Download Studying Human Neurodegenerative Diseases in Flies and Worms

Journal of Neuropathology and Experimental Neurology
Copyright q 2000 by the American Association of Neuropathologists
Vol. 59, No. 10
October, 2000
pp. 847 856
Studying Human Neurodegenerative Diseases in Flies and Worms
MEL B. FEANY, MD, PHD
Abstract. Invertebrate models of several human neurodegenerative diseases have recently been described. These models
faithfully replicate key neuropathological features of the human disorders. Because the basic cell biology of the nervous
system is very similar in vertebrates and invertebrates, the sophisticated and rapid genetic analysis feasible in Drosophila and
C. elegans promises significant insight into human neurodegenerative syndromes. In addition, the short lifespan, small size,
and ease of culturing make worms and flies ideal for drug testing.
Key Words:
Alzheimer disease; C. elegans; Drosophila; Neurodegeneration; Parkinson disease; Polyglutamine.
INTRODUCTION
Animal models provide critical testing grounds for
both models of disease pathogenesis and potential therapeutic agents. Vertebrates, particularly mice, have provided many useful models of nervous system diseases.
However, genetic analysis in vertebrate systems is still
slow in comparison with simpler organisms like the nematode Caenorhabditis elegans, or the fruit fly Drosophila
melanogaster. Recent insight into the molecular basis of
human diseases, and in particular neurodegenerative diseases, has motivated the production of faithful invertebrate models of neurologic diseases.
Modeling diseases in simple invertebrate systems is attractive because genetic interactions can be used to define
cellular cascades mediating, for example, death of dopaminergic neurons in Parkinson disease. The obvious
differences between mammals and invertebrates represent
potential pitfalls to modeling complex human disorders
like neurodegenerative diseases in flies or worms. Fortunately, despite marked anatomic divergence, basic cell
biological processes are highly conserved between the 2
classes of organisms. Processes as diverse as development and learning and memory not only share mechanistic similarities between vertebrates and invertebrates,
but also often show conservation at the level of specific
second messenger systems and transcription factors as
well. With the sequencing of the entire genomes of C.
elegans and Drosophila, similarities can now be evaluated on the whole genome level. Orthologs exist for many
proteins involved in human neurologic diseases (Table 1)
(1).
Worms and flies have dominated invertebrate models
of human disease. Both C. elegans and Drosophila have
the rapid generation time, well-developed genetics, and
transgenic technology required to select mutations based
From the Department of Pathology, Division of Neuropathology,
Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts.
Correspondence to: Mel B. Feany, MD, PhD, Department of Pathology, Division of Neuropathology, Brigham and Women’s Hospital, Harvard Medical School, 221 Longwood Ave., Boston, MA 02115.
on phenotype, define interacting loci, and express human
genes of interest. The 2 organisms are complementary in
many ways. The C. elegans nervous system is simpler
than the Drosophila nervous system. In C. elegans, each
neuron can be identified individually. More sophisticated
behaviors may be monitored in Drosophila, correlating
with the additional complexity of the nervous system.
Worms live 3 wk, while flies have a life span of 2
months. Transgenic animals are relatively easy to produce
in both system, but ‘‘knock-in’’ technology is primitive
in both systems. A variety of techniques exist to modulate
transgene expression. Tissue specific and inducible promotors are widely available, allowing tissue and developmentally specific effects to be defined.
Several approaches have been taken to modeling human diseases in invertebrates. If an ortholog is known
from sequence information (Table 1), mutations can be
identified in the invertebrate gene, phenotypes sought,
and biochemical cascades defined. Human counterparts
of novel components of the cellular pathways delineated
in the invertebrate model organism will be attractive candidates for roles in the human disorder. In practice, substantial insights often obtain when a gene mutated in a
human disease is cloned and immediately identified as a
component of a well-characterized invertebrate genetic
pathway.
Traditional genetics relies on phenotypes rather than
sequence information. Novel mutations are produced that
show features of a specific disease process. The causative
genetic defects may then be relevant to neurologic disorders with the same characteristics. For instance, if mutant flies are selected for adult onset degeneration of the
nervous system, human homologs of the Drosophila gene
products are plausible candidates for involvement in neurodegenerative diseases.
Alternatively, more directed strategies may be employed. If human disease gene products are toxic, relevant models may be produced by simply expressing the
human protein in the model organism. Given the advanced transgenic technology available in flies and
worms, such expression experiments are technically
straightforward. To date, expressing human proteins
847
848
FEANY
TABLE 1
Neurodegenerative Disease Gene Orthologs
Ortholog present
Disease
Gene
Alzheimer disease
Amyloid precursor protein
Presenilin
Parkin
UCHL1
Huntingtin
Ataxin-2
Ataxin-6
Tau
Superoxide dismutase
Parkinson disease
Polyglutamine diseases
‘‘Tau-opathies’’
Amyotrophic lateral sclerosis
TABLE 2
Invertebrate Models of Human Neurodegenerative
Diseases
Disease modeled
Huntingtin disease
Parkinson disease
Spinocerebellar ataxia type 1
Spinocerebellar ataxia type 3
Pathology in invertebrates
Flies:
Retinal degeneration
Nuclear inclusions
Worms:
Neuronal degeneration
Non-nuclear aggregates
Flies:
Dopaminergic neurodegeneration
Filamentous cytoplasmic
inclusions
Flies:
Retinal abnormalities
Nuclear inclusions
Non-neuronal cell death
Flies:
Retinal degeneration
Nuclear inclusions
Non-neuronal cell death
linked to neurodegenerative diseases has provided the
best invertebrate models in terms of both fidelity to the
human disorder, and subsequent genetic insights (Table
2).
Parkinson Disease
Mutations in the a-synuclein gene have been linked to
familial Parkinson disease (2, 3). a-synuclein is an abundant neuronal protein of unknown function that does not
have a clear ortholog in Drosophila or C. elegans (1).
Modeling Parkinson disease by manipulating the endogenous a-synuclein gene in invertebrates is therefore not
an attractive option. However, if disease is produced in
humans by the action of a toxic protein, it may not be
necessary or even desirable, to manipulate the invertebrate homolog of the human disease-related gene. Instead, expression of the toxic human protein model organism may model the disorder. It is not clear that
J Neuropathol Exp Neurol, Vol 59, October, 2000
Drosophila
C. elegans
1
1
1
1
1
1
1
1
1
1
1
2
1
2
2
1
1
1
mutations in a-synuclein cause Parkinson disease through
a toxic mechanism; however, the dominant mode of inheritance is consistent with a gain of function property.
