amplify the mtDNA D-Loop
region by PCR. After amplification, the 261 base pairs
fragments were purified and directly sequenced.
Acta Neuropathol (2009) 118:697–709
pattern, gyration and foliation. After trimming of the fixed
brains in transverse plane, all but one dog showed reddishbrown discoloured, sunken or cavitated lesions of the midthalamus that extended non-contiguously into the striatum,
and, caudally, into the tegmentum of the medulla oblongata
and that occasionally were surrounded by gliotic scar tisse
(Fig. 1). Throughout all these brain regions, the lesions had
an oblique dorsolateral to ventromedial orientation. Within
the medulla oblongata, the malacia appeared V-shaped due
to fusion in the midline.
Results
Morphological investigations
CNS pathology
On gross examination, all brains appeared fully developed
and presented with a normal leptomeningeal vascular
123
Fig. 1 Gross findings in SNE after trimming in transverse sections
(a), sectioning and staining with H&E (b) and LFB (c). Most
strikingly, the mid-thalamic nuclei show a bilateral cyst-like rarefaction of the neuropil (arrows). Further malacic foci and/or cavitations
are observed in the neocortex of the parietal and temporal lobes (black
arrowheads) that occasionally extend into the adjacent subcortical
white matter of the corona radiata or centrum semiovale (white
arrowheads). Scale bar 1 cm
Acta Neuropathol (2009) 118:697–709
701
Within the neopallium of all affected dogs, a multifocal,
bilaterally symmetric softening and/or brownish discolouration at the grey:white-boundary was observed. These
changes were confined to the parenchyma that surrounded
the tips of the sulci (Fig. 1). Softening was most prominent
in the frontal and parietotemporal lobes, in proximity to the
cruciate sulcus and the suprasylvian sulcus. Accomplishing
the caudal fossa changes, two dogs presented with a
markedly blurred grey:white-boundary within the central
parts of the vermis.
In all affected dogs, histopathology was characterised by
varying degrees of neuronal loss, glial degeneration and
vascular proliferation. These changes showed a highly
consistent spatial distribution and some stage-specific
characteristics (Table 2).
The damage of basal nuclei, thalamus and brain stem
tegmentum was principally based on loss of nerve fibres
and glia cells (Fig. 2), whereas cerebrocortical (Fig. 3b)
and cerebellar lesions (Fig. 4) showed widespread nerve
cell necroses with secondary glial changes.
Throughout all sections, cerebrocortical necrosis was
accompanied by an extensive vascular prominence arising
from the watershed areas (Table 2; Fig. 3a). Vascular
hypertrophy in diencephalon and brain stem, on the other
hand, was restricted to the active stages of tissue necrosis.
The cerebellum did not exhibit vascular changes.
Grossly cavitating lesions of the thalamus and basal
nuclei corresponded to liquefaction of the mid-thalamic
and juxta-capsular neuropil with some remaining, poorly
cellular gliovascular trabecules (Fig. 2a–d), a varying
degree of invasion by foamy macrophages (gitter cells)
(Fig. 2e) and a peripheral fibrous gliosis. Special stains
revealed a majority of these trabecules containing myelinated (Fig. 2b) and/or unmyelinated nerve fibres. Further
naked axons were identified traversing some of the virtually empty spaces (Fig. 2d).
In one dog without gross thalamic changes, histology
showed a rather well circumscribed, active grey matter
degeneration and necrosis accompanied by severe oedema
and astrocytosis, moderate microglial proliferation and
infiltration by a moderate number of lymphocytes, plasma
cells and histiocytes with some interspersed groups of
gitter cells. Astroglial changes were accomplished by mild
to moderate recruitment of occasionally vacuolated gemistocytes (Fig. 2g). Thereby, relationship between the
numbers of astrocytes and macrophages was a reflection of
the duration of the process.
Apart from reactive phenotypes, numerous Alzheimer
type II cells were noted within necrotising foci and the
adjacent neuropil (Fig. 2f).
Similarly active stages of grey matter degeneration and/
or necrosis were confined to the periphery of the cavitations in all but one other dog, and to the lateroventral
aspects of the caput nuclei caudati (5/11), putamen (7/11)
as well as the mesencephalic (4/11) and pontine tegmentum
(10/11), and the medullary reticular formation (11/11)
(Table 2).
