Cosponsors of this publication
CONTENTS
Produced by the Science/AAAS Custom Publishing Office
INTRODUCTIONS
2 Understanding the Aging Brain By Alan Leshner
3 Neurodegenerative Research in China: Promise and Challenge
THERAPEUTIC STRATEGIES
NEURODEGENERATION RESEARCH
48 Deep-Brain Stimulation Research in China
4 Molecular Mechanisms Underlying Neuronal Death and
Synaptic Dysfunction in Neurodegenerative Diseases
Kwok-On Lai and Nancy Y. Ip
Dong Chen, Haigang Ren, Qingsong Hu, Feng Gao,
and Guanghui Wang
Rui Sheng and Zheng-Hong Qin
Ya Ke, Ka-Chun Wu, and Zhong-Ming Qian
16 Novel Pathways Regulating Function and Metabolism of ß-Amyloid Precursor Protein in Alzheimer’s Disease
Yun-Wu Zhang, Guojun Bu, and Huaxi Xu
Yingjun Liu and Jiawei Zhou
Chengyuan Tang, Tongmei Zhang, and Zhuohua Zhang
27 Animal Models of Huntington’s Disease
Xiao-Jiang Li and Shihua Li
29 FMRP Regulates Microtubule Network Formation, Neurogenesis, and DNA Damage Response
59 Novel Therapeutic Strategies for Amyotrophic
Lateral Sclerosis
25 Proteolytic Pathways in Parkinson’s Disease
Yan Zheng, Jun Jia, and Xiao-Min Wang
Weidong Le and Xiaojie Zhang
62 Novel Neuroprotective Strategy for Stroke—
Activating Inherent Neuronal Survival Mechanisms
Jian-Zhi Wang, Qing Tian, and Di Gao
22 Dopaminergic Modulation of Astrocyte Functions in the
Pathogenesis of Parkinson’s Disease
18 Role of Tau Hyperphosphorylation in Alzheimer’s
Disease-Associated Neurodegeneration
Jun Liu and Sheng-Di Chen
56 Exploring Therapeutic Mechanisms of Traditional
Chinese Medicine Extracts and Electroacupuncture
in Parkinson’s Disease
13 Iron and Neurodegeneration
Changhai Tian and Jialin C. Zheng
53 Potential Neuroprotective Therapies for
Parkinson’s Disease
10 The Multiple Roles of Autophagy in Neural Function
and Disease
Bomin Sun and Wei Liu
50 Neural Progenitors by Direct Reprogramming:
Strategies for the Treatment of Parkinson’s and
Alzheimer’s Diseases
7 The Role of Autophagy, Apoptosis, Neuroinflammation,
and Mitochondrial Dysfunction in Parkinson’s Disease
By Xiong-Li Yang, Zhuohua Zhang, and Beisha Tang
Aiyu Yao and Yong Q. Zhang
GENETIC AND CLINICAL STUDIES
Ming Chen, Bo-Xing Li, and Tian-Ming Gao
About the Cover: The
ink and wash style of
painting (Shui-mo) was
developed about 1,500
years ago in China. In
this painting, the artist
has used a withered
tree to signify degenerating neurons. The red
characters (in ancient
Chinese calligraphy,
Zhuan-shu) translate
as “neurodegenerative diseases.” Credit:
Zaihao Hu
32 Progress in Treating Hereditary Ataxia in Mainland China
Hong Jiang, Junling Wang, Juan Du, Ranhui Duan,
Jiada Li, and Beisha Tang
35 Genetic Etiology of Parkinson’s Disease in China
Jifeng Guo, Xinxiang Yan, Danling Wang, Qian Xu, Beisha Tang, and Zhuohua Zhang
37 PKD and PRRT2-related Paroxysmal Diseases in China
Jun-Ling Wang, Nan Li, Hong Jiang, Lu Shen, Kun Xia, and Beisha Tang
41 Clinical and Molecular Biological Studies of Amyotrophic
Lateral Sclerosis in China
Zhangyu Zou and Liying Cui
43 Brainnetome Studies of Alzheimer’s Disease Using Neuroimaging
Tianzi Jiang, Yong Liu, and Bing Liu
46 Fragile X Syndrome in China: A Clinical Review
Ranhui Duan
This booklet was produced by the Science/AAAS Custom Publishing
Office and sponsored by the State Key Laboratory of Medical Genetics
in China. Materials that appear in this booklet were commissioned, edited,
and published by the Science/AAAS Custom Publishing Office and were
not reviewed or assessed by the Science Editorial staff. Articles in this
booklet can be cited using the following format: [AUTHOR NAME(S)],
[ARTICLE TITLE] in Pathways to Cures: Neurodegenerative Diseases in
China, S. Sanders, Z. Zhang, B. Tang, Eds. (Science/AAAS, Washington,
DC, 2013), pp. [xx-xx].