In addition, a-synuclein is a prominent component of
Lewy bodies, the abnormal neuronal cytoplasmic inclusion characteristic of Parkinson disease. Thus, abnormal
aggregation of the protein may be related to disease pathogenesis. a-synuclein is present in Lewy bodies not only
in the rare familial cases of Parkinson disease linked to
mutation in the a-synuclein gene, but also in inclusions
in the much more common sporadic form of the disease
as well. A common mechanism may therefore operate in
the familial and sporadic forms of the disease. The effects
of expressing human a-synuclein in Drosophila (4) and
mice (5) support these speculations.
When normal or mutant human a-synuclein is expressed in the nervous system of flies, 3 features of Parkinson disease are reproduced: adult-onset loss of dopaminergic neurons, filamentous intraneuronal inclusions
containing a-synuclein, and progressive locomotor dysfunction. Dopaminergic neurons degenerate at about 30
days of age in the transgenic flies. Since flies live only
60 days, the effects of genetic manipulation, like expression of human a-synuclein, can be observed over the entire lifespan of the organism in a relatively short period
of time. Nonneuronal cells are not affected by expression
of normal or mutant human a-synuclein.
The effect on dopaminergic neurons is at least relatively specific. The majority of non-dopaminergic neurons are spared in aged transgenic flies. There is also
specificity for particular subsets of dopaminergic neurons. Dopaminergic neurons in the dorsomedial region of
the fly’s brain appear particularly vulnerable to the effect
of human a-synuclein. In Parkinson disease, subsets of
dopaminergic neurons within the substantia nigra also
show selective vulnerability. Cells in the ventral lateral
portion of the nucleus are more liable to degeneration
than neurons in other areas. The molecular basis of enhanced susceptibility of certain dopaminergic neurons in
Parkinson disease is unknown. The presence of a similar
NEURODEGENERATIVE DISEASE IN INVERTEBRATES
Fig. 1. a-synuclein immunoreactivity in transgenic Drosophila is initially diffuse (A), but becomes punctate as the
animals age (B, arrow).
phenomena in the Drosophila model suggests that the
molecular basis of increased susceptibility within subsets
of dopaminergic neurons may be amenable to genetic
analysis.
a-synuclein containing cytoplasmic inclusion bodies
are formed in transgenic Drosophila (Fig. 1B). The staining pattern for a-synuclein is initially diffuse and cytoplasmic in transgenic flies (Fig. 1A); however, over time
a-synuclein immunoreactive puncta appear (Fig. 1B).
The time course of inclusion formation roughly parallels
dopaminergic neuronal loss, suggesting that the 2 phenomena may be causally related. On electron microscopic
analysis, the a-synuclein inclusions are found in neuronal
cell bodies and are composed of filaments with diameters
from 7 to 10 nm, admixed with granular material similar
to human Lewy bodies.
Loss of dopaminergic neurons and inclusion formation
is accompanied by locomotor dysfunction. Flies normally
exhibit vigorous climbing activity when subjected to a
gentle tap. Over time, a-synuclein transgenic flies lose
their climbing ability more rapidly than normal flies.
These findings demonstrate a behavioral consequence of
expressing human a-synuclein in Drosophila, plausibly
mediated by loss of dopaminergic neurons and/or inclusion formation. All 3 phenomena, dopaminergic neurodegeneration, inclusion formation, and locomotor dysfunction occur over the same time course. In addition, all
3 abnormalities manifest with expression of normal as
849
well as mutant human a-synuclein, supporting a common
pathogenetic mechanism in idiopathic Parkinson disease
and the rare familial variants linked to mutations in the
a-synuclein gene. Expression of human a-synuclein in
transgenic mice also produces dopaminergic degeneration, a-synuclein inclusion formation, and locomotor abnormalities (5).
Although dopaminergic neurons appear particularly
sensitive to a-synuclein toxicity in flies, other cells in the
nervous system of transgenic Drosophila can also be affected by expression of human a-synuclein. Expression
of normal or mutant a-synuclein specifically in the retina
produces age dependent retinal degeneration. Thus, in
flies, as in people, a-synuclein-related degenerative
changes show relative rather than absolute specificity for
dopaminergic neurons. The Drosophila retina is probably
somewhat resistant to a-synuclein toxicity because about
5-fold more transgenic a-synuclein is produced in the eye
than the brain. The presence of retinal degeneration in asynuclein transgenic flies is an important technical advantage. Harnessing the powerful genetics of invertebrate
model organisms is the ultimate goal of modeling human
neurodegenerative diseases in these animals. In practice,
this means creating additional mutations that modify relevant phenotypes. Large numbers of randomly generated
mutations must be examined to identify the infrequent
mutations that modify a-synuclein toxicity. An easily detected phenotype, ideally one that can be tested quickly
in live flies, is required for successful generation of sufficient modifiers to delineate biochemical cascades. Molecular cloning of such modifiers should identify novel
proteins mediating a-synuclein toxicity in Drosophila.
The strong conservation between vertebrate and invertebrate systems suggests that mammalian homologs of
Drosophila modifiers will encode proteins important in
the pathogenesis of the human disorder. Invertebrate
models will thus act as discovery tools to identify molecules underlying pathology in human disease. A critical
test of this scheme is identification of genetic mutations
in mammalian homologs of genes identified in invertebrates in familial forms of the human disease. Thus,
Fig. 2. Comparison of normal fly eye (A) with eyes from flies expressing unexpanded ataxin-1 (B), and expanded ataxin-1
(C). Expression of both normal and mutant forms of ataxin-1 produce rough eyes with small necrotic areas (black spots in B and
C). D: Nuclear inclusion in transgenic Drosophila expressing expanded ataxin-1 in the retina (arrow).
J Neuropathol Exp Neurol, Vol 59, October, 2000
850
FEANY
mammalian homologs of modifiers of a-synuclein toxicity defined in Drosophila will be candidates for Parkinson disease mutations.
Polyglutamine Expansion Disorders: Huntington
Disease and the Spinocerebellar Ataxias
Direct expression of a toxic human protein in invertebrates has provided a number of useful models of disorders caused by the expansion of polyglutamine-encoding CAG repeat sequences (6–8). Diseases such as
Huntington disease and the autosomal dominant cerebellar ataxias are caused by expansion of a trinucleotide sequence within the coding region of unrelated proteins.
The expanded sequence encodes polyglutamine stretches.
The toxic dominant nature of CAG expansion mutations
has been established in mouse models. In transgenic
mice, expression of the expanded, but not unexpanded,
protein produces toxicity. Inactivation of the mouse homologs does not mimic the human disease. The strong
genetic evidence for a toxic dominant gain of function
mechanism encouraged several groups to model polyglutamine disorders in invertebrates. Warrick et al (6) demonstrated that expanded, but not unexpanded, ataxin-3
produces age-dependent degeneration of the Drosophila
retina. Although both neuronal and nonneuronal tissue is
affected by expression of expanded ataxin, neurons appear particularly vulnerable to the toxic action of the mutant protein. The retinal abnormality can be at least partially rescued by expression of the baculovirus
antiapoptotic protein p35, suggesting that cell death occurs via an apoptotic mechanism. In contrast, expression
of expanded huntingtin in the Drosophila eye produces
retinal degeneration that is not rescued by p35 (7). Differences in the expression systems used by the 2 groups
could account for the inconsistency of p35 rescue. Alternatively, the differential effect of p35 may indicate that
polyglutamine expansions act via different mechanisms
depending on their protein contexts. Protein context must
play an important role in the polyglutamine disorders,
because distinct clinical and pathologic syndromes are
produced by polyglutamine expansions within different
proteins showing widespread, overlapping expression
patterns (reviewed in 9).
Similar experiments have been carried out in C. elegans. Faber et al. (8) expressed expanded and unexpanded forms of huntingtin in nematode sensory neurons.
They observed huntingtin-induced neuronal degeneration, but only after coexpression of another toxic protein.
Induction of degeneration also required very long trinucleotide expansion. In C. elegans, 150 repeat, but not 95
repeat mutations caused neurodegeneration. In humans,
more than about 38 repeats usually produces disease, and
a repeat length of 95 would be expected to produce severe, juvenile onset Huntington disease. ced-3 caspase
J Neuropathol Exp Neurol, Vol 59, October, 2000
function was required for neurodegeneration, implicating
apoptosis as the mechanism of cell death.
Protein aggregation and formation of nuclear inclusions are prominent features of human and transgenic
mouse pathology in polyglutamine expansion diseases.
Both Drosophila and C. elegans develop abnormal protein aggregates, but only the Drosophila models show
nuclear localization of the aggregated protein (Fig. 2D)
(6–8). In both mice and Drosophila, inclusion formation
precedes detectable pathology (6, 7, 10). In contrast, a
functional assay for sensory neuron function in C. elegans reveals neuronal dysfunction before cytoplasmic
protein aggregates are seen. Although nuclear inclusions
are a prominent feature of polyglutamine diseases in humans, mice, and flies, the relationship of inclusion formation to disease pathogenesis remains a subject of debate. Recent work in transgenic mice (11) and cell culture
(12) has shown that polyglutamine containing proteins
can induce neuronal dysfunction in the absence of large
nuclear aggregates. It remains to be determined if toxicity
requires formation of smaller aggregates not readily detectable at the light microscopic level.
Invertebrate models may not only reproduce key aspects of human disorders, but may also suggest new hypotheses regarding the pathogenesis of the human disease. Expression of expanded ataxin-1 in Drosophila
produces abnormalities in retinal (Fig. 2C) and other tissues (4), consistent with results with other mutant polyglutamine-containing proteins. Unexpectedly, however,
unexpanded ataxin-1 is also quite toxic (Fig. 2B). Normal
sequence ataxin-1 may have neurotoxic properties unrelated to the polyglutamine expansion. Expression of
unexpanded ataxin-1 does not have a toxic effect on
mouse Purkinje cells. Unexpanded ataxin-1 may be toxic
to Drosophila, but not mice. Alternatively, ataxin-1 toxicity might manifest only with higher concentrations of
the protein, or with aggregation. Ataxin-1 toxicity might
then play a role in human disease following aggregation
of the protein, alteration of the normal subcellular localization, or increase in concentration. The Drosophila results suggest that expanded ataxin-1 might exert a dominant negative effect. It would be of interest to determine
if toxicity of expanded ataxin-1 requires, or is influenced
by, the presence of unexpanded protein. A possible dominant negative effect could be tested using mice lacking
endogenous ataxin-1.
Thus far, invertebrate models of polyglutamine diseases appear promising: flies and worms expressing mutant
proteins show both degeneration of the nervous system
and abnormal protein aggregates. However, the major
utility of the invertebrate systems is in defining proteins
important in the pathogenesis of the human disorders.
Two groups have approached the problem with different
methods. First, Warrick et al (13) used the candidate approach to test proteins they suspected might be involved
NEURODEGENERATIVE DISEASE IN INVERTEBRATES
in the disease based on current models of disease pathogenesis. These authors showed that expression of the
molecular chaperone HSP70 suppresses mutant ataxin-3
induced neurodegeneration in the Drosophila retina.
Strikingly, chaperone proteins were also implicated in
polyglutamine toxicity using an unbiased genetic approach. Kazemi-Esfarjani and Benzer (14) screened for
novel mutations that suppress polyglutamine toxicity in
Drosophila. They identified a gene homologous to human
heat shock protein 40/HDJ1 and a separate locus that also
encodes a predicted chaperone-related J domain. The
Drosophila modifier results correlate well with the presence of protein aggregates in polyglutamine diseases and
studies in mammalian systems implicating heat shock
chaperones polyglutamine diseases (15).
Alzheimer Disease
The impact of invertebrate work on the polyglutamine
expansion diseases has come primarily from animals
overexpressing mutant human proteins. In contrast, contributions of invertebrates to the study of Alzheimer disease have obtained almost exclusively from a more traditional approach. Genetic experiments have delineated
basic biologic pathways involved in development and
normal physiological processes, and these results have
been related to Alzheimer disease pathogenesis. Genetic
evidence that the presenilins influence Notch signaling
provided the first intersection between signaling cascades
defined in invertebrates and human neurodegenerative
disease. The Notch family of transmembrane receptors
mediate a variety of cell interactions during development
in Drosophila and C. elegans (reviewed in 16). More
than 80% of familial Alzheimer disease cases are linked
to mutations in the presenilin genes, which encode multiple transmembrane domain proteins (17–19). Presenilins
play a key role in Notch signaling. Work in C. elegans
first implicated presenilins in Notch functioning. One of
the two C. elegans presenilin genes, sel-12, was first defined as a suppressor of a hypermorphic allele of lin-12,
a C. elegans Notch protein (20). Mice with inactivated
presenilin or Notch genes show similar malformations
(21, 22). Recently, the role of presenilin in Notch processing has been further defined. Signaling through the
Notch receptor requires ligand-dependent cleavage of the
Notch protein. Cleavage of the Notch protein occurs
within the transmembrane domain, and is presenilin dependent (23–25).