In cavitating burnt-out lesions the blood vessels were
neither increased nor hypertrophic. In active areas, on the
other hand, small blood vessels were congested, hypertrophic, thicker than normal and often appeared to be in
excess, although this latter feature was judged to be largely
due to collapse of the intervening neuropil from loss of
tissue.
Notably, in all these areas, neuronal changes were
restricted to a few chromatolytic perikarya and some axonal spheroids, whereas most of the even free-floating nerve
cells were spared from significant cytopathological abnormalities (Fig. 2e).
The middle and inner cortical laminae surrounding the
tips of many neopallial sulci revealed multiple infarcts with
extensive spongiosis due to enlarged perivascular and
perineuronal spaces, vasogenic oedema, nerve cell loss,
fading neurons and eosinophilic nerve cell necroses followed by an extensive astrogliosis. Microglia were
activated though not much increased.
Moreover, there was a marked excess of capillaries and
all small blood vessels were prominent due to endothelial
Table 2 Spatial distribution of neuropathological lesions
Cerebral
cortex
Basal
nuclei
Thalamus
Corpora
geniculata lat.
Mesencephalic and pontine
formatio reticularis
Medullary formatio
reticularis
Cerebellar
vermis
Cavitation
(?)
?
???
-
(?)
(?)
-
Malacia
??
??
???
-
?
??
-
Spongiosis
???
??
???
?
??
???
-
Neuronal necrosis
??
(?)
?
?
?
(?)
???
Glial degeneration
Gliosis
??
??
??
-
???
??
?
?
??
??
???
??
(?)
???
Vascular proliferation
???
(?)
???
?
??
??
-
- absent, (?) occasionally seen, ? mild, ?? moderate, ??? severe
123
702
Fig. 2 ‘‘Burnt-out’’ lesions of
the thalamus present
histologically as empty spaces
with/without a central island of
remaining parenchyma that are
transversed by trabecules
(a; H&E, black arrowhead).
These consist of gliovascular
strands containing myelinated
(b; WSS) and unmyelinated
nerve fibres and capillaries
(c; von Willebrand factor).
Bodian0 s staining displays also
some naked intralesional axons
spanning the cavities (d).
Regardless of the degree of
parenchyma loss, the remaining
neurons look remarkably
unaffected (e; H&E, arrow).
Resorption of the tissue debris is
carried out by gitter cells
(e; arrowheads). Numerous
Alzheimer type II cells are
observed within the periphery of
the lesion and the adjacent
neuropil (f; arrowheads).
Parenchyma destruction quite
often recruits vacuolated
gemistocytes (g; arrowheads).
Scale bars 750 lm in (a),
150 lm in (b), 45 lm in (c),
75 lm in (d), 35 lm in (e),
20 lm in (f) and (g)
123
Acta Neuropathol (2009) 118:697–709
Acta Neuropathol (2009) 118:697–709
swelling, hyperplasia of the myoendothelial complexes and
media, and occasional perivascular mononuclear aggregates (Fig. 3a).
703
Fig. 3 a A well-circumscribed, extensive vascular prominence
(arrowheads) and gliosis is noted in the affected cerebral cortex, in
addition to the neuronal loss. b At more chronic stages, cyst-like
cavitations (asterisk) develop that replace the middle and inner
cortical layers and the superficial parts of the neighbouring subcortical white matter, including the U-fibres. SC sulcus. Scale bars:
250 lm in (a), 2 mm in (b)
Four animals further showed a cyst-like cavitation of the
inner cortex and its adjacent subcortical white matter, again
with spanning gliovascular trabecules and some scattered
gitter cells (Fig. 3b).
In three dogs, eosinophilic nerve cell necroses and a
very mild astrogliosis and astrocytosis were observed in the
dorsal, magnocellular part of the lateral geniculate nuclei.