Editors: Sean Sanders, Ph.D.; Zhuohua Zhang, Ph.D.; Beisha Tang, M.D.
Assistant Editor: Jia-Da Li, Ph.D.
Proofreader/Copyeditor: Yuse Lajiminmuhip; Designer: Amy Hardcastle
© 2013 by The American Association for the Advancement of Science.
All rights reserved. 20 December 2013
1
Produced by the Science/AAAS Custom Publishing Office
Understanding the Aging Brain
As treatments
improve,
survival to old
age becomes
more common,
bringing about a
shift from diseases of youth
to diseases of
old age.
2
In most cultures around the
world, the brain is recognized to
be the center of our minds and
our personality, defining who,
and even what, we are. We also
depend on our brains to help us
navigate a complex world of social interactions, physical obstacles, and emotional challenges. It
is therefore understandable that
diseases affecting our cognitive
ability are regarded as among
the most devastating and debilitating as well as being frightening
on a personal level.
Standards of health care are
changing rapidly across the
globe and probably nowhere
more quickly than in China. As
treatments improve, survival to
old age becomes more common,
bringing about a shift from diseases of youth to diseases of old
age, including a spectrum of neurodegenerative diseases such as
Parkinson’s, Alzheimer’s, amyotrophic lateral sclerosis, and Huntington’s. For this reason, there is
some urgency among researchers to gain a better understanding of the etiology and biological
basis of these disorders in order
to affect preventative measures
or even possible cures as well
as find ways to improve the quality of life for current and future
sufferers.
This booklet looks at the current state of neurodegenerative research in China, outlining
recent progress in the field and
future prospects. It has been
divided into three sections that
mirror the progress of research
from discovery in the laboratory
to clinical testing to cure. The
first section provides a sampling of the promising basic research coming out of China, highlighting potential targets
for deeper investigation and for possible development of
therapeutic interventions. Section two covers progress
in understanding the genetic basis of neurodegenerative
disorders as well as reviewing some recent clinical studies that are helping researchers understand which treatments might be most effective and why. The third and final
section of this booklet highlights some promising therapies for a number of neurodegenerative disorders, including deep brain stimulation and certain traditional Chinese
medicines.
It is apparent that a few broad conclusions can be drawn
from the work contained in this booklet. First, it is clear
that although the brains from autopsied disease sufferers
provide many clues, they are a poor substitute for being
able to observe the disease progress in real-time. The
development of better, noninvasive or minimally invasive
monitoring technologies is therefore a priority, as is the
creation of more physiologically and biologically accurate
animal models. It is clear that in China, progress is already
being made in both areas. Second, researchers need to
look more broadly for both targets and cures. Autophagy,
a process more recently accepted as playing a key role
in neurodegeneration, promises to provide a number of
potential drug targets. Traditional
Chinese medicine seems to be
providing some additional promising potential therapeutics that warrant deeper investigation, with a
number having been used to treat
the symptoms of these diseases
for centuries, apparently with good
effect.
It is likely that the research world
will follow with interest the progress in China over the coming decades as the country aggressively
pushes its research agenda, determined to make its mark in the field
of neurodegenerative diseases,
amongst others.
Alan Leshner, Ph.D.