Like Notch, processing of the amyloid precursor protein involves cleavage within the transmembrane domain.
Many lines of evidence implicate altered regulation of
the intratransmembrane cleavage in the pathogenesis of
Alzheimer disease. Cleavage within the transmembrane
domain of the amyloid precursor protein by gamma-secretase is required to produce a set of peptides 39 to 43
851
amino acids in length (Ab peptides) that form the extracellular congophilic amyloid plaques characteristic of Alzheimer disease. Approximately 2% to 3% of familial
Alzheimer disease is caused by mutations in the amyloid
precursor protein that act to increase Ab. In many systems, presenilin mutations increase Ab accumulation, and
presenilins likely represent gamma secretase (26).
Although intramembranous proteolysis provides the
most obvious analogy between the Notch signaling pathway and neurodegenerative disease, other elements of the
same cascade may prove equally relevant. Presenilin associates with beta-catenin and glycogen synthase kinase3beta (GSK-3beta). Both beta-catenin and GSK-3beta are
modulated by from the Notch and the related Wingless/
Wnt pathways (reviewed in 27). Presenilin stabilizes
beta-catenin. Alzheimer disease associated mutations reduce the stabilization provided by presenilin, and may
predispose to beta-amyloid induced apoptosis (28). In
contrast, mutant presenilins bind avidly to GSK-3beta
and enhance the ability of the kinase to phosphorylate
tau (29). Phosphorylation of tau by GSK-3beta has been
implicated in neurofibrillary tangle formation. Thus, presenilin mutations may have multiple effects in the Notch/
Wingless signaling pathways that predispose to the diverse cellular pathologies associated with Alzheimer
disease: neuronal loss, amyloid plaque formation, and development of neurofibrillary tangles.
Mutations in Notch, and the Notch ligand Jagged, have
been documented in human disease. CADASIL, or cerebral autosomal dominant arteriopathy with subcortical
infarcts and leukoencephalopathy, is a small vessel vasculopathy characterized by multiple subcortical infarcts
and a clinical syndrome that often includes dementia.
Missense mutations in Notch3 are present in CADASIL
patients (30). T-cell lymphoblastic leukemia/lymphoma
can be associated with translocations at the Notch locus
(31). Jagged mutations cause Alagille syndrome, an autosomal dominant disorder including neonatal jaundice,
intrahepatic cholestasis, and developmental anomalies of
the liver, heart, vertebrae, eyes, and face (32).
The Drosophila and C. elegans amyloid precursor protein homologs do not contain sequences homologous to
human Ab (33, 34). Since production and deposition of
Ab appears central in the development of Alzheimer disease, manipulating the endogenous amyloid precursor
protein ortholog in these organisms holds little promise
for modeling the human disorder. However, an expression
strategy relying on direct toxicity of a human protein
does appear useful. Expression of a signal peptide-Ab
fusion construct in C. elegans produces large Congophilic, Ab-rich deposits in muscle (35). The ability of
forced expression of the Ab peptide to produce an amyloid-like substance in worms has been used to assay the
effects of various amino acid substitutions on in vivo
amyloid production (36). If the system can be expanded
J Neuropathol Exp Neurol, Vol 59, October, 2000
852
FEANY
to monitor neurotoxicity of Ab as well, the inherent potential of the system to define elements required for Ab
toxicity genetically could be exploited. In addition, the
relationship between neurotoxicity and aggregation could
be addressed.
Retinitis Pigmentosa
The parallels between retinal degeneration in humans
and Drosophila provides one of the best arguments that,
despite differences in the organization and function of the
nervous system, insights derived from the invertebrate
system will be directly applicable to human disease. Mutations in human rhodopsin cause retinal degeneration, or
retinitis pigmentosa (37). Despite clear anatomic and biochemical differences between Drosophila and mammalian eyes, mutations in Drosophila opsin also cause retinal degeneration (38). Retinal degeneration in fly
rhodopsin mutants, and in other Drosophila retinal degeneration mutants as well, can be blocked by expression
of the baculovirus anti-apoptotic protein p35 (39). These
findings suggest that modulation of apoptosis may be
therapeutic in retinitis pigmentosa.
The phototransduction cascade is initiated through similar photosensitive opsins in mammals and Drosophila.
The opsins then influence divergent cellular second messenger cascades. In vertebrates, the light signal activates
a heterotrimeric G-protein that then stimulates cGMP
phosphodiesterase. Decreased cGMP levels lead to the
closure of plasma membrane cation channels and photoreceptor hyperpolarization. In contrast, the light response in Drosophila involves activation of phospholipase C, production of inositol triphosphate, calcium
release, and eventual depolarization of the photoreceptor.
Mutations in a number of the proteins involved in the
Drosophila phototransduction cascade produce retinal degeneration: phospholipase C (norpA), diacylglycerol kinase (rdgA), phosphatidylinositol transfer protein (rdgB),
and a calcium-dependent serine/threonine protein phosphatase (rdgC). Mammalian homologs of many of these
proteins have been identified. Mammalian rdgB can substitute functionally for the fly protein, and is specifically
expressed in the retina (40). A number of retinal diseases
colocalize genetically with mammalian rdgB (41). A human retinal-restricted phosphatase, shares unique sequence motifs with rdgC (42). Despite differences in the
phototransduction cascade, homologous proteins are
clearly used in the same organ and are attractive disease
gene candidates. The parallel with the eyeless/aniridia
story is striking (reviewed in 43). Mutations in the Pax6 transcription factor cause eyeless phenotypes in humans
(Aniridia), mice (Small eye), and Drosophila (eyeless).
Strikingly, both vertebrate and invertebrate Pax-6 can induce eye formation when expressed ectopically (44). The
same transcription factor directs the development of a
J Neuropathol Exp Neurol, Vol 59, October, 2000
homologous organ, despite the marked structural dissimilarities between eyes in flies and people.
‘‘Failures’’: Amyotrophic Lateral Sclerosis and Prion
Diseases
Attempts have been made to model both amyotrophic
lateral sclerosis (ALS) and prion diseases in flies. Both
diseases appear to be caused by a toxic dominant gain of
function mechanism and appear ideal candidates for modeling in invertebrate systems. However, successful models have not yet been produced. Dominant mutations in
the Cu/Zn superoxide dismutase (SOD) gene cause about
20% of familial ALS (reviewed in 45). Several lines of
evidence support a novel toxic gain of function mechanism for pathogenic SOD mutations. Expression of mutant SOD proteins in transgenic mice creates a motor neuron degeneration syndrome with many similarities to the
human disease. Mice with no endogenous Cu/Zn SOD
are healthy and have no spontaneous abnormality of motor neurons. Some SOD-linked mutant proteins can inhibit the activity of normal SOD, but a dominant gain of
function mechanism of disease pathogenesis appears unlikely because neuron disease produced by expression of
mutant SOD is not influenced by endogenous SOD protein. Progression and pathologic features of the disease
are the same regardless of levels of normal SOD. The
precise nature of the toxicity endowed by ALS-linked
mutations remains unclear. Increased peroxidase activity
of mutant enzymes has been described, and may relate to
motor neuron dysfunction (46).
Modeling SOD-related pathologies in invertebrates appears an attractive prospect. Both C. elegans and Drosophila have clear Cu/Zn SOD orthologs. A bovine transgene can rescue SOD null mutations in Drosophila (47),
showing functional as well as sequence homology. In
both organisms, SOD can clearly influence the regulation
of lifespan (48, 49). Missense mutations in Drosophila
SOD produce neurodegeneration (50). The retinas of flies
carrying a biochemically null SOD missense mutation
develop normally, but show marked degeneration of neuronal and nonneuronal cells after 7 days. Effects on motor
neurons and other areas of the nervous system were not
reported, nor was the presence of inclusion bodies. Point
mutations, but not deletions, show biochemical evidence
of a dominant negative effect. However, unlike murine
ALS models, normal SOD rescues, rather than enhances,
the phenotypic abnormalities seen in the mutant flies. Endogenous point mutations in the Drosophila SOD gene
cause neurodegeneration, a feature of ALS. In addition,
mutant enzymes show a dominant negative biochemical
effect that is observed in selected human disease-linked
SOD mutations. Overall, however, mutations in the endogenous fly SOD gene fail to replicate key aspects of
the human disease. The fly mutations are recessive and
NEURODEGENERATIVE DISEASE IN INVERTEBRATES
there is no evidence for a toxic effect of the mutant proteins independent of the normal SOD enzyme. Drosophila SOD may be different enough from the human enzyme that analogous toxic properties cannot be endowed
on the fly enzyme. Alternatively, the genetic selection
used to isolate the Drosophila mutants may bear little
relationship to the factors producing pathogenic human
mutations.
If human mutations are indeed a special set of genetic
alterations endowing a specific toxic property on the abnormal enzymes, expressing the mutant human SOD
might model ALS in invertebrates. Transgenic mice expressing disease-linked SOD mutant proteins develop a
motor neuron degeneration syndrome that resembles
ALS. One human mutant protein has been expressed in
Drosophila (51). The G41S mutation was expressed in
Drosophila motor neurons using the control of elements
specific for motor neurons. No abnormalities were detected in flies producing the human mutant. The results
were not altered by the presence or absence of endogenous Drosophila SOD, or by the transgenic expression
of normal human SOD. Interestingly, Drosophila lifespan
can be increased by expression of normal human SOD
selectively in motor neurons (52). Motor neuron senescence may be a key factor limiting fly survival, and by
analogy other organisms as well. Like the normal protein,
human SOD carrying the disease-linked G41S mutation
increases fly lifespan.
In contrast to the results modeling polyglutamine disorders in invertebrates by expressing a toxic human protein, the Drosophila ALS model appears disappointing.
Neither mutations in the endogenous fly SOD gene nor
expression of a toxic human transgenes recapitulate key
features of the human disorder. However, before abandoning the invertebrate approach to ALS entirely, several
points should be addressed. First, only a single SOD mutant was assayed in Drosophila. Age of onset and severity
of disease vary substantially in transgenic mice expressing different disease-linked SOD mutations. Given the
short life spans of flies and worms, and the many other
differences between flies, worms, and people, a ‘‘supertoxic’’ SOD mutant might be required to model ALS. In
C. elegans, a very long repeat length Huntingtin allele is
needed to produce effects. Second, expression of mutant
SOD might be more toxic in other tissues. Although the
primary pathology in ALS involves upper and lower motor neurons, the cell type in which mutant SOD exerts its
effects is unclear. Of course, cofactors necessary for the
development of human disease may simply be absent or
inadequately conserved in invertebrates. Perceived and
actual differences between the nervous systems of mammals and invertebrates have limited the application of
simple genetic systems to human disease in the past and
represent a legitimate concern.
853
As for many of the neurodegenerative disorders discussed above, prion protein pathogenesis probably also
involves a toxic, novel property of the abnormal protein.
However, at least 1 attempt to model spongiform encephalopathy in Drosophila was unsuccessful (53). Expression of wild type Syrian hamster prion protein produced no detectable phenotype in flies. In addition, no
protease resistant protein was seen. Prion proteins from
other species, and mutant forms of prion proteins, might
prove more toxic. In the experiments expressing hamster
prion protein in flies, repeated heat shocks were used to
drive expression of a transgene controlled by a heat shock
promotor. An expression system that does not rely on
heat shock may be more suitable because induced heat
shock proteins could act as chaperones to mitigate the
effect of abnormally folded proteins.
Mutations in Invertebrate Genes Producing
Neurodegeneration
A number of mutants exist in flies and worms that were
selected specifically for adult onset degeneration of the
brain. These mutants represent an important complement
to transgenic animals created to model specific diseases.
One of the most powerful features of genetic analysis is
the unbiased nature of the approach. In theory, any gene
product that contributes to a particular cellular process
should be identifiable by random mutagenesis. Practical
limitations include cases in which mutations of 1 copy
of a gene are lethal, and are therefore difficult to recover.
Genetic redundancy represents a more common limitation. Functional complementation by related proteins may
mask relevant phenotypes, even though the encoded gene
product is important in the process being studied. In spite
of these caveats, genetic dissection remains one of the
best methods of nonhypothesis driven analysis.
The Drosophila mutant bubblegum is the only mutant
selected for neurodegeneration that has shown a biochemical or molecular relationship to a human disease
gene product (54). bubblegum was originally isolated on
the basis of reduced lifespan. Histologic analysis of the
brain reveals degeneration of the retina and optic lobes.
The visual system is normal at eclosion, but over time
axonal swelling, neuropil vacuolization, and cell loss are
seen. Cloning the bubblegum gene revealed a connection
to human disease. bubblegum corresponds to very long
chain fatty acid (VLCFA) acyl coenzyme A synthetase.
Intriguingly, reduced VLCFA acyl coenzyme A synthetase activity is present in adrenoleukodystrophy patients.
Very long chain fatty acid levels are elevated in bubblegum flies. In addition, feeding flies with monounsaturated
fatty acids (‘‘Lorenzo’s oil’’) normalized VLCFA levels,
as is seen in patients with adrenoleukodystrophy. Remarkably, feeding the same fatty acids to developing flies
J Neuropathol Exp Neurol, Vol 59, October, 2000
854
FEANY
not only decreased VLCFA levels, but prevented neurodegeneration as well. Attempts to prevent or retard manifestations of adrenoleukodystrophy in patients by lowering VLCFA levels have so far been unrewarding.
Given the biochemical similarities, bubblegum has
been proposed as a model for adrenoleukodystrophy.
Similarities between the Drosophila mutant and the human disease include elevation of VLCFAs, normalization
of fatty acid levels with dietary treatment, and loss of
VLCFA acyl coenzyme A synthetase activity. However,
VLCFAs are increased in not only adrenoleukodystrophy,
but in 9 of 15 other known peroxisome disorders, including Zellweger’s syndrome, acyl coenzyme A oxidase
deficiency, and neonatal adrenoleukodystrophy. Adrenoleukodystrophy is not linked genetically to VLCFA acyl
coenzyme A synthetase. Instead, mutations in a transmembrane transporter of the ABC superfamily whose
function and substrate specificity cause the disease. There
is no apparent correlation between VLCFA levels in patients (or mice with inactivation of the corresponding
ABC-type transporter) and disease symptoms, progression, or neuropathology. Finally, the pathology of the
Drosophila mutants bears little resemblance to the human
disorder. Adrenoleukodystrophy manifests as a disease of
myelinating glia in the central and peripheral nervous
system. Neuronal loss occurs late in the human disorder
and may be secondary to white matter damage. Ultrastructural examination of tissue from patients with adrenoleukodystrophy reveals characteristic trilaminar inclusions composed of accumulated fatty acids. Similar
structures were not observed in the bubblegum mutant
flies. bubblegum mutants do show peroxisomal dysfunction associated with neuronal degeneration. Perhaps the
mutant may best serve as a general model of peroxisomal
disorders. Prevention of neurodegeneration in the fly
model by biochemical normalization suggests that the
strategy may eventually be successful a human disease,
despite the poor results to date.
Direct examination of brain histology has also been
used to isolate Drosophila neurodegeneration mutants.
The swiss cheese mutant appeared in a screen for mutants
showing anatomical brain defects in the adult fly (55).
Histologic sections show widespread vacuoles that increase in number over time. Lifespan is decreased by
approximately two thirds to one half in various mutant
alleles. The first ultrastructural abnormality seen is abnormal, multilayered wrapping of neurons by glial processes. Subsequently, neuronal loss and widespread neuropil vacuolization occurs. Although the initial pathology
present in swiss cheese mutant flies is glial, the gene
product is present primarily in neurons (56). The glial
abnormalities observed could be secondary to neuronal
pathology. In humans, concentric Schwann cell hypertrophy, or so-called ‘‘onion bulb’’ formation, is prominent
J Neuropathol Exp Neurol, Vol 59, October, 2000
in primary disorders of the myelin sheath like CharcotMarie-Tooth disease. However, concentric hypertrophy is
a nonspecific response to injury that may be seen in a
wide range of acquired and inherited disorders.
The swiss cheese gene encodes a novel protein with 2
domains homologous to the regulatory subunit of protein
kinase A. The swiss cheese protein also shows similarity
to human neuropathy target esterase, a protein implicated
in neuronal degeneration produced by organophosphates.
Although the precise role of the protein remains unclear,
these studies strongly implicate the swiss cheese protein
in the neurodegenerative process.
Glial abnormalities have also been documented in the
Drosophila neurodegenerative mutant drop-dead (57).
Glia in drop-dead mutant brains demonstrate shortened
cell processes rather than the excess wrapping observed
in the swiss cheese mutant. As for swiss cheese mutants,
glia in drop-dead flies are affected prior to the onset of
neuronal loss. The nature of the protein encoded by the
drop-dead locus, and its expression pattern, will be help
define the precise role of glial pathology to neurodegeneration in the drop-dead mutant. It is intriguing that 2
independently isolated Drosophila neurodegeneration
mutants appear to exert their primary effect on glial cells.
These results suggest that brain degeneration in Drosophila may be fundamentally different in some aspects
than central nervous system degenerations in humans. Alternatively, perhaps early glial pathology should be reevaluated in common human neurodegenerative diseases.
C. elegans has provided one of the most satisfying
models of neurodegeneration from the mechanistic point
of view. Dominant mutations in 3 worm genes, deg-1,
mec-4, and mec-10, cause degeneration of touch-receptor
neurons (58). The 3 C. elegans degenerin genes encode
proteins with homology to subunits of amiloride-sensitive
sodium channels found in epithelia and various types of
neurons. Identification of the degenerins as channel components suggested that cell death results from dysregulated ion flux through the mutant channels. This hypothesis is supported by the observation that mammalian
degenerin is activated by mutations causing degeneration
in C. elegans (59). Mammalian degenerin remains an attractive disease gene candidate, but no linkage to a human disorder has yet been reported.
Ion channel abnormalities can also produce muscle disease in worms and humans. The C. elegans degenerin
unc-105 forms cation ion channels. These channels are
activated by mutations that cause muscle degeneration or
hypercontraction (60). Although a direct unc-105 mammalian homolog has not been described, a variety of human myopathies, including hyperkalemic periodic paralysis and myotonia congenita, are caused by mutations in
ion channels.
NEURODEGENERATIVE DISEASE IN INVERTEBRATES
Summary
Promising new models now exist for neurodegeneration diseases in flies and worms. The most successful
models have used direct expression of toxic human proteins in invertebrate organisms to model the cognate human disorders. These models faithfully replicate many
aspects of the human syndromes and provide great promise for further genetic dissection of fundamental pathogenetic mechanisms. In addition, small, short-lived flies
and worms may be ideal drug testing substrates. As the
biochemical and cell biological similarities between vertebrates and invertebrates continue to emerge, additional
fly and worm disease models will doubtless continue to
enrich the study of neurodegenerative disease.
REFERENCES
1. Rubin GM, Yandell MD, Wortman JR, et al. Comparative genomics
of the eukaryotes. Science 2000;287:2204–15
2. Polymeropoulos MH, Lavedan C, Leroy E, et al. Mutation in the
a-synuclein gene identified in families with Parkinson’s disease.
Science 1997;276:2045–47
3. Krüger R, Kuhn W, Müller T, et al. Ala30Pro mutation in the gene
encoding a-synuclein in Parkinson’s disease. Nat Genet 1998;18:
106–8
4. Feany MB, Bender WW. A Drosophila model of Parkinson’s disease. Nature 2000;404:394–98
5. Masliah E, Rockenstein E, Veinbergs I, et al. Dopaminergic loss
and inclusion body formation in a-synuclein mice: Implications for
neurodegenerative disorders. Science 2000;287:1265–69
6. Warrick JM, Paulson HL, Gray-Board GL, et al. Expanded polyglutamine protein forms nuclear inclusions and causes neural degeneration in Drosophila. Cell 1998;93:939–49
7. Jackson GR, Salecker I, Dong X, et al. Polyglutamine-expanded
human huntingtin transgenes induce degeneration of Drosophila
photoreceptor neurons. Neuron 1998;21:633–42
8. Faber PW, Alter JR, MacDonald ME, Hart AC. Polyglutamine-mediated dysfunction and apoptotic death of a Caenorhabditis elegans
sensory neuron. Proc Natl Acad Sci USA 1999;96:179–84
9. Bates GP, Mangiarini L, Davies SW. Transgenic mice in the study
of polyglutamine repeat expansion diseases. Brain Pathol 1998;8:
699–714
10. Davies SW, Turmaine M, Cozens BA, et al. Formation of neuronal
intranuclear inclusions underlies the neurological dysfunction in
mice transgenic for the HD mutation. Cell 1997;90:537–48
11. Klement IA, Skinner PJ, Kaytor MD, et al. Ataxin-1 nuclear localization and aggregation: role in polyglutamine-induced disease in
SCA1 transgenic mice. Cell 1998;95:41–53
12. Saudou F, Finkbeiner S, Devys D, Greenberg ME. Huntingtin acts
in the nucleus to induce apoptosis but death does not correlate with
the formation of intranuclear inclusions. Cell 1998;95:55–66
13. Warrick JM, Chan HY, Gray-Board GL, Chai Y, Paulson HL, Bonini NM. Suppression of polyglutamine-mediated neurodegeneration
in Drosophila by the molecular chaperone HSP70. Nat Genet 1999;
23:425–28
14. Kazemi-Esfarjani P, Benzer S. Genetic suppression of polyglutamine toxicity in Drosophila. Science 2000;287:1837–40
15. Zoghbi HY, Orr HT. Polyglutamine diseases: Protein cleavage and
aggregation. Curr Opin Neurobiol 1999;9:566–70
16. Artavanis-Tsakonas S, Rand MD, Lake RJ. Notch signaling: Cell
fate control and signal integration in development. Science 1999;
284:770–76
855
17. Levy-Lahad E, Wijsman EM, Nemens E, et al. A familial Alzheimer’s disease locus on Chromosome 1. Science 1995;269:970–77
18. Rogaev EI, Sherrington R, Roageava EA, et al. Familial Alzheimer’s disease in kindreds with missense mutations in a gene on
chromosome 1 related to the Alzheimer’s disease type 3 gene. Nature 1995;376:775–78
19. Sherrington R, Rogaev EI, Liang Y, et al. Cloning of a gene bearing
missense mutations in early-onset familial Alzheimer’s disease. Nature 1995;375:754–60
20. Levitan D, Greenwald I. Facilitation of lin-12-mediated signalling
by sel-12, a Caenorhabditis elegans S182 Alzheimer’s disease
gene. Nature 1995;377:351–54
21. Swiatek PJ, Lindsell CE, del Amo FF, Weinmaster G, Gridley T.
Notch1 is essential for postimplantation development in mice.
Genes Dev 1994;8:707–19
22. Shen J, Bronson RT, Chen DF, Xia W, Selkoe DJ, Tonegawa S.
Skeletal and CNS defects in Presenilin-1-deficient mice. Cell 1997;
89:629–39
23. De Strooper B, Annaert W, Cupers P, et al. A presenilin-1-dependent gamma-secretase-like protease mediates release of Notch intracellular domain. Nature 1999;398:518–22
24. Struhl G, Greenwald I. Presenilin is required for activity and nuclear access of Notch in Drosophila. Nature 1999;398:522–25
25. Ye Y, Lukinova N, Fortini ME. Neurogenic phenotypes and altered
Notch processing in Drosophila Presenilin mutants. Nature 1999;
398:525–29
26. Wolfe MS, Xia W, Ostaszewski BL, Diehl TS, Kimberly WT, Selkoe DJ. Two transmembrane aspartates in presenilin-1 required for
presenilin endoproteolysis and gamma-secretase activity. Nature
1999;398:513–17
27. Anderton BH. Alzheimer’s disease: Clues from flies and worms.
Curr Biol 1999;9:106–9
28. Zhang Z, Hartmann H, Do VM, et al. Destabilization of beta-catenin by mutations in presenilin-1 potentiates neuronal apoptosis.
Nature 1998;395:698–702
29. Takashima A, Murayama M, Murayama O, et al. Presenilin 1 associates with glycogen synthase kinase-3beta and its substrate tau.
Proc Natl Acad Sci USA 1998;95:9637–41
30. Joutel A, Corpechot C, Ducros A, et al. Notch3 mutations in CADASIL, a hereditary adult-onset conditioncausing stroke and dementia. Nature 1996;383:707–10
31. Ellisen LW, Bird J, West DC, et al. TAN-1, the human homolog of
the Drosophila notch gene, is broken by chromosomal translocations in T lymphoblastic neoplasms. Cell 1991;66:649–61
32. Li L, Krantz ID, Deng Y, et al. Alagille syndrome is caused by
mutations in human Jagged1, which encodes a ligand for Notch1.
Nat Genet 1997;16:243–51
33. Martin-Morris LE, White K. The Drosophila transcript encoded by
the beta-amyloid protein precursor-like gene is restricted to the nervous system. Development 1990;110:185–95
34. Daigle I, Li C. apl-1, a Caenorhabditis elegans gene encoding a
protein related to the human beta-amyloid protein precursor. Proc
Natl Acad Sci USA 1993;90:12045–49
35. Link CD. Expression of human beta-amyloid peptide in transgenic
Caenorhabditis elegans. Proc Natl Acad Sci USA 1995;92:9368–72
36. Fay DS, Fluet A, Johnson CJ, Link CD. In vivo aggregation of
beta-amyloid peptide variants. J Neurochem 1998;71:1616–25
37. Dryja TP, McGee TL, Reichel E, et al. A point mutation of the
rhodopsin gene in one form of retinitis pigmentosa. Nature 1990;
343:364–66
38. Steele F, O’Tousa JE. Rhodopsin activation causes retinal degeneration in Drosophila rdgC mutant. Neuron 1990;4:883–90
39. Davidson FF, Steller H. Blocking apoptosis prevents blindness in
Drosophila retinal degeneration mutants. Nature 1998;391:587–91
J Neuropathol Exp Neurol, Vol 59, October, 2000
856
FEANY
40. Chang JT, Milligan S, Li Y, et al. Mammalian homolog of Drosophila retinal degeneration B rescues the mutant fly phenotype. J
Neurosci 1997;17:5881–90
41. Guo J, Yu FX. Cloning and characterization of human homologue
of Drosophila retinal degeneration B: A candidate gene for degenerative retinal diseases. Dev Genet 1997;20:235–45
42. Montini E, Rugarli EI, Van de Vosse E, et al. A novel human
serine-threonine phosphatase related to the Drosophila retinal degeneration C (rdgC) gene is selectively expressed in sensory neurons of neural crest origin. Hum Mol Genet 1997;6:1137–45
43. Gehring WJ. The master control gene for morphogenesis and evolution of the eye. Genes Cells 1996;1:11–5
44. Halder G, Callaerts P, Gehring WJ. Induction of ectopic eyes by
targeted expression of the eyeless gene in Drosophila. Science
1995;267:1788–92
45. Tu PH, Gurney ME, Julien JP, Lee VM, Trojanowski JQ. Oxidative
stress, mutant SOD1, and neurofilament pathology in transgenic
mousemodels of human motor neuron disease. Lab Invest 1997;76:
441–56
46. Wiedau-Pazos M, Goto JJ, Rabizadeh S, et al. Altered reactivity of
superoxide dismutase in familial amyotrophic lateral sclerosis. Science 1996;271:515–18
47. Reveillaud I, Phillips J, Duyf B, Hilliker A, Kongpachith A, Fleming JE. Phenotypic rescue by a bovine transgene in a Cu/Zn superoxide dismutase-null mutant of Drosophila melanogaster. Mol
Cell Biol 1994;14:1302–7
48. Phillips JP, Campbell SD, Michaud D, Charbonneau M, Hilliker
AJ. Null mutation of copper/zinc superoxide dismutase in Drosophila confers hypersensitivity to paraquat and reduced longevity.
Proc Natl Acad Sci USA 1989;86:2761–65
49. Larsen PL. Aging and resistance to oxidative damage in Caenorhabditis elegans. Proc Natl Acad Sci USA 1993;90:8905–9
J Neuropathol Exp Neurol, Vol 59, October, 2000
50. Phillips JP, Tainer JA, Getzoff ED, Boulianne GL, Kirby K, Hilliker
AJ. Subunit-destabilizing mutations in Drosophila copper/zinc superoxide dismutase: Neuropathology and a model of dimer dysequilibrium. Proc Natl Acad Sci USA 1995;92:8574–78
51. Elia AJ, Parkes TL, Kirby K, et al. Expression of human FALS
SOD in motorneurons of Drosophila. Free Radic Biol Med 1999;
26:1332–38
52. Parkes TL, Elia AJ, Dickinson D, Hilliker AJ, Phillips JP, Boulianne
GL. Extension of Drosophila lifespan by overexpression of human
SOD1 in motorneurons. Nat Genet 1998;19:171–74
53. Raeber AJ, Muramoto T, Kornberg TB, Prusiner SB. Expression and
targeting of Syrian hamster prion protein induced by heat shock in
transgenic Drosophila melanogaster. Mech Dev 1995;51:317–27
54. Min KT, Benzer S. Preventing neurodegeneration in the Drosophila
mutant bubblegum. Science 1999;284:1985–88
55. Heisenberg M, Böhl K. Isolation of anatomical brain mutants of
Drosophila by histological means. Z Naturforsch [C] 1979;34:143–47
56. Kretzschmar D, Hasan G, Sharma S, Heisenberg M, Benzer S. The
swiss cheese mutant causes glial hyperwrapping and brain degeneration in Drosophila. J Neurosci 1997;17:7425–32
57. Buchanan RL, Benzer S. Defective glia in the Drosophila brain
degeneration mutant drop-dead. Neuron 1993;10:839–50
58. Chalfie M, Wolinsky E. The identification and suppression of inherited neurodegeneration in Caenorhabditis elegans. Nature 1990;
345:410–16
59. Waldmann R, Champigny G, Voilley N, Lauritzen I, Lazdunski M.
The mammalian degenerin MDEG, an amiloride-sensitive cation
channel activated by mutations causing neurodegeneration in Caenorhabditis elegans. J Biol Chem 1996;271:10433–36
60. Garcia-Anoveros J, Garcia JA, Liu JD, Corey DP. The nematode
degenerin UNC-105 forms ion channels that are activated by degeneration- or hypercontraction-causing mutations. Neuron 1998;
20:1231–41