All dogs presented with a severe cerebellar cortical
degeneration confined to the spinocerebellar aspects of the
vermis. It was characterised by a substantial loss of Purkinje cells (PC) in central areas, indicated by a markedly
reduced density of PC (Fig. 4) and multiple empty baskets
upon Bodian0 s stain. PC loss tapered off towards the more
lateral aspects of the vermis where active neuronal
degeneration was observed. Corresponding to the severity
of PC damage, a moderate to severe Bergmann0 s gliosis,
replacement oedema, and neuronal depletion and gliosis of
the underlying granular cell layer was seen (Fig. 4). Axonal torpedos were rarely recognised.
Apart from thinning of the associated foliary white
matter, and a few Wallerian-like fibre degenerations, the
cerebellar white matter and cerebellar roof nuclei were
histologically inconspicuous. One of the dogs showed a
bilateral marked neuronal depopulation and astrogliosis of
the caudal olivary nuclei.
In all animals, the spinal cord was grossly unremarkable.
On histological inspection, however, seven dogs showed a
mild to moderate, bilaterally symmetrical spongiosis of the
deep dorsolateral white matter and the sulcomarginal tracts
due to multiple Wallerian-like fibre degeneration with
dilated myelin sheaths and occasional macrophages within
the myelin tubes. Grey matter changes and alterations of
the spinal nerve roots and rootlets were not recognised.
Fig. 4 The cerebellum shows a patchy Purkinje cell (PC) loss.
Compared to the relatively spared areas of the dorsal (declive, folium,
tuber) and ventral vermis (lingula, nodulus), rostral and caudal tips of
the folia and the cerebellar hemispheres (a), the cortex of the central
vermis presents with a severe PC degeneration and loss that progresses
from lateral (b) to medial (c). PC degeneration is accompanied by
increasing granule cell depletion in the granular layer (GL) and a
proliferation of Bergmann glia (BG). Accumulation of glial filaments
and loss of granule cell and PC dendrites results in a radially striped
appearance of the molecular layer (ML). Scale bar 50 lm
123
704
Acta Neuropathol (2009) 118:697–709
Fig. 5 Mild unspecific
mitochondrial changes in
muscle (a) and nerve (b–d)
fibres comprise enlarged and
irregularly shaped mitochondria
with distorted cristae (a–c;
asterisks), trilaminar inclusions
(a; white arrowhead),
multivacuolated mitochondria
with blebbing of mitochondrial
membranes (b; black
arrowheads), abnormal
compartment formation and
accumulation of electron dense,
sometimes paracrystalline, and
curvilinear material (c; white
arrowhead), axoplasmic
accumulation of large
mitochondria with irregular
cristae formation (d).
Magnification: 956,025 in (a),
941,500 in (b) and (c), 963,125
in (d)
Muscle pathology
Muscle changes were absent in two Terriers and mild in the
remaining dogs. In the latter, both cryostat sections and
paraffin-embedded tissue presented with multiple, randomly distributed single muscle fibres and small fibre
groups undergoing mild anguloid atrophy. Cryosections
did not show pathological sarcoplasmic storage of neutral
lipids or diastase-resistant polysaccharides. Distribution,
density and size of mitochondria appeared normal upon
Gomori0 s trichrome staining. In particular, ragged red
fibres were not recognised.
group, consisting of age-matched toy breed dogs (Fig. 6;
Table 3). The measured activities were 54 and 41% of the
lower limit of the reference range for complex I and IV,
respectively, if normalised for the mitochondrial marker
enzyme citrate synthase. In the other animal (VI), the
activities of these complexes were less-severely decreased
(81 and 90% of the lower limit of control range). The
activities of the complexes II/III were normal in both cases.
Genetic analyses
The evidence of a combined deficiency of complexes I and
IV of the mitochondrial respiratory chain in the two dogs
Electron microscopy
On ultrastructural level, several mitochondria within striated muscle cells appeared moderately enlarged. Moreover,
compared to age-matched control dogs, nerve cells and
myocytes contained an increased number of dysmorphic
mitochondria with abnormal cristae, compartmentalisation,
electron dense deposits and numerous trilaminar and
pentalaminar membranous inclusions (Fig. 5).
Mitochondrial respiratory chain activities
In order to quantify the mitochondrial respiratory chain
activities, the enzymes of two canine skeletal muscle
homogenates were determined using standard spectrophotometric methods. In animal III, the activities of complexes
I and IV were significantly lower than in our normal control
123
Fig. 6 Box plot displaying the muscle enzyme activities of respiratory chain complexes I through IV in dog III (square), dog IV (circle)
and the controls of age-matched toy breed dogs (boxes)
Acta Neuropathol (2009) 118:697–709
705
could be suggestive of a mitochondrial tRNA mutation.
We, therefore, screened all 22 mitochondrial tRNA-genes
for pathogenic mutations by direct sequencing of PCRamplified mtDNA fragments.
Table 3 Respiratory chain complex activities normalised for citrate
synthase activity
Animal III
Animal IV
Reference
range
Complex I [U/g NCP]
10.0
29.8
19–44
Complex I [U/U CS]
0.12
0.18
0.22–0.36
Complex II ? III [U/g NCP]
8.4
13.1
10–30
Complex II ? III [U/U CS]
0.1
0.08
0.08–0.21
Complex IV [U/g NCP]
59
250
135–260
Complex IV [U/U CS]
0.71
1.5
1.7–2.4
Citrate synthase (CS) [U]
83
162
56–142
U international units, g gram, NCP non-collagen protein, CS citrate
synthase
A homoplasmic nucleotide exchange from A to G was
identified at nucleotide position (np) 2691 of the Canis
familiaris mtDNA (Fig. 7). This A-to-G transition is
located within the gene for tRNA-Leu(UUR) (Fig. 8) at an
evolutionary highly conserved position, as demonstrated by
the alignment of mammalian mitochondrial tRNA genes
(Table 4).
Moreover, the transition corresponds exactly to the socalled MELAS-mutation at np 3243 of the human mtDNA
(Fig. 8) [17]. Four out of the eleven affected Yorkshire
Terriers were carriers of this A2691G mutation that was
absent in our preliminary control group, consisting of 11
individuals. However, enlargement of our control group
(n = 42) revealed the A2691G transition in three healthy
dogs (3/42, 7.1%).
Other relevant mtDNA mutations were ruled out by
sequencing, while major rearrangements of the mtDNA,
such as deletions and duplication, were excluded by
Southern blot analysis of DNA extracted from muscle
tissue.
Discussion
Fig. 7 Four out of 11 SNE-affected Yorkshire terriers and 2 out of 42
control dogs had a A-to-G-transition at? position np2691 coding for
tRNA-Leu(UUR)
During the past two decades, our knowledge of the
molecular mechanisms underlying mitochondrial disorders
in humans has increased considerably. It has also been
proposed that certain neurodegenerative diseases in
domestic animals may resemble human mitochondriopathies. However, in most cases there has been insufficient
enzymological and molecular evidence to convincingly
prove this assumption [1–3, 13, 20, 34]. To close one of
Fig. 8 Canine A2691G
transition corresponds exactly to
human MELAS-mutation
A3243G. Both mutations are
confined to the evolutionary
highly conserved D-loop of the
mitochondrial tRNA for leucine.
However, note the different
number of Watson–Crick pairs
at the D-stem (red lines).
Replacement of A to G within
the human tRNA results in the
palindrome sequence GGGCCC
(green rectangle) that is not
flanked by a Watson–Crick pair
at the D-stem. Thereby, a
dimerisation of two mutated
human tRNAs becomes possible
which is likely to interfere with
the aminoacid binding
123
A
A
A
CCCTAAT
CCCTAAT
CCTCT
CCTCT
TTCAAAT
TTCAACT
AGAGG
AGAGG
TTCCC
TTACC
CCTTG
CTTTA
CTTAAAC
CTTAAAA
TAAGA
TAAAA
G
G
TTGC
TTGC
AGCCAAGTAA
AGACCGGTAA
GCAG
GCAG
ATTAGGG
ATTAAGG
ATTAAGG
Rattus norvegicus
Sus scrofa
TG
TG
A
TCTTAAC
CCCTAAT
CCTCT
CCTCT
TTCAATT
TTCAAAT
AGAGG
AGAGG
CAGTC
TTCCC
CCTTG
CTTTA
CTTAAAA
CTTAAAC
TAAGA
TAAAA
A
G
TTGC
TCGC
AGCCCGGTAA
AGCCAGGAAA
GCAG
GCAG
GTTAAGA
Mus musculus
TG
A
Homo sapiens
TG
TCTTAAC
CCTCT
TTCAATT
AGAGG
TAGTC
CTTTA
CTTAAAA
TAAAA
A
TCGC
AGCCCGGTAA
GCAG
GTTAAGA
Gorilla gorilla
TG
CCTTAAC
CCCTAAC
CCTCT
CCTCT
TTCAAAT
TTCAATT
AGAGG
AGAGA
TATCC
CTATC
CTTTA
TTTTA
CTTAAAC
CTTAAAC
TAAAA
TAAAA
A
G
CTGC
TTGC
AGCCCGGTAA
GGCCCGGTAA
GCAG
GCAG
GTTAGGG
Canis familiaris
TG
GTTAAGG
Bos taurus
TG
ACC-stem
T-stem
T-loop
T-stem
AC-stem
AC-loop
AC-stem
D-stem
D-Loop
D-stem
ACC-stem
tRNA-LeuUUR
Table 4 Alignment of different mammalian sequences for the tRNA-Leu gene (abridged version)
A
Acta Neuropathol (2009) 118:697–709
A
706
123
these gaps, we investigated the enzymatic and genetic
background of a Leigh-like encephalopathy in Yorkshire
Terriers. A similar subacute necrotising brain disease had
been observed previously in Alaskan Huskies, which was
reported to mirror human Leigh syndrome (LS) in terms of
both clinical presentation and neuropathology [3, 34].
Similar to the human counterpart, affected Huskies showed
a V-shaped bilateral and symmetrical malacia in basal
nuclei, thalamus and brain stem that crosses grey:whiteboundaries [3]. As in the Yorkshire Terriers (Fig. 9), the
grey matter of the mid-thalamus was most severely rarified
and often replaced by fluid filled cavities.
From the diencephalon the process extended rostrally to
the extrapyramidal basal nuclei of the striatum and caudally to the tegmentum of the medulla oblongata. Amongst
these areas, the malacia appeared contiguous in severely
and chronically affected Huskies, whereas, even at late
stages, the lesions in the Yorkshire Terriers were interrupted by interposition of healthy appearing or less affected
parenchyma.
Apart from these differences, relating to the degree or
stage of brain damage, the Yorkshire Terrier SNE does not
involve the caudal colliculi, which were commonly affected in the reported Huskies. Further spatial differences
were not encountered. In turn, cerebrocortical infarcts are a
consistent feature of SNE in both breeds and they can be
observed in human Leigh-patients [5, 29]. Throughout both
species and canine breeds the subcortical white matter is
affected to a far lesser extent [10, 29].
Another common feature in SNE and LS is the patchy
Purkinje cell (PC) degeneration. The combination of PC
loss and damage to the precerebellar olivary nuclei, postulated for affected humans [5], was seen in one Yorkshire
Terrier only and data from Huskies [3] do not suggest a
nuclear damage at all. Hence, the causative role of loss of
climbing fibres in development of excitotoxic PC death in
LS [5] has to be questioned.
Apart from the slight variations mentioned above, the
spatial distribution pattern of both canine forms of SNE is
very similar. The investigated Alaskan Huskies only differ
from the Yorkshire terriers in the chronicity of the pathological features, which may be explained by the rapidly
progressive clinical course in the latter. The monophasic
and fatal disease development in Yorkshire Terriers
contrasts heavily with the chronic, often static and remittent-relapsing courses that may even remain sublethal in
Alaskan Huskies [3]. One sled dog presented to us (SM)
had a survival time of more than 8 years, defined as the
interval between clinical onset and neurological decompensation (unpublished data). Most of the Yorkshire
Terriers reported here had to be euthanised due to severe
deterioration within a few weeks to months after clinical
diagnosis. It is therefore very likely that the uniform
Acta Neuropathol (2009) 118:697–709
707
Fig. 9 Summary of the spatial
lesion pattern in the eleven
SNE-affected Yorkshire
Terriers
neuropathological picture seen in SNE results from a
phenotypic convergence rather than an identical molecular
disease mechanism.
On the other hand, the monophasic and invariably rapid
SNE course in Yorkshire terriers may facilitate the monitoring of therapeutic efficacy in translational research.
In Alaskan Huskies, the presumptive diagnosis of a
mitochondrial defect was based on the neuropathological
resemblance to LS [34]. Preliminary electron microscopic
investigations revealed unspecific mitochondrial changes
[34]. Likewise, ultrastructural findings in SNE-affected
Yorkshire Terriers were limited to mild mitochondrial
abnormalities that, on their own, would not be sufficient to
state a mitochondrial disease.
However, the enzymological analysis of two muscle
biopsies uncovered a significantly compromised mitochondrial respiratory chain in these animals. A combined
defect of more than one respiratory chain complex is highly
suggestive of a genetic abnormality that impairs mitochondrial protein synthesis, rather than a mutation in
structural proteins, such as individual complex subunits or
assembly factors [26].
In humans, combined defects have been observed in
association with mitochondrial tRNA-mutations, and,
conversely, mutations in tRNA-genes frequently cause
encephalopathies such as LS or MELAS (Mitochondrial
Encephalomyopathy Lactic Acidosis and Stroke-like episodes) [4, 18, 21, 33].
Sequencing of all canine mitochondrial tRNA genes
revealed an A to G transition at np 2691 (an evolutionary
conserved position within the mitochondrial tRNALeu(UUR) gene) in four affected dogs (Table 4). This
nucleotide exchange corresponds exactly to the so-called
MELAS mutation (np3243) of human mtDNA [17], which
has been associated with a number of fatal multisystemic
disorders such as MELAS, MERFF (Myoclonus Epilepsy
with Ragged Red Fibres), MELAS plus MERFF, PEO
(Progressive External Ophthalmoplegia), MDM (Mitochondrial Diabetes Mellitus), HCM (Hypertrophic
Cardiomyopathy) and LS [4, 9, 12, 22, 23, 25, 31, 32]. The
causative link between A2691G and SNE, however,
became unlikely after evidence of the same exchange in a
set of healthy dogs of the control group, and after the
haplotype lineage analysis sorted all carriers in one
phylogenetic clade.
The fact that the same, evolutionary conserved mutation
within dogs appears to be neutral whereas in humans it
leads to a severe systemic disorder can be explained by
differences regarding the secondary cloverleaf structure
and the tertiary L-shape of the human and canine tRNALeu(UUR).
The three-dimensional conformation of the human
tRNA for leucine is stabilised by just two Watson-Crickpairs in the D-stem that cannot counteract a dimerisation of
two mutated tRNAs with identical palindrome sequences
(…GGGCCC…) (Fig. 8). Aminoacylation of the mutated
and dimerised tRNAs is significantly attenuated and is,
therefore, responsible for the loss of function [14, 17, 27,
36, 37]. In contrast, the canine tRNA D-stem is stabilised
by two additional Watson-Crick pairs (Fig. 8) and probably
retains its normal L-shaped tertiary structure even in
presence of A2691G. Amongst mammalian species, the
D-stem also contains four Watson–Crick pairs in felids and
rhinoceros, three pairs in the majority of remaining
123
708
mammals and two only in humans and non-human primates such as bonobos and gorillas [15]. These structural
characteristics may provide an explanation for the different
impact of the tRNA-Leu(UUR) point-mutation in humans
and dogs and might predict in which species corresponding
tRNA-Leu(UUR) mutations may be pathogenic.
Considering further genetic causes of SNE, we ruled out
mutations in other mitochondrial tRNA genes, the ATP6
gene and major deletions and duplications of the mtDNA.
Hence, SNE in Yorkshire Terriers may be caused by a
nuclear gene defect. Likewise, nuclear mutations are
responsible for the majority of LS in people. Impairment of
the mitochondrial transcription and translation apparatus,
which is entirely nuclear encoded, also can lead to combined respiratory chain deficiencies. In particular, mutation
and dysfunction of the mitochondrial elongation factor
EFG1 has been associated with early-onset LS and combined respiratory chain defects [6, 30].
Taken together, canine SNE exhibits major neuropathological characteristics of human LS. To date, is has
been observed in Alaskan Huskies and Yorkshire terriers.
In the latter, SNE is one of the most prevalent neurodegenerative diseases and exhibits a consistent clinical course
and morphology. SNE-affected Yorkshire terriers are easily recognised due to conspicuous gait abnormalities and
behavioural changes with/without seizures and visual deficits. As in humans and Huskies, a clinical diagnosis
reliably can be made upon MR imaging, showing symmetric lesions affecting the basal nuclei, thalamus and
brain stem [29, 34]. Further work is necessary to clarify the
genetic background and the resemblance to other canine
SNEs and LS. Identification of the disease-causing mutation is indispensable for detecting carriers, which is
mandatory for both establishing a feasible animal model
and elimination of the disease from the dog population.
Acknowledgments We thank Bettina Treske and Karin Stingl for
excellent technical assistance. This work was supported by the
Deutsche Forschungsgesellschaft to M.FB. (Ba1438/3 and 4) and to
SH. (Ho2374/1-1).
References
1. Braund KG (2003) Degenerative disorders of the central nervous
system. International Veterinary Information Service. Ithaca.
http://www.ivis.org.
2. Brenner O, de Lahunta A, Cummings JF, Summers BA,
Monachelli M (1997) A canine encephalomyelopathy with
morphological abnormalities in mitochondria. Acta Neuropathol
(Berl) 94:390–397
3. Brenner O, Wakshlag JJ, Summers BA, de Lahunta A (2000)
Alaskan Husky encephalopathy—a canine neurodegenerative
disorder resembling subacute necrotizing encephalomyelopathy
(Leigh syndrome). Acta Neuropathol (Berl) 100:50–62
123
Acta Neuropathol (2009) 118:697–709
4. Campos Y, Martin MA, Lorenzo G, Aparicio M, Cabello A,
Arenas J (1996) Sporadic MERRF/MELAS overlap syndrome
associated with the 3243 tRNA(Leu(UUR)) mutation of mitochondrial DNA. Mus Nerv 19:187–190
5. Cavanagh JB, Harding BN (1994) Pathogenic factors underlying
the lesions in Leigh’s disease: tissue responses to cellular energy
deprivation and their clinico- pathological consequences. Brain
117(Pt 6):1357–1376
6. Coenen MJ, Antonicka H, Ugalde C et al (2004) Mutant mitochondrial elongation factor G1 and combined oxidative
phosphorylation deficiency. N Engl J Med 351:2080–2086
7. Darin N, Oldfors A, Moslemi AR, Holme E, Tulinius M (2001)
The incidence of mitochondrial encephalomyopathies in childhood: clinical features and morphological, biochemical, and DNA
anbormalities. Ann Neurol 49:377–383
8. DiMauro S, De Vivo DC (1996) Genetic heterogeneity in Leigh
syndrome. Ann Neurol 40:5–7
9. Fabrizi GM, Cardaioli E, Grieco GS et al (1996) The A to G
transition at nt 3243 of the mitochondrial tRNALeu(UUR) may
cause an MERRF syndrome. J Neurol Neurosurg Psychiatry
61:47–51
10. Filosto M, Tomelleri G, Tonin P et al (2007) Neuropathology of
mitochondrial diseases. Biosci Rep 27:23–30
11. Fischer JC, Ruitenbeek W, Gabreels FJ et al (1986) A mitochondrial encephalomyopathy: the first case with an established
defect at the level of coenzyme Q. Eur J Pediatr 144:441–444
12. Goto Y, Nonaka I, Horai S (1990) A mutation in the tRNA
(Leu)(UUR) gene associated with the MELAS subgroup of
mitochondrial encephalomyopathies. Nature 348:651–653
13. Gruber AD, Wessmann A, Vandevelde M, Summers BA, Tipold
A (2002) Mitochondriopathy with regional encephalic mineralization in a Jack Russell Terrier. Vet Pathol 39:732–736
14. Hao R, Yao YN, Zheng YG, Xu MG, Wang ED (2004) Reduction
of mitochondrial tRNALeu(UUR) aminoacylation by some
MELAS-associated mutations. FEBS Lett 578:135–139
15. Helm M, Brule H, Friede D, Giege R, Putz D, Florentz C (2000)
Search for characteristic structural features of mammalian mitochondrial tRNAs. RNA 6:1356–1379
16. Holt IJ, Harding AE, Morgan-Hughes JA (1988) Deletions of
muscle mitochondrial DNA in patients with mitochondrial
myopathies. Nature 331:717–719
17. Kern AD, Kondrashov FA (2004) Mechanisms and convergence
of compensatory evolution in mammalian mitochondrial tRNAs.
Nat Genet 36:1207–1212
18. Koga Y, Akita Y, Takane N, Sato Y, Kato H (2000) Heterogeneous presentation in A3243G mutation in the mitochondrial
tRNA(Leu(UUR)) gene. Arch Dis Child 82:407–411
19. Leigh D (1951) Subacute necrotizing encephalomyelopathy in an
infant. J Neurol Neurosurg Psychiatry 14:216–221
20. Li FY, Cuddon PA, Song J et al (2006) Canine spongiform leukoencephalomyelopathy is associated with a missense mutation
in cytochrome b. Neurobiol Dis 21:35–42
21. McFarland R, Clark KM, Morris AA et al (2002) Multiple neonatal deaths due to a homoplasmic mitochondrial DNA mutation.
Nat Genet 30:145–146
22. Moraes CT, Ciacci F, Silvestri G et al (1993) Atypical clinical
presentations associated with the MELAS mutation at position
3243 of human mitochondrial DNA. Neuromuscul Disord 3:43–
50
23. Obayashi T, Hattori K, Sugiyama S et al (1992) Point mutations
in mitochondrial DNA in patients with hypertrophic cardiomyopathy. Am Heart J 124:1263–1269
24. Rahman S, Blok RB, Dahl HH et al (1996) Leigh syndrome:
clinical features and biochemical and DNA abnormalities. Ann
Neurol 39:343–351
Acta Neuropathol (2009) 118:697–709
25. Silvestri G, Bertini E, Servidei S et al (1997) Maternally inherited
cardiomyopathy: a new phenotype associated with the A to G AT
nt.3243 of mitochondrial DNA (MELAS mutation). Mus Nerv
20:221–225
26. Smeitink J, van den Heuvel L, DiMauro S (2001) The genetics
and pathology of oxidative phosphorylation. Nat Rev Genet
2:342–352
27. Sohm B, Frugier M, Brule H, Olszak K, Przykorska A, Florentz C
(2003) Towards understanding human mitochondrial leucine
aminoacylation identity. J Mol Biol 328:995–1010
28. Southern EM (1975) Detection of specific sequences among
DNA fragments separated by gel electrophoresis. J Mol Biol
98:503–517
29. Tanji K, Kunimatsu T, Vu TH, Bonilla E (2001) Neuropathological features of mitochondrial disorders. Semin Cell Dev Biol
12:429–439
30. Valente L, Tiranti V, Marsano RM et al (2007) Infantile
encephalopathy and defective mitochondrial DNA translation in
patients with mutations of mitochondrial elongation factors EFG1
and EFTu. Am J Hum Genet 80:44–58
709
31. van den Ouweland JM, Lemkes HH, Ruitenbeek W et al (1992)
Mutation in mitochondrial tRNA(Leu)(UUR) gene in a large
pedigree with maternally transmitted type II diabetes mellitus and
deafness. Nat Genet 1:368–371
32. Verma A, Moraes CT, Shebert RT, Bradley WG (1996) A
MERRF/PEO overlap syndrome associated with the mitochondrial DNA 3243 mutation. Neurology 46:1334–1336
33. Vilarinho L, Maia C, Coelho T, Coutinho P, Santorelli FM (1997)
Heterogeneous presentation in Leigh syndrome. J Inherit Metab
Dis 20:704–705
34. Wakshlag JJ, de Lahunta A, Robinson T et al (1999) Subacute
necrotising encephalopathy in an Alaskan husky. J Small Anim
Pract 40:585–589
35. Wallace DC, Singh G, Lott MT et al (1988) Mitochondrial DNA
mutation associated with Leber’s hereditary optic neuropathy.
Science 242:1427–1430
36. Wittenhagen LM, Kelley SO (2002) Dimerization of a pathogenic
human mitochondrial tRNA. Nat Struct Biol 9:586–590
37. Wittenhagen LM, Kelley SO (2003) Impact of disease-related
mitochondrial mutations on tRNA structure and function. Trends
Biochem Sci 28:605–611
123