CEO, AAAS
Executive Publisher, Science
Neurodegenerative Research in China:
Promise and Challenge
The population of China is aging. An ever-increasing number of Chinese are afflicted with neurodegenerative diseases that result from the gradual and progressive loss of
neural cells, leading to nervous system dysfunction. Despite decades of basic and clinical endeavors, understanding of the etiology of these diseases and finding cures remains a great challenge.
It is well accepted that genetic factors contribute to the
pathogenesis of neurodegenerative diseases. Genetic
studies have defined the incidence of known diseaserelated genes in the Chinese population and identified a
number of new causative genes. These findings lay the
foundation for the design of genetic diagnoses of neurodegenerative diseases and the basis for dissecting and understanding the pathological mechanisms. Unfortunately,
neurodegenerative diseases are most often genetically
complex, with only a few being monogenic. The real difficulty therefore is defining the genetic background of the
sporadic cases. Unbiased genome-wide association studies have identified a number of novel loci with associations to Alzheimer’s and Parkinson’s diseases, providing a
possible explanation for the heritability of certain diseases.
Whole genome and exome analysis through next generation sequencing will likely generate further insights into the
molecular pathways.
Results from human genetic
studies offer an entry point to define the molecular etiology of neurodegenerative diseases. Functional investigations suggest that
the causative genes in many neurodegenerative diseases appear to
be distinct, but are in fact functionally linked, indicating a potential
disease mechanism. In Alzheimer’s disease, multiple causative
and susceptibility genes regulate
β-amyloid metabolism, emphasizing the important role of this protein
in pathogenesis. Genes found to
be related to Parkinson’s disease
play roles in maintaining mitochondrial homeostasis and modulating
protein degradation, highlighting
the involvement of these essential
cellular pathways in the pathogen-
esis of this disease. Recently,
neurodegenerative diseases have
been shown to involve not only
neurons, but also the supporting
cells such as astrocytes. These
studies open new avenues of research into disease etiology and
provide potential targets for treatment.
The eventual goal of neurodegenerative disease research is
to develop effective therapies.
Different strategies are currently
being pursued by Chinese scientists, ranging from identification
of neuroprotective molecules to
the generation of cells for neuronal replacement. Notably, many
novel compounds from Chinese
herbal medicines appear to
show promise in, for example,
improving dementia caused by
Alzheimer’s disease and relieving the symptoms of Parkinson’s
disease.
There are still many questions to
be answered before neurodegenerative disease can be regarded
as fully treatable. Nevertheless,
incremental scientific advances
over the coming years will likely
continue to challenge our previous thinking and shift the paradigm of current treatments. The
pursuit of a deeper understanding of the cellular and genetic
processes involved in neurodegenerative diseases will undoubtedly prove beneficial to patients in
China and worldwide.
Xiong-Li Yang1
Zhuohua Zhang2
Beisha Tang2,3
Results from
human genetic
studies offer
an entry point
to define the
molecular
etiology of
neurodegenerative diseases.
Institute of Neurobiology,
Institutes of Brain Science and
State Key Laboratory of Medical
Neurobiology, Fudan University,
Shanghai, China
2
State Key Laboratory of Medical
Genetics, Xiangya Medical School,
Central South University, Changsha,
Hunan, China
3
Department of Neurology, Xiangya
Hospital, Central South University,
Changsha, Hunan, China
1
3
NEURODEGENERATION RESEARCH
Produced by the Science/AAAS Custom Publishing Office
Molecular Mechanisms Underlying Neuronal Death and
Synaptic Dysfunction in Neurodegenerative Diseases
Kwok-On Lai and Nancy Y. Ip*
The number of people afflicted with neurodegenerative diseases
will increase as the population ages. The National Institute of
Aging at the U.S. National institutes of Health estimated that 524
million people were aged 65 or older in 2010 (8% of the world’s
population), and the aging population is projected to grow to
nearly two billion by 2050 (16% of the world’s population), with
most of the increase in developing countries (www.nia.nih.gov/
research/publication/global-health-and-aging/humanitys-aging).
The implications of an aging population are profound: as this
segment of the population lives long enough to suffer serious
age-related chronic diseases, it will spur a massive demand for
health care.
Among the brain diseases and disorders in the elderly, neurodegenerative diseases such as Alzheimer’s disease (AD) and
Parkinson’s disease (PD) represent a leading cause of mortality.
Although therapeutic drugs for these diseases are available, they
merely alleviate the symptoms and, in some cases, are only effective in the early stages of the disease. The current obstacle to
developing effective treatments for these neurodegenerative diseases is the lack of understanding of disease pathophysiology.
Delineating the complex biological processes at the molecular
level within the diseased brain is therefore crucial to uncovering
new drug targets against which more effective therapeutic drugs
can be developed. This article summarizes our efforts to elucidate the molecular mechanisms that underlie the neuronal death
and synaptic dysfunction in these diseases, with particular emphasis on the role of a key serine/threonine kinase in the brain:
cyclin-dependent kinase 5 (Cdk5).
Alzheimer’s Disease
AD is the most common form of dementia, characterized by loss
of cognitive function and memory. The disease generally affects
individuals over the age of 60, and almost half of the population over the age of 85 eventually succumbs to it. AD is clinically
characterized by progressive cognitive impairment and memory
loss (1). The key pathologies of the patient’s postmortem brains
include extensive neuronal loss, disposition of senile plaques
that consist of the accumulation of b-amyloid peptides (Ab), and
Division of Life Science, Molecular Neuroscience Center and
State Key Laboratory of Molecular Neuroscience, The Hong
Kong University of Science and Technology, Hong Kong
*
Corresponding Author: [email protected]
4
appearance of neurofibrillary tangles, which are composed of
hyperphosphorylated tau. Emerging evidence indicates that dysregulation of Cdk5 activity is one of the major culprits responsible
for neuronal death and loss of synapses in AD (2).
Role of Cdk5 in Tau Hyperphosphorylation
and Neuronal Death
The catalytic activity of Cdk5 requires its binding to either one
of the two activators: p35 or p39. Calpain-mediated cleavage of
p35 and p39 to the more stable p25 and p29, respectively, is
proposed to account for the aberrant activity of Cdk5 in AD (3, 4).
Hyperactivity of Cdk5 is involved in multiple aspects of AD pathogenesis. Cdk5 can directly phosphorylate tau, which results in
destabilization of microtubules and is therefore believed to exert
toxic effects through disruption of axonal transport in neurons
(5). Indeed, knockdown of Cdk5 expression by RNAi reduces
tau phosphorylation and decreases the number of neurofibrillary
tangles in triple transgenic (3x-Tg) AD mice (6), suggesting that
Cdk5 is the main kinase causing tau hyperphosphorylation in AD.
In addition to tau pathology, Cdk5 also mediates neuronal
death induced by Ab, since inhibition of Cdk5-p25 activity prevents Ab-induced apoptosis in cortical neurons (7). Mechanistically, Cdk5 might regulate neuronal death by activating downstream kinases such as p38 mitogen-activated protein kinase
(MAPK) and c-Jun N-terminal kinase (JNK) (8, 9). In addition,
our laboratory and others have found that Cdk5 can promote AD
pathogenesis through the regulation of gene transcription. The
transcription factor STAT3 can be phosphorylated by Cdk5 at
Ser-727 both in vitro and in vivo in the brains of rodent models,
and the transcriptional activity of STAT3 is enhanced upon this
phosphorylation event (10). This leads to increased transcription
of b-secretase 1 mRNA, which encodes the enzyme b-secretase
that participates in the amyloidogenic cleavage of amyloid precursor protein (APP) to the Ab peptide. Consistent with the notion
that Cdk5 is involved in Ab generation, enhanced Ab production
is observed in mice overexpressing p25, and the increased APP
processing can be reversed by a Cdk5 inhibitor (11). STAT3 is
also involved in Ab-induced neuronal death. Enhanced tyrosine
phosphorylation of STAT3, which indicates activation of the transcription factor, is observed in cultured cortical neurons upon Ab
treatment or in the hippocampi of AD patients. Suppression of
STAT3—either by RNAi or using a peptide that blocks STAT3
dimerization—can abolish Ab-induced caspase activity and neuronal apoptosis (12). These findings underscore that STAT3 and
Figure 1. Schematic diagram illustrating the multiple roles of Cdk5 in the pathogenesis of AD and PD. Cdk5 phosphorylates multiple targets in diverse cellular
processes, including neuronal death, microtubule disassembly, Ab production, and synaptic loss. EndoB1, endophilin B1; UVRAG, UV radiation resistance-associated
gene protein; JNK, c-Jun N-terminal kinase; STAT3, signal transducer and activator of transcription 3; BACE1, b-secretase; APP, amyloid precursor protein; AMPAR,
α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor; NR2B, N-methyl D-aspartate receptor subtype 2B; EphA4, Eph receptor A4.
the transcription of target genes, such as iNOS and TRAIL, are
crucial mediators of neuronal loss in response to Ab.
While the aforementioned studies demonstrate that Cdk5
is required for neuronal apoptosis in AD, it should be emphasized that Cdk5 also paradoxically promotes neuronal survival
during brain development. We have previously shown that the
anti-apoptotic protein Bcl-2 can be phosphorylated by Cdk5 at
Ser-70. Notably, expression of a phosphodeficient S70A Bcl-2
mutant increases nuclear fragmentation (a sign of apoptosis)
in cultured retinal neurons, but it does not further exacerbate
apoptosis in Cdk5-/- neurons, indicating that the absence of Bcl-2
phosphorylation at Ser-70 contributes to the neuronal death of
neurons lacking Cdk5 (13). Other studies also found that Cdk5 in
the nucleus promotes neuronal survival by preventing cell cycle
reentry (14, 15). It therefore appears that an intricate balance in
Cdk5 activity is critical to neuronal survival: either too much or
too little Cdk5 activity can lead to neuronal death through distinct
signaling pathways.
Role of Cdk5 in Synaptic Dysfunction
Emerging evidence indicates that loss of synapses and impaired
synaptic plasticity in the hippocampus precedes neuronal death
in AD, and the memory deficit seen is largely attributable to
the synaptic abnormalities. Natural accumulation of soluble Ab
oligomers accounts for the progressive loss of synapses of hippocampal neurons (16). Precise synaptic insertion and removal
of the N-methyl-D-aspartate (NMDA) and α-amino-3-hydroxy-5-
methyl-4-isoxazolepropionic acid (AMPA) subtypes of glutamate
receptors is essential for synaptic transmission and plasticity, and hence higher cognitive functions such as learning and
memory (17). Ab affects neurotransmission and synaptic plasticity by inducing the internalization and removal of NMDA and
AMPA receptors from synapses (18, 19). Synaptic strength can
also be modulated by regulating the number and morphology of
dendritic spines—specialized structures on the neuronal processes of postsynaptic neurons where excitatory synapses are
localized. While a typical mature spine contains a bulbous mushroom-shaped head separated from the dendritic shaft by a short
spine neck, treatment using Ab peptides leads to elimination of
dendritic spines, and the remaining spines become abnormally
long and irregular (20). It is generally believed that the reduction
of synaptic AMPA and NMDA receptors, and the loss of mature
dendritic spines, are central to the cognitive impairment in AD.
Many synaptic proteins have recently been identified to be
substrates of Cdk5 (21). Similar to the situation in neuronal survival and death discussed above, a balance in Cdk5 activity is
critical to normal synapse function. On the one hand, Cdk5 is
required for synaptic plasticity and memory formation through
phosphorylation of downstream targets such as the brain-derived neurotrophic factor receptor TrkB (22) or the regulation
of the cyclic AMP pathway (23). On the other hand, conditional
knockout mice with reduced Cdk5 expression show enhanced
memory as a result of reduced degradation of the NMDA receptor subunit GluN2 (24). Indeed, hyperactivity of Cdk5 in AD
5
NEURODEGENERATION RESEARCH
Produced by the Science/AAAS Custom Publishing Office
might not only contribute to neuronal death and toxicity through
tau hyperphosphorylation, but is also likely responsible for the
impaired synapse function in the early stages of the disease.
For example, phosphorylation of postsynaptic scaffold proteins
such as PSD-95 and GKAP by Cdk5 might account for the loss
of AMPA receptors induced by Aß (25, 26). Moreover, a previous study demonstrated that Cdk5 promotes spine elimination
through the regulation of the actin cytoskeleton. Upon activation
of the receptor tyrosine kinase EphA4 by the membrane-bound
ligand ephrinA1, Cdk5 is activated, which in turn phosphorylates and activates the guanine nucleotide exchange factor
ephexin1. This leads to activation of the small GTPase RhoA,
and triggers spine retraction (27). We also found that activation
of EphA4 leads to degradation of AMPA receptors by the proteasome (28). These studies indicate that EphA4 is an important
negative regulator of excitatory neurotransmission. It would be
interesting to further investigate whether the spine elimination
and loss of AMPA receptors mediated by EphA4 signaling might
contribute to the loss of synapses in AD.
Parkinson’s Disease
PD is a progressive neurodegenerative disorder that is associated with cognitive, emotional, and movement disorders, including resting tremor, rigidity, bradykinesia, and postural instability.
The characteristic pathology of PD includes loss of dopaminergic neurons in the substantia nigra pars compacta and the presence of intracytoplasmic inclusions known as Lewy bodies. The
etiology and the pathogenesis of PD are not completely understood but genetic factors might account for the progression of
the disease (29). In particular, two missense point mutations in
the α-synuclein protein (A53T and A30P) are linked to the earlyonset of the familial PD (30, 31).
Deregulation of autophagy is an emerging cellular mechanism that underlies the pathophysiology of PD (32). Autophagy is a homeostatic pathway involved in protein degradation and organelle turnover through the lysosomal machinery.
The autophagic pathway is responsible for the clearance of
α-synuclein, which is the major constituent of Lewy bodies.
Moreover, the mutations of α-synuclein observed in familial PD
attenuate its degradation by chaperone-mediated autophagy
(33). While these findings suggest that impaired autophagy
might contribute to the aberrant accumulation of α-synuclein
in Lewy bodies observed in PD, hyperactivation of the autophagic pathway is associated with neuronal death in PD
models (34, 35). Understanding the regulation of autophagy
in PD will therefore provide insights into potential targets for
drug screening that could ameliorate PD symptoms or disease
progression.
We recently delineated an unexpected role for Cdk5 in regulating neuronal autophagy through the protein endophilin B1
(EndoB1). A protein exhibiting lipid-binding properties, EndoB1
6
is required for autophagy induction and cell death regulation in fibroblasts (36). We found that EndoB1 is phosphorylated by Cdk5
at Thr-145. Knockdown of EnboB1 in cultured cortical neurons
by RNAi abolished starvation-induced autophagy, as indicated
by the formation of microtubule-associated protein 1 light chain
3 (LC3)-II-positive autophagosomes. Notably, the impaired autophagy could be rescued by coexpression of the wild type, but
not the phosphodeficient T145A mutant form of EndoB1 (37).
Phosphorylation by Cdk5 promotes EndoB1 dimerization, which
then triggers formation of autophagosomes through the recruitment of the UV radiation resistance-associated gene protein and
Beclin 1. Interestingly, increased Thr-145 phosphorylation of EndoB1 is detected in mice injected with 1-methyl-4-phenyl-1,2,3,6tetrahydropyridine (MPTP) and also in α-synucleinA53T mice, two
well-established animal models of PD. Furthermore, knockdown
of either EndoB1 or Cdk5, or overexpression of the T145A mutant of EndoB1, abolished the α-synucleinA53T mutant-induced
autophagy and neuronal death (37). Our findings therefore establish the involvement of Cdk5-mediated phosphorylation of EndoB1 and autophagy deregulation in the pathophysiology of PD.
In conclusion, we and others have demonstrated a complex
but essential role for Cdk5 in the pathogenesis of AD and PD.
Aberrant Cdk5 activity contributes to different aspects of disease
progression, including neuronal death and synaptic dysfunction,
through phosphorylation of distinct substrates. The identification of small molecule inhibitors of Cdk5 would be a promising
strategy for drug discovery targeting these neurodegenerative
diseases.
REFERENCES
1. D. M. Walsh, D. J. Selkoe, Neuron 44, 181 (2004).
2. Z. H. Cheung, N. Y. Ip, Trends Cell Biol. 22, 169 (2012).
3. G. N. Patrick et al., Nature 402, 615 (1999).
4. H. Patzke, L. H. Tsai, J. Biol. Chem. 277, 8054 (2002).
5. D. B. Evans et al., J. Biol. Chem. 275, 24977 (2000).
6. D. Piedrahita et al., J. Neurosci. 30, 13966 (2010).
7. Y. L. Zheng et al., J. Biol. Chem. 285, 34202 (2010).
8. K. H. Sun, H. G. Lee, M. A. Smith, K. Shah, Mol. Biol. Cell 20, 4611 (2009).
9. K. H. Chang et al., J. Neurochem. 113, 1221 (2010).
10. A. K. Y. Fu et al., Proc. Natl. Acad. Sci. U.S.A. 101, 6728 (2004).
11. Y. Wen et al., Neuron 57, 680 (2008).
12. J. Wan et al., J. Neurosci. 30, 6873 (2010).
13. Z. H. Cheung, K. Gong, N. Y. Ip, J. Neurosci. 28, 4872 (2008).
14. J. Zhang, K. Herrup, Cell Cycle 7, 3487 (2008).
20. P. N. Lacor, J. Neurosci. 24, 796 (2007).
34. E. H. Baehrecke, Nat. Rev. Mol. Cell Biol. 6, 505 (2005).
21. K. O. Lai, N. Y. Ip, Biochim. Biophys. Acta 1792, 741 (2009).
35. M. C. Maiuri, E. Zalckvar, A. Kimchi, G. Kroemer, Nat. Rev. Mol. 22. K. O. Lai et al., Nat. Neurosci. 15, 1506 (2012).
23. J. S. Guan et al., PLOS ONE 6, e25735 (2011).
36. Y. Takahashi et al., Nat. Cell Biol. 9, 1142 (2007).
24. A. H. Hawasli et al., Nat. Neurosci. 10, 880 (2007).
37. A. S. Wong et al., Nat. Cell Biol. 13, 568 (2011).
17. J. D. Shepherd, R. L. Huganir, Annu. Rev. Cell Dev. Biol. 23, 613 (2007).
18. H. Hsieh et al., Neuron 52, 831 (2006).
Cell Biol. 8, 741 (2007).
25. F. Roselli, P. Livrea, O. F. Almeida, PLOS ONE 6, e23097 (2011).
26. F. Roselli et al., J. Neurosci. 25, 11061 (2005).
Acknowledgments: We thank all members of the Ip laboratory and are
27. W. Y. Fu et al., Nat. Neurosci. 10, 67 (2007).
grateful to Ka-Chun Lok for preparing the figure in this review. The
28. A. K. Fu et al., Nat. Neurosci. 14, 181 (2011).
studies were supported in part by the Research Grants Council of Hong
29. E. Broussolle, S. Thobois, Rev. Neurol. 158, 11 (2002).
Kong (661109, 661309, 660810, 661010, 661111, and 661212), and the
30. R. Kruger et al., Nat. Genet. 18, 106 (1998).
Theme-based Research Scheme (T13-607/12R), as well as the National
31. M. H. Polymeropoulos et al., Science 276, 2045 (1997).
Key Basic Research Program of China (2013CB530900), the Innovation
32. T. Pan et al., Neurobiol. Dis. 32, 16 (2008).
and Technology Fund for State Key Laboratory (ITCPT/17-9), the
33. A. M. Cuervo, L. Stefanis, R. Fredenburg, P. T. Lansbury, D. Sulzer, Shenzhen Peacock Plan, and the SH Ho Foundation.
The Role of Autophagy, Apoptosis, Neuroinflammation,
and Mitochondrial Dysfunction in Parkinson’s Disease
Dong Chen, Haigang Ren, Qingsong Hu, Feng Gao, and Guanghