Formation of polyglutamine inclusions in non-CNS tissue ,

advertisement
 1999 Oxford University Press
Human Molecular Genetics, 1999, Vol. 8, No. 5
813–822
Formation of polyglutamine inclusions in non-CNS
tissue
Kirupa Sathasivam+, Carl Hobbs1,+, Mark Turmaine2, Laura Mangiarini,
Amarbirpal Mahal, Fabien Bertaux, Erich E. Wanker3, Pat Doherty1,
Stephen W. Davies2 and Gillian P. Bates*
Medical and Molecular Genetics, GKT Medical and Dental School, King’s College, 8th Floor, Guy’s Tower, Guy’s
Hospital, London SE1 9RT, UK, 1Experimental Pathology, GKT Medical and Dental School, King’s College, 4th
Floor, Hodgkin Building, Guy’s Hospital, London SE1 9RT, UK, 2Department of Anatomy and Developmental
Biology, University College London, Gower Street, London WC1E 6BT, UK and 3Max Planck Institute for
Molecular Genetics, Ihnestrasse 73, Berlin Dahlem, Germany
Received November 19, 1998; Revised and Accepted February 11, 1999
Huntington’s disease (HD) is one of a class of inherited
progressive neurodegenerative disorders that are
caused by a CAG/polyglutamine repeat expansion. We
have previously generated mice that are transgenic for
exon 1 of the HD gene carrying highly expanded CAG
repeats which develop a progressive movement disorder and weight loss with similarities to HD. Neuronal
inclusions composed of the exon 1 protein and ubiquitin are present in specific brain regions prior to onset
of the phenotype, which in turn occurs long before
specific neurodegeneration can be detected. In this report we have extended the search for polyglutamine inclusions to non-neuronal tissues. Outside the central
nervous system (CNS), inclusions were identified in a
variety of post-mitotic cells. This is consistent with a
concentration-dependent nucleation and aggregation
model of inclusion formation and indicates that brainspecific factors are not necessary for this process. To
possibly gain insights into the wasting that is observed
in the human disease, we have conducted a detailed
analysis of the timing and progression of inclusion
formation in skeletal muscle and an investigation into
the cause of the severe muscle atrophy that occurs in
the mouse model. The formation of inclusions in nonCNS tissues will be particularly useful with respect to
in vivo monitoring of pharmaceutical agents selected
for their ability to prevent polyglutamine aggregation
in vitro, without the requirement that the agent can
cross the blood–brain barrier in the first instance.
INTRODUCTION
Huntington’s disease (HD) is one of a family of inherited, late
onset, neurodegenerative disorders that are caused by a polygluta-
mine (polyQ) expansion. The other members of this group
include dentatorubral pallidoluysian atrophy (DRPLA), spinal
and bulbar muscular atrophy (SBMA) and the spinocerebellar
ataxias (SCA1–3, 6 and 7) (1). Comparison of the genetics and
molecular biology of these diseases shows that they have a
number of features in common, namely a dominant pattern of
inheritance, anticipation and the broadly comparable sizes of the
normal and expanded repeat ranges. The proteins containing
these polyQ expansions are unrelated, in most cases of unknown
function and have extensively overlapping expression patterns
(1,2).
We have previously generated a mouse model of HD that is
transgenic for exon 1 of the human HD gene, containing highly
expanded CAG repeats, under the control of the HD promoter.
Three lines [R6/1, (CAG)115; R6/2, (CAG)145; R6/5,
(CAG)128–156] develop a progressive neurological phenotype,
the symptoms of which include a movement disorder and weight
loss with similarities to HD (3,4). On the basis of home cage
behaviour, line R6/2 has an age of onset of ∼2 months. However,
a more recent detailed assessment of motor function has shown
that deficits can be detected at between 5 and 6 weeks and that the
age at which a phenotype is apparent is dependent upon the
sensitivity of the test and the difficulty of a particular task (5,6).
The disease progresses rapidly such that mice are rarely kept
beyond 14 weeks of age.
Immunocytochemistry with antibodies to the truncated exon 1
protein and to ubiquitin identified neuronal intranuclear inclusions in brains from all symptomatic lines (7). At the ultrastructural level, inclusions can be observed in the absence of immunolabelling as pale staining granular and fibrillar structures devoid of
a membrane. Only a single nuclear inclusion is detected per
neuron. In line R6/2, inclusions are first identified in the cortex
at 3.5 weeks and in the striatum at 4.5 weeks, prior to the onset
of symptoms (7). In contrast, selective cell death is not observed
in these structures until 14–17 weeks of age, suggesting that the
symptoms are due to a neuronal dysfunction rather than as a result
of neurodegeneration (M. Turmaine, A. Raza, A. Mahal, L.
Mangiarini, G.P. Bates and S.W. Davies, submitted for publica-
*To whom correspondence should be addressed. Tel: +44 171 955 4485; Fax: +44 171 955 4444; Email: gillian.bates@kcl.ac.uk
+These authors contributed equally to this work
804
Human Molecular Genetics, 1999, Vol. 8, No. 5
tion). Coincident with the first appearance of inclusions, a
decrease in brain weight is observed (7). It has not yet been
determined whether this is the consequence of cell loss, cell
shrinkage, loss of processes or a reduction in the intercellular
volume. In the two lines showing a slower progression, R6/1 and
R6/5, inclusions are also observed in neurites and occasionally in
astrocytes (M. Turmaine, A. Raza, A. Mahal, L. Mangiarini, G.P.
Bates and S.W. Davies, submitted for publication).
Insights into the structure of the polyQ inclusions have come
from in vitro analysis of GST–huntingtin exon 1 fusion proteins
(8). It was found that GST fusion proteins with (CAG)83 and
(CAG)121 spontaneously formed highly insoluble aggregates
with a fibrillar morphology. Removal of the GST tag from
proteins with (CAG)51 also resulted in amyloid-like fibrils that
showed green birefringence when stained with Congo red and
examined under polarized light. Aggregation was not observed
with proteins containing (CAG)20 or (CAG)30 repeats in the
non-pathogenic range. This is consistent with the proposal by
Perutz et al. (9) that polyQ stretches can form polar zippers via
hydrogen bonding in a cross-β-pleated amyloid-like structure. A
detailed analysis of the concentration and polyQ length dependence of aggregation has shown that the ability to form these highly
ordered structures closely correlates with the polyQ pathogenic
threshold (10).
Neuronal inclusions are now clearly established as the pathological signature of polyQ disease and have been identified in
post-mortem brains from HD (11–13), SCA1 (14), SCA3 (15),
SCA7 (16), DRPLA (12,17) and SBMA patients (18). In all cases
they are composed of the specific polyQ-containing protein and
ubiquitin. Inclusions in affected neurons of SCA1 patients and an
SCA1 transgenic mouse model stain positive for antibodies that
detect the 20S proteasome and the molecular chaperone
HDJ-2/HSDJ (19). Similarly, inclusions in HD patient brains and
transgenic mice contain the 20S proteasome and several components of the 19S and 11S activation complexes (S.W. Davies,
unpublished data).
The factors that govern the cell specificity of both inclusion
formation and neurodegeneration for each of the polyQ diseases
are not understood. The ectopic expression of a polyQ expansion
within the hprt gene, introduced by gene targeting, has demonstrated that polyQ tracts are capable of aggregation in a protein
context other than that already associated with disease (20). These
mice developed a progressive neurological disorder associated
with neuronal intranuclear inclusions, although neurodegeneration was absent. This clearly demonstrates that interactions
specific to the polyQ diseases are not necessary for the formation
of polyQ aggregation in neuronal populations.
In this report, we describe the distribution of polyQ inclusions
in tissues outside the central nervous system (CNS) in the R6
transgenic mouse lines. We have conducted a detailed analysis of
the timing and formation of inclusions in non-neuronal tissues in
relation to organ shrinkage/atrophy. We have carried out a
detailed investigation into the basis of the observed muscle
atrophy in order to: (i) provide insights into the loss of muscle
bulk observed in HD (21,22); and (ii) provide a parallel analysis
of the formation and consequence of neuronal inclusions. The
identification of inclusions in non-CNS tissues will be useful for
monitoring in vivo trials of pharmaceutical agents designed to
prevent polyQ aggregation.
RESULTS
Detection of polyQ inclusions in non-CNS tissues
In order to identify inclusions in non-CNS tissues, an R6/2 male
mouse at 14 weeks of age was perfusion fixed and individual
organs were paraffin wax embedded, sectioned and processed for
immunohistochemistry using the CAG53b and anti-ubiquitin
antibodies. Tissues and cell types in which inclusions were
detected are summarized in Table 1 and examples are illustrated
in Figure 1. In all cases, inclusions were only detected in cell
nuclei; in most cases a single inclusion was present although some
nuclei contained up to three inclusions (especially in cells in the
liver, kidney and skeletal muscle). Inclusions were found in the
muscle fibres of skeletal and cardiac muscle but were not present
in the smooth muscle of the stomach wall. In the adrenal glands,
inclusions were restricted to the medulla and the reticula layer of
the cortex and they were not present in the glomerulosa and
fasciculata layers. In the pancreas they were restricted to the islets
of Langerhans, in the liver to the hepatocytes and in the kidney
were identified at a low level in the tubular, interstitial and
glomerular cells. They were found in the neuronal ganglion cells
of the myenteric plexus of the stomach wall and Meissner’s
plexus of the duodenum. Inclusions were absent from skin,
spleen, testis, coagulation gland, the mucosal cells of the stomach,
duodenum, colon and buccal cavity and the white adipose tissue.
Table 1. Intranuclear inclusion-positive cell types
Tissue
Inclusion-positive cell type
Skeletal muscle
Muscle fibres
Heart
Cardiac muscle
Liver
Hepatocytes
Adrenal glands
Medullary and reticula cells
Pancreas
Islets of Langerhans
Kidney
Tubular/interstitial/glomerular cells
Stomach wall
Myenteric plexus
Duodenum
Meissner’s plexus
We had been specifically interested in investigating the
presence of inclusions in tissues accessible by biopsy which could
have an important application in monitoring clinical trials. Of the
possible candidate tissues, skin, mucosal cells and skeletal
muscle, inclusions were only identified in muscle.
Ultrastructure of skeletal muscle inclusions
An ultrastructural analysis of nuclear inclusions in skeletal
muscle was performed. The inclusions appear as pale staining
granular and fibrillar structures identical to those previously
observed in neurons (Fig. 2). This electron microscopical analysis
suggested that at least in some cases the appearance of more than
one inclusion per nucleus at the light microscopy level could be
accounted for by single inclusions in adjacent nuclei.
Shrinkage/atrophy of organs in R6/2 transgenic mice
A series of R6/2 male mice had previously been used to determine
the timing of the progressive decrease in brain weight with respect
805
Human
Genetics,
1999,
8, No.
NucleicMolecular
Acids Research,
1994,
Vol. Vol.
22, No.
1 5
805
Figure 1. Intranuclear inclusions in non-CNS tissues. Sections were from an 8-week-old (a) and 14-week-old (b–f) R6/2 mouse immunoprobed with anti-ubiquitin
antibodies (a) and CAG53b anti-huntingtin antibodies (b–f). Intranuclear inclusions are present in skeletal muscle (a), pancreas (b), liver (c), adrenal gland (d) and
kidney (e) but not in spleen (f). Nuclei were counterstained with methyl green. Scale bar, 10 µm.
Figure 2. Ultrastructural analysis of nuclear inclusions in skeletal muscle from a 12-week-old R6/2 mouse. (a) Muscle fibre nucleus containing a nuclear inclusion
(arrow) adjacent to the more darkly stained nucleolus. (b) Adjacent nuclei each containing a single inclusion. The morphology of the muscle fibres, including the
mitochondria and motor end plate, are perfectly normal. Scale bar, 2 µm.
to the loss in body weight (7). This comprised seven or eight mice
at each time point ranging from 2 to 12 weeks. It was found that
the decrease in brain weight occurred after 4 weeks, considerably
earlier than the decrease in body weight, which commenced at ∼7
weeks. This study has now been extended to determine the onset
and extent of atrophy in other organs. To this end, the heart, liver,
spleen, kidneys, thymus, testes and quadriceps and soleus
muscles were dissected from the original series of male R6/2 mice
and weighed (Fig. 3). The most dramatic atrophy occurred in the
testes, which, like brain, began to decrease in weight after 4
weeks. By 8 weeks, the transgenic testes were 88% of the weight
of the non-transgenic testes, after which point they plummeted in
weight to 15% at 12 weeks. This testicular atrophy most likely
underlies the loss of fertility that is encountered in the R6/2 males.
Approximately 50% of R6/2 males are sterile, and of those that
do breed, all are sterile by 8 weeks of age. The quadriceps and
soleus muscles appear to develop normally until 6 weeks of age,
after which time they atrophy to 42 and 47% of the weight of those
dissected from the non-transgenic mice at 12 weeks. Similarly,
weight loss begins in heart and kidney at 6 weeks and in liver at
806
Human Molecular Genetics, 1999, Vol. 8, No. 5
Figure 3. Progressive organ atrophy in line R6/2. The quadriceps and soleus muscles (a), testes (b), liver (c), kidneys (d) and spleen (e) were dissected and weighed
from a series of R6/2 male mice ranging in age from 2 to 12 weeks. Weight loss was detected in testes from 4 weeks, both muscles and kidney from 6 weeks and liver
from 8 weeks. Transgenic tissue is represented by a solid line and non-transgenic tissue by a dotted line.
8 weeks, decreasing to 55, 50 and 75% of the non-transgenic
organs at 12 weeks, respectively.
Appearance of nuclear inclusions in non-CNS tissues
An analysis of the onset and progression of inclusion formation
was conducted in longitudinal sections taken from snap-frozen
quadriceps muscle of R6/2 mice that were 2, 4, 6, 8, 12 and 14
weeks of age. In each case, three transgenic mice and a single
non-transgenic littermate control were processed at each time
point. Inclusions were not present at 4 weeks but could be seen
by 6 weeks of age. Overall, a focal distribution of nuclei
containing inclusions was observed, making quantification
difficult. To gain an estimate of the frequency of inclusions, the
number of inclusions present in ∼300 nuclei in adjacent fields in
8 µm longitudinal sections was counted for each mouse. This
suggested that the number of inclusions increased steadily from
6 (5 ± 1%) to 12 weeks (18 ± 3%). However, this analysis
provides only an approximation and a more accurate figure will
require extensive quantification over several muscle types.
To determine the appearance of inclusions in the other tissues,
two transgenic male mice and one non-transgenic littermate were
perfusion fixed and paraffin embedded sections were prepared at
4, 6, 8 and 10 weeks of age. In all cases, inclusions were not
present at 4 weeks and were readily apparent by 8 weeks. At
6 weeks they were more prominent in pancreas, kidney and
adrenals than in liver.
807
Human
Genetics,
1999,
8, No.
NucleicMolecular
Acids Research,
1994,
Vol. Vol.
22, No.
1 5
807
Investigation into the basis of the muscle atrophy in
lines R6/1 and R6/2
comparatively uniform atrophy of all muscle fibres had occurred
in the transgenics.
In order to shed light on the loss of muscle bulk observed in the
human disease, we set out to uncover the basis of the muscle
atrophy that occurs in the R6 transgenic lines. Transverse sections
from the quadriceps muscle of 14-week-old R6/2 and 15-monthold R6/1 mice were cut from snap-frozen tissue and stained using
haematoxylin and eosin (H&E), haematoxylin–van Gieson
(HVG) and the periodic acid–Schiff (PAS) assay. Acid phosphatase and NADH-tetrazolium reductase (NADH-TR) enzyme
activities were measured and immunohistochemistry was performed with the anti-N-cam antibody (H28).
There was little evidence of myopathy as indicated by muscle
fibre regeneration. If present, internal nuclei can be detected
(particularly noticeable in the PAS stained sections) (Fig. 4).
Whilst the frequency of internal nuclei was slightly elevated (4%
in R6/1 and 1% in R6/2 as compared with 1 and 0.3% in their
respective controls), this was not apparent to an extent that would
be indicative of myopathy. Similarly, vesicular nuclei or fibre
splitting could not be detected, which would also provide
indications of myopathy. Regenerating fibres are uniformly
immunoreactive for the anti-N-cam antibody and only 2% in
R6/1 and 0.3% in R6/2 were evident as compared with 0.7 and
0.3% in their respective controls. The muscle atrophy is not
caused by denervation. Examination of all sections revealed no
evidence of a focal atrophy and the motor endplate distribution,
as demonstrated by anti-N-cam immunohistochemistry, also
appeared normal (Fig. 4). We could find no evidence of cellular
infiltration in response to necrotic fibres on the basis of acid
phosphatase activity or of the presence of granular or basophilic
fibres on the H&E stained sections. The HVG stained sections
showed no fibrosis, as is evident from the proliferation of the
endomysial or perimysial connective tissue, or of the presence of
myofibril central cores (Fig. 4). The PAS stained tissue showed
no dramatic increase in the glycogen content of the transgene
muscles or abnormalities in the basal membrane. In contrast, the
transgene sections showed little staining, possibly reflecting the
emaciation that had occurred in these mice. The PAS staining can
distinguish type 1 and type 2 fibres on the basis of glycogen
content and this indicated that there was probably no selective
atrophy or specific fibre type grouping (Fig. 4), although this was
pursued in more depth as outlined below.
Ultrastructural analysis
Muscle fibre type analysis
To further examine the possibility of a specific fibre type atrophy
or fibre type grouping, NADH-TR activity was measured and
immunohistochemistry was performed with antibodies to the
specific myosin heavy chains present in the various muscle fibre
types. Sections from the quadriceps muscle of a 14-week-old
R6/2 mouse and 15-month-old R6/1 mouse were cut from
snap-frozen tissue as before and the immunoreactivity to
anti-type 2A myosin is shown in Figure 5. A morphometric
analysis was also conducted. This analysis demonstrated that
there was no fibre type grouping. The lesser fibre diameter was
measured for the sections through the quadriceps muscle from the
14-week-old R6/2 mouse and from the extensor digitorum longus
muscle of the R6/1 mouse. The normal distribution evident in
these graphs (Fig. 5e and f) shows quite clearly that a
Ultrastructural analysis of R6/2 quadriceps muscle at 14 weeks
showed no evidence of myopathy or neuropathy. The only feature
that distinguished it from non-transgenic muscle was the presence
of intranuclear inclusions (Fig. 2). In contrast, the ultrastructure
of skeletal muscle from the 15 month R6/1 mouse did show
evidence of degeneration (Fig. 6). Non-specific ultrastructural
changes included enlarged mitochondria with a well-preserved
proliferation of cristae, less elaborate motor endplates with a
reduction in the number of post-synaptic folds and endplateassociated fibrous structures.
DISCUSSION
We have extended the search for polyQ inclusions to the
non-CNS tissues of a transgenic mouse model of HD. Nuclear
inclusions were found in post-mitotic cells of a number of tissues,
including skeletal and cardiac muscle, liver, kidney, adrenal
glands and pancreas. In most cases, the presence of inclusions
correlates with a progressive decrease in size of the respective
organ. An extensive histological analysis of skeletal muscle
found no evidence of myopathy or neuropathy, but instead that the
decrease in muscle bulk was caused by a uniform shrinkage
across all muscle fibre types.
Inclusion formation
The molecular pathogenesis of polyQ disease is triggered by the
gain of function that the polyQ expansion imparts to the protein
in question. The gain of this new function must correlate precisely
with the threshold at which the repeat expands into the pathogenic
range and any proposed molecular mechanism of pathogenesis
must account for the late onset of the disease. It has previously
been shown by in vitro analysis that the likely gain of function is
the ability of expanded polyQ repeats to form amyloid-like
aggregates (8). More recently, a detailed study of the concentration and polyQ length dependence of the aggregation of polyQ
tracts in the context of the HD protein has shown that aggregation
occurs at a polyQ repeat length that shows a striking correlation
with the pathogenic threshold observed in HD (10).
The nucleation and aggregation model for the formation of
inclusions predicts that an initial slow nucleation step, possibly in
equilibrium with the soluble form of the protein, is followed by
a more rapid aggregation step (23). The predicted kinetics could
account for the late onset of the disease both in humans and in the
transgenic mice. Evidence from (i) immunohistochemistry on
post-mortem brain material with antibodies spanning the huntingtin protein (11–13), (ii) cell culture models transfected with
various portions of huntingtin (24–28) and (iii) in vitro aggregation experiments (8) indicates that it is an N-terminal fragment of
huntingtin that is the precursor of aggregation and that is present
in the inclusions. Therefore, the molecular pathogenesis of the
human disease would require a cleavage/proteolysis step whereas, in the R6 mouse lines, the transgene only codes for the
N-terminus, therefore by-passing this initial step in the disease
pathogenesis. Therefore, in the mice the rate of formation of
aggregates could be predicted to depend on polyQ length and the
concentration of the transgene protein whereas in the human
808
Human Molecular Genetics, 1999, Vol. 8, No. 5
Figure 4. Histological and immunocytochemical analysis of skeletal muscle from lines R6/1 and R6/2. Transverse frozen sections through quadriceps muscle from
a 15-month-old non-transgenic littermate (a, e and i), a 15-month-old R6/1 transgenic (b, f and j), a 14-week-old non-transgenic littermate (c, g and k) and a
14-week-old R6/2 transgenic (d, h and l) were stained for PAS (a–d), immunoprobed with an anti-N-cam antibody (e–h) and stained for HVG (i–l). Nuclei were
counterstained with haematoxylin. Scale bar, 40 µm
disease the cleavage/proteolytic event that generates the Nterminal precursor may provide an additional rate-limiting step.
The precise nature of this fragment and the cleavage/proteolysis
step involved in the disease are unknown.
Distribution of inclusions
The formation of inclusions in neurons is likely to occur as a result
of the terminally differentiated nature of these cells allowing the
critical concentration for aggregation to be reached. A mitotically
active cell might be expected to dilute the concentration of the
precursor molecule at each cell division. However, the identification of inclusions in non-proliferating astrocytes (S.W. Davies,
unpublished data) indicates that they can form in mitotically
active cells. Therefore, should inclusions be present in peripheral
tissues, they could be predicted to occur similarly in post-mitotic
cells and possibly mitotically active cells dependent upon the
level of expression of the protein. An expression profile of
huntingtin outside the CNS has not been studied in depth. In situ
hybridization to human tissues showed expression of the HD gene
to be uniformly low in the acinar, ductal and islet of Langerhans
cells of the pancreas and hepatocytes, bile ducts and connective
tissues of the liver. Developing sperm in testes expressed variable
levels of HD mRNA, with immature spermatogonia expressing
higher levels than maturing spermatids (29). It has not been
possible to determine the level of the transgene protein by
immunohistochemistry in the R6 lines as it cannot be distinguished from the mouse protein with available antibodies.
However, if the level of transgene protein in non-CNS tissues is
similarly low, a distribution of inclusions in post-mitotic cells, as
identified in this report, is consistent with the reliance of inclusion
formation upon the local concentration of the transgene protein.
In all cases, inclusions were identified in nuclei. In neurons, the
normal distribution of huntingtin is reported to be extranuclear
(30–32) and the mechanism by which the transgene protein enters
the nucleus is unknown. The subcellular localization of huntingtin in peripheral tissues has been the subject of much less
extensive scrutiny and a nuclear localization in some cell types
cannot be ruled out. Indeed, a nuclear localization of huntingtin
in cell cultures established from mouse embryonic fibroblasts,
human skin fibroblasts and mouse neuroblastoma cells has
809
Human
Genetics,
1999,
8, No.
NucleicMolecular
Acids Research,
1994,
Vol. Vol.
22, No.
1 5
e
809
f
Figure 5. Type 2A fibre typing and muscle fibre morphometry in R6/1 and R6/2 skeletal muscle. Frozen transverse sections from a 15-month-old non-transgenic
littermate (a), a 15-month-old R6/1 transgenic (b), a 14-week-old non-transgenic littermate (c) and a 14-week-old R6/2 transgenic (d) were immunoprobed with an
anti-myosin heavy chain type 2A antibody. This shows no evidence of type 2A fibre atrophy. Scale bar, 40 µm. The lesser fibre diameter was measured for the extensor
digitorum longus muscle from an R6/1 mouse at 15 months (e) and the quadriceps muscle from an R6/2 mouse at 14 weeks (f). Dotted line, non-transgenic littermates;
solid line, transgenics.
Figure 6. Ultrastructure of degenerative muscle in line R6/1. Non-specific ultrastructural changes which may consist of fibrillar structures (F), changes in motor end
plate morphology (M) (a) reassociated muscle fibres (R) and enlarged mitochondria (b). Scale bar, 2 µm.
previously been documented (33). Still, irrespective of the normal
localization of huntingtin in these cell types, in all cases the
transgene product can clearly enter and accumulate in the
nucleus.
The relationship between neuronal inclusions and
pathology
The phenotype of the R6 lines is caused by neurodysfunction and
not neurodegeneration. In line R6/2, inclusions are present in the
810
Human Molecular Genetics, 1999, Vol. 8, No. 5
cortex and striatum (3–4 weeks) prior to the onset of symptoms
(∼8 weeks), which is long before specific neurodegeneration can
be seen in these brain regions (14–17 weeks). Changes in the
expression levels of specific neurotransmitter receptors can be
detected from as early as 4 weeks (34), providing an insight into
a possible mechanism of dysfunction. Shortly after the onset of
symptoms, inclusions can be detected in the majority of, if not all,
neurons. However, within the lifetime of the R6/2 mouse, cell
death is the consequence of polyQ aggregation in only a small
number of neurons residing in the cortex, striatum and cerebellum
(S.W. Davies, unpublished data).
Immunohistochemical analysis of HD, DRPLA, SCA3 and
SCA7 post-mortem brains has indicated that the distribution of
inclusions overlaps with, but is also wider than, the classically
described neuropathology (11,12,16). Therefore, in the lifetime
of the patient, polyQ aggregation and the eventual presence of an
inclusion, as detected by light microscopy, does not necessarily
lead to the degeneration of the neuron in which it has formed. The
pattern of neurodegeneration may arise from a selective susceptibility to polyQ aggregate toxicity.
The relationship between non-neuronal inclusions and
pathology
Within the skeletal muscle of the R6/2 mice, there was a
remarkable correlation between the appearance of inclusions at
6 weeks and the onset of muscle atrophy. We have conducted an
extensive histological study and found no evidence of myopathy
or neuropathy and at close to end stage the only consistent
difference between the R6/2 muscle and that of the nontransgenic littermates was the size of the muscle fibres. It would
appear that the muscle fibres have undergone a uniform
shrinkage. We are not aware of other diseases in which a loss of
muscle bulk is the result of this uniform atrophy that encompasses
all muscle fibre types. Therefore, as in brain, degenerative
changes are only observed after a protracted time course, in this
case in the form of non-specific ultrastructural changes in a
15-month-old R6/1 mouse. These findings are in contrast to the
muscle pathology observed in other myopathies that are caused
by amyloid inclusions. Sporadic inclusion body myositis (s-IBM)
is characterized by the presence of amyloid inclusions in the
nucleus and cytoplasm of muscle fibres which have been found
to contain β-amyloid, two other epitopes of the amyloid precursor
protein, hyperphosphorylated tau, α1-antichymotrypsin, apolipoprotein E, ubiquitin, presenilin-1 and cellular prion protein
(35). There are three sequential and overlapping aspects of
muscle fibre destruction in s-IBM: an attack on muscle fibres by
cytotoxic T lymphocytes, vacuolar degeneration and a denervation atrophy of the muscle fibres. The hereditary inclusion body
myopathies are similar to the s-IBMs except that they lack the
lymphocytic inflammation (36). Recent transgenic models of
these disorders have demonstrated a complex myopathy (37,38).
Similarly, oculopharyngeal muscular dystrophy is caused by
modest expansions of a GCG/alanine repeat expansion in the
poly(A)-binding protein 2 gene, resulting in intranuclear inclusions within muscle fibres (39). This disease is also characterized
morphologically by the presence of rimmed vacuoles (40). A
uniform shrinkage of the muscle fibres has not been reported for
any of these disesaes.
In addition to brain and muscle, we have found that the liver,
kidney, heart and testis of the R6/2 mice also decrease in size.
With the exception of testis, these organs contain inclusions in a
subset of cell types, the appearance of which broadly correlates
with the onset of weight loss. Is it possible that the presence of
inclusions can lead to cell shrinkage? Comparison with other
mouse models may argue against this as a general cell autonomous mechanism. No loss in brain weight was reported in mice
in which the hprt gene had been targeted with a polyQ expansion,
despite an extensive distribution of neuronal inclusions (20).
Recent reports of diabetes in R6/2 mice (41) are probably
accounted for by the identification of inclusions in the pancreatic
islets. It is likely that the presence of the polyQ inclusions directly
leads to cellular dysfunction in these cells. In 72% of cases, type II
diabetes is associated with the deposition of amylin between the
endocrine cells and capillaries, usually penetrating into deep
invaginations of the plasma membrane of the B cells. It is thought
that the accumulation of amyloid in the islets is likely to impair
islet function and may be a causal factor in the development of
type II diabetes (42). Consistent with this hypothesis, mutations
in the amylin gene are responsible for an early onset form of
type II diabetes (43). The deposition of polyQ amyloid in
pancreatic islets may similarly lead to an impairment of islet
function. It is interesting that an increase in the frequency of
diabetes in HD patients, as compared with the normal population,
has been reported (44), although it is not known whether this is
related to polyQ amyloid deposition. However, the diabetes may
contribute to the loss of body weight observed in these mice.
Non-CNS pathology and inclusions in HD
There is little documentation of peripheral organ weight and/or
histology in HD. One detailed study of liver morphology was
conducted in response to the observation that HD patients
commonly presented with a minor disturbance of liver function
(45). This paper reported that in HD the liver tends to be smaller
than normal and the histology suggested that hepatocytes in HD
have a shortened survival time. In addition, a recent review (46)
discussed an unpublished study of 122 HD cases in which organ
weight was found to be significantly reduced for cases with
increased severity of brain involvement (E.D. Bird, C. Hall and
R.H. Myers, unpublished data). Weight loss was not uniform
across all tissues and was most evident in the liver and heart and
least evident in the kidney.
A recent study of the peripheral organs in SBMA identified
non-neural inclusions in scrotal skin, dermis, kidney, heart and
testis but not in spleen, liver and muscle (47). In this case, the cells
containing inclusions were all mitotic cells, capable of mitosis in
adulthood. Inclusions were nuclear and contained the N-terminus
of the androgen receptor (AR). The authors suggest that this
pattern could not be explained by expression levels alone and
could be influenced by the presence of a protease that cleaves the
N-terminal fragment from the full-length protein, specific
proteins that interact with the mutant AR or different AR
functional activities among the different tissues. It is not known
whether inclusions form in HD peripheral organs as there are no
ultrastructural reports or examples of immunocytochemistry
using anti-N-terminal huntingtin antibodies to post-mortem HD
tissues outside the CNS. As discussed above, in the R6 lines the
probable initial step in the pathogenic pathway is missing, that of
cleavage/proteolysis to generate the aggregation precursor.
Therefore, any tissue variations in the distribution of factors
involved in the generation of this fragment would be expected to
811
Human
Genetics,
1999,
8, No.
NucleicMolecular
Acids Research,
1994,
Vol. Vol.
22, No.
1 5
impact on the cell specificity of inclusions. In addition, the R6
transgenic lines are clearly an extreme version of the disease, with
polyQ expansions greater than that generally found in juvenile
HD. Therefore, it may be necessary to look at juvenile cases to
determine whether the distribution of non-neural inclusions seen
in the mice might be mirrored in patients, and these studies will
be undertaken when suitable tissues become available.
Conclusion
PolyQ inclusions have been identified in a number of post-mitotic
non-CNS cells in a mouse model of HD. The most immediate
application of these findings will be with respect to testing
therapeutic interventions. A major effort will be invested in the
identification of small molecules that can prevent or slow down
the kinetics of aggregate formation. Nuclear inclusions in
non-CNS tissues will be used to assess whether compounds,
selected for their ability to prevent aggregation in vitro, can
prevent or slow down inclusion formation in vivo without having
to worry in the first instance whether they are capable of crossing
the blood–brain barrier.
MATERIALS AND METHODS
Mice
Transgenic mouse lines R6/1 and R6/2 [Jackson codes B6CBATgN(HDexon1)61 and B6CBA-TgN(HDexon1)62] can be obtained from the Induced Mutant Resource (Jackson Laboratory,
Bar Harbor, ME). Lines R6/1 and R6/2 were maintained by
backcrossing to (C57BL/6×CBA)F1 mice. Genotyping and CAG
repeat sizing was as previously described (4). The timing and
progression of organ atrophy were assessed by weighing organs
dissected from a series of male mice that had been perfusion fixed
with 4% paraformaldehyde.
Antibodies, immunocytochemistry, histology and
ultrastructure
The sources and working dilutions of antibodies were as follows:
anti-ubiquitin (1:2000; Dako, Copenhagen, Denmark); CAG53b
(1:5000; 7); anti-N-cam H28 (1:500; Boehringer Mannheim,
Mannheim, Germany); anti-myosin heavy chain type 1 (A4.84),
type 2 (N3.36) and type 2A (A474) skeletal muscle fibre
antibodies (Simon Hughes, King’s College, London, UK).
Immunohistochemistry was performed as previously described
(48) on 5 µm paraffin embedded peripheral tissue sections (49)
from mice that had been perfusion fixed with paraformaldehyde/
lysine/periodate fixative, with the exception of muscle for which
8 µm frozen sections were prepared (48). Ultrastructural analysis
(48) and histological stains (49) were carried out as published.
Morphometry
The lesser muscle fibre diameter was measured by a semiautomated approach using the colour freelance program (Foster
Findlay Associates, Newcastle upon Tyne, UK).
ABBREVIATIONS
AR, androgen receptor; CNS, central nervous system; DRPLA,
dentatorubral pallidoluysian atrophy; HD, Huntington’s disease;
H&E, haematoxylin and eosin; HVG, haematoxylin–van Gieson;
811
NADH-TR, NADH-tetrazolium reductase; PAS, periodic acid–
Schiff base; polyQ, polyglutamine; s-IBM, sporadic inclusion
body myositis; SBMA, spinal and bulbar muscular atrophy; SCA,
spinocerebellar ataxia.
ACKNOWLEDGEMENTS
We wish to thank Anne Young, Stefan Buk, Rick H. Myers, Jean
Bolt, Jean Louis Mandel, Adriano Aguzzi and Hans Lehrach for
enlightening discussions, and Robin Poston for help with the
muscle fibre morphometry. This work was supported by grants
from the Wellcome Trust, European Union (BMH4 CT96 0244),
Hereditary Disease Foundation (in the form of an award to G.P.B.
from Harry Lieberman), the Huntington’s Disease Society of
America and the Special Trustees of Guy’s Hospital.
REFERENCES
1. Bates, G.P., Mangiarini, L. and Davies, S.W. (1998) Transgenic mice in the
study of polyglutamine repeat expansion diseases. Brain Pathol., 8, 699–714.
2. Wells, R. and Warren, S. (1998) Genetic Instabilities and Hereditary
Neurological Diseases. Academic Press, San Diego, CA.
3. Mangiarini, L., Sathasivam, K., Seller, M., Cozens, B., Harper, A.,
Hetherington, C., Lawton, M., Trottier, Y., Lehrach, H., Davies, S.W. and
Bates, G.P. (1996) Exon 1 of the Huntington’s disease gene containing a
highly expanded CAG repeat is sufficient to cause a progressive neurological
phenotype in transgenic mice. Cell, 87, 493–506.
4. Mangiarini, L., Sathasivam, K., Mahal, A., Mott, R., Seller, M. and Bates,
G.P. (1997) Instability of highly expanded CAG repeats in transgenic mice is
related to expression of the transgene. Nature Genet., 15, 197–200.
5. Carter, R.J., Lione, L.A., Humby, T., Mangiarini, L., Mahal, A., Bates, G.P.,
Morten, A.J. and Dunnett, S.B. (1999) Characterisation of progressive motor
deficits in mice transgenic for the human Huntington’s disease mutation. J.
Neurosci., in press.
6. File, S., Mahal, A., Mangiarini, L. and Bates, G.P. (1998) Striking changes in
anxiety in Huntingon’s disease transgenic mice. Brain Res., 805, 234–240.
7. Davies, S.W., Turmaine, M., Cozens, B.A., DiFiglia, M., Sharp, A.H., Ross,
C.A., Scherzinger, E., Wanker, E.E., Mangiarini, L. and Bates, G.P. (1997)
Formation of neuronal intranuclear inclusions (NII) underlies the neurological dysfunction in mice transgenic for the HD mutation. Cell, 90, 537–548.
8. Scherzinger, E., Lurz, R., Turmaine, M., Mangiarini, L., Hollenbach, B.,
Hasenbank, R., Bates, G.P., Davies, S.W., Lehrach, H. and Wanker, E.E.
(1997) Huntingtin encoded polyglutamine expansions form amyloid-like
protein aggregates in vitro and in vivo. Cell, 90, 549–558.
9. Perutz, M.F., Johnson, T., Suzuki, M. and Finch, J.T. (1994) Glutamine
repeats as polar zippers: their possible role in inherited neurodegenerative
diseases. Proc. Natl Acad. Sci. USA, 91, 5355–5358.
10. Scherzinger, E., Sittler, A., Heiser, V., Schweiger, K., Hasenbank, R., Bates,
G.P., Lehrach, H. and Wanker, E.E. (1999) Self-assembly of polyglutaminecontaining huntingtin fragments into amyloid-like fibrils: implications for
Huntington’s disease pathology. Proc. Natl Acad. Sci. USA, in press.
11. DiFiglia, M., Sapp, E., Chase, K.O., Davies, S.W., Bates, G.P., Vonsattel, J.-P.
and Aronin, N. (1997) Aggregation of huntingtin in neuronal intranuclear
inclusions and dystrophic neurites in brain. Science, 277, 1990–1993.
12. Becher, M.W., Kotzuk, J.A., Sharp, A.H., Davies, S.W., Bates, G.P., Price,
D.L. and Ross, C.A. (1998) Intranuclear neuronal inclusions in Huntington’s
disease and dentatorubral pallidoluysian atrophy: correlation between the
density of inclusions and IT15 CAG repeat length. Neurobiol. Dis., 4,
387–395.
13. Gourfinkel-An, I., Cancel, G., Duyckaerts, C., Faucheux, B., Hauw, J.-J.,
Trottier, Y., Brice, A., Agid, Y. and Hirsch, E.C. (1998) Neuronal distribution
of intranuclear inclusions in Huntington’s disease with adult onset. NeuroReport, 9, 1823–1826.
14. Skinner, P.J., Koshy, B.T., Cummings, C.J., Klement, I.A., Helin, K.,
Servadio, A., Zoghbi, H.Y. and Orr, H.T. (1997) Ataxin-1 with an expanded
glutamine tract alters nuclear matrix-associated structures. Nature, 389,
971–974.
15. Paulson, H.L., Perez, M.K., Trottier, Y., Trojanowski, J.Q., Subramony, S.H.,
Das, S.S., Vig, P., Mandel, J.-L., Fischbeck, K.H. and Pittman, R.N. (1997)
Intranuclear inclusions of expanded polyglutamine protein in spinocerebellar
ataxia type 3. Neuron, 19, 1–20.
812
Human Molecular Genetics, 1999, Vol. 8, No. 5
16. Holmberg, M., Duyckaerts, C., Durr, A., Cancel, G., Gourfinkel-An, I.,
Damier, P., Faucheux, B., Trottier, Y., Hirsch, E.C., Agid, Y. and Brice, A.
(1998) Spinocerebellar ataxia type 7 (SCA7): a neurodegenerative disorder
with neuronal intranuclear inclusions. Hum. Mol. Genet., 7, 913–918.
17. Igarashi, S., Koide, R., Shimohata, T., Yamada, M., Hayashi, Y., Takano, H.,
Date, H., Oyake, M., Sato, T., Sato, A., Egawa, S., Ikeuchi, T., Tanaka, H.,
Nakano, R., Tanaka, K., Hozumi, I., Inuzuka, T., Takahashi, H. and Tsuji, S.
(1998) Suppression of aggregate formation and apoptosis by transglutaminase inhibitors in cells expressing truncated DRPLA protein with an
expanded polyglutamine stretch. Nature Genet., 18, 111–117.
18. Li, M., Miwa, S., Kobayashi, Y., Merry, D.E., Yamamoto, M., Tanaka, F.,
Doyu, M., Hashizume, Y., Fischbeck, K.H. and Sobue, G. (1998) Nuclear
inclusions of the androgen receptor protein in spinal and bulbar muscular
atrophy. Ann. Neurol., 44, 249–254.
19. Cummings, C.J., Mancini, M.A., Antalffy, B., DeFranco, D.B., Orr, H.T. and
Zoghbi, H.Y. (1998) Chaperone suppression of aggregation and altered
subcellular proteasome localisation imply protein misfolding in SCA1.
Nature Genet., 19, 148–154.
20. Ordway, J.M., Tallaksen-Greene, S., Gutekunst, C.-A., Bernstein, E.M.,
Cearley, J.A., Wiener, H.W., Dure IV, L.S., Lindsey, R., Hersch, S.M., Jope,
R.S., Albin, R.L. and Detloff, P.J. (1997) Ectopically expressed CAG repeats
cause intranuclear inclusions and a progressive late onset neurological
phenotype in the mouse. Cell, 91, 753–763.
21. Quarrell, O.W.J. and Harper, P.S. (1996) The clinical neurology of
Huntington’s disease. In Harper, P.S. (ed.), Huntington’s Disease, 2nd Edn.
W.B. Saunders, London, UK, Vol. 31, pp. 31–72.
22. Sanberg, P.R., Fibiger, H.C. and Mark, R.F. (1981) Body weight and dietary
factors on Huntington’s disease patients compared with matched controls.
Med. J. Aust., 1, 407–409.
23. Lansbury, P.T. (1997) Structural neurology: are seeds at the root of neuronal
degeneration. Neuron, 19, 1151–1154.
24. Martindale, D., Hackman, A., Wieczorek, A., Ellerby, L., Wellington, C.,
McCutcheon, K., Singaraja, R., Kazemi-Esfarjani, P., Devon, R., Kim, S.U.,
Bredesen, D.E., Tufaro, F. and Hayden, M.R. (1998) Length of huntingtin and
its polyglutamine tract influences localisation and frequency of intracellular
aggregates. Nature Genet., 18, 150–154.
25. Li, S.-H. and Li, X.-J. (1998) Aggregation of N-terminal huntingtin is
dependent on the length of its glutamine repeats. Hum. Mol. Genet., 7,
777–782.
26. Cooper, J.K., Schilling, G., Peters, M.F., Herring, W.J., Sharp, A.H.,
Kaminsky, Z., Masone, J., Khan, F.A., Delanoy, M., Borchelt, D.R., Dawson,
V.L., Dawson, T.M. and Ross, C.A. (1998) Truncated N-terminal fragments
of huntingtin with expanded glutamine repeats form nuclear and cytoplasmic
aggregates in cell culture. Hum. Mol. Genet., 7, 783–790.
27. Hackam, A.S., Singaraja, R., Wellington, C.L., Metzler, M., McCutcheon, K.,
Zhang, T., Kalchman, M. and Hayden, M.R. (1998) The influence of
huntingtin protein size on nuclear localisation and cellular toxicity. J. Cell
Biol., 141, 1097–1105.
28. Lunkes, A. and Mandel, J.-L. (1998) A cellular model that recapitulates major
pathogenic steps of Huntington’s disease. Hum. Mol. Genet., 7, 1355–1361.
29. Strong, T.V., Tagle, D.A., Valdes, J.M., Elmer, L.W., Boehm, K., Swaroop,
M., Kaatz, K.W., Collins, F.S. and Albin, R.L. (1993) Widespread expression
of the human and rat Huntington’s disease gene in brain and nonneuronal
tissues. Nature Genet., 5, 259–263.
30. DiFiglia, M., Sapp, E., Chase, K., Schwarz, C., Meloni, A., Young, C., Martin,
E., Vonstattel, J.-P., Carraway, R., Reeves, S.A., Boyce, F.M. and Aronin, N.
(1995) Huntingtin is a cytoplasmic protein associated with vesicles in human
and rat brain neurons. Neuron, 14, 1075–1081.
31. Sharp, A.H., Loev, S.J., Schilling, G., Li, S.-H., Li, X.-J., Bao, J., Wagster,
M.V., Kotzuk, J.A., Steiner, J.P., Lo, A., Hedreen, J., Sisodia, S., Snyder, S.H.,
Dawson, T.M., Ryugo, D.K. and Ross, C.A. (1995) Widespread expression of
Huntington’s disease gene (IT15) protein product. Neuron, 14, 1065–1074.
32. Trottier, Y., Devys, D., Imbert, G., Sandou, F., An, I., Lutz, Y., Weber, C.,
Agid, Y., Hirsch, E.C. and Mandel, J.-L. (1995) Cellular localisation of the
Huntington’s disease protein and discrimination of the normal and mutated
forms. Nature Genet., 10, 104–110.
33. de Rooij, K.E., Dorsman, J.C., Smoor, M.A., den Dunnen, J.T. and van
Ommen, G.-J. (1996) Subcellular localisation of the Huntington’s disease
gene product in cell lines by immunofluorescence and biochemical
subcellular fractionation. Hum. Mol. Genet., 5, 1093–1099.
34. Cha, J.-H.J., Kosinski, C.M., Kerner, J.A., Alsdorf, S.A., Mangiarini, L.,
Davies, S.W., Penney, J.B., Bates, G.P. and Young, A.B. (1998) Altered brain
neurotransmitter receptors in transgenic mice expressing a portion of an
abnormal human Huntington’s disease gene. Proc. Natl Acad. Sci. USA, 95,
6480–6485.
35. Askanas, V. and Engel, W.K. (1998) Sporadic inclusion-body myositis and its
similarities to Alzheimer disease brain. Scand. J. Rheumatol., 27, 389–405.
36. Askanas, V. and Engel, W.K. (1997) Recent progress and classification of the
hereditary inclusion-body myopathies. Acta Myol., 1, 21–25.
37. Jin, L.-W., Hearn, M.G., Ogburn, C.E., Dang, N., Nochlin, D., Ladiges, W.C.
and Martin, G.M. (1998) Transgenic mice over-expressing the C-99 fragment
of βPP with an α-secretase site mutation develop a myopathy similar to
human inclusion body myositis. Am. J. Pathol., 153, 1679–1686.
38. Fukuchi, K.-I., Pham, D., Hart, M., Li, L. and Lindsey, J.R. (1998) Amyloid-β
deposition in skeletal muscle of transgenic mice. Am. J. Pathol., 153,
1687–1693.
39. Brais, B., Bouchard, J.-P., Xie, Y.-G., Rochefort, D.L., Chretien, N., Tome,
F.M.S., Lafreniere, R.G., Rommens, J.M., Uyama, E., Nohira, O., Blumen, S.,
Korcyn, A.D., Heutink, P., Mathieu, J., Duranceau, A., Codere, F., Fardeau,
M. and Rouleau, G.A. (1998) Short GCG expansions in the PABP2 gene
cause oculopharyngeal muscular dystrophy. Nature Genet., 18, 164–167.
40. Tome, F.M.S., Chateau, D., Helbling-Leclerc, A. and Fardeau, M. (1997)
Morphological changes in muscle fibres in oculopharyngeal muscular
dystrophy. Neuromusc. Disord., 7, S63–S69.
41. Hulbert, M.S., Zhou, W., Wasmeier, C., Kaddis, F.G., Hutton, J.C. and Freed,
C.R. (1999) Mice transgenic for an expanded CAG repeat in the Huntington’s
disease gene develop diabetes. Diabetes, 48, 649–651.
42. Clark, A., Lewis, C.E., Willis, A.C., Cooper, G.J.S., Morris, J.F. and Reid,
K.B.M. (1987) Islet amyloid formed from diabetes-associated peptide may
be pathogenic in type-2 diabetes. Lancet, 2, 231–234.
43. Sakagashira, S., Sanke, T., Hanabusa, T., Shimomura, H., Ohagi, S.,
Kumagaye, K.Y., Nakajima, K. and Nanjo, K. (1996) Missense mutation of
amylin gene (S20G) in Japanese NIDDM patients. Diabetes, 45, 1279–1281.
44. Farrer, L.A. (1985) Diabetes mellitus in Huntington’s disease. Clin. Genet.,
27, 62–67.
45. Bolt, J.M.W. and Lewis, G.P. (1973) Huntington’s chorea, a study of liver
function and histology. Q. J. Med., 165, 151–174.
46. Myers, R.H., Marans, K.S. and MacDonald, M.E. (1998) Huntington’s
disease. In Wells, R.D. and Warren, S.T. (eds), Genetic Instabilities and
Hereditary Neurological Diseases. Academic Press, San Diego, CA, pp.
301–323.
47. Li, M., Nakagomi, Y., Kobayashi, Y., Merry, D.E., Tanaka, F., Doyu, M.,
Mitsuma, T., Hashizume, Y., Fischbeck, K.H. and Sobue, G. (1998)
Nonneural nuclear inclusions of androgen receptor protein in spinal and
bulbar muscular atrophy. Am. J. Pathol., 153, 695–701.
48. Davies, S.W., Sathasivam, K., Hobbs, C., Doherty, P., Mangiarini, L.,
Scherzinger, E., Wanker, E.E. and Bates, G.P. (1999) Methods to detect
polyglutamine aggregation in mouse models. Methods Enzymol., in press.
49. Bancroft, J.D. and Stevens, A. (1990) Theory and Practice of Histological
Techniques, 3rd Edn. Churchill Livingstone, London, UK.
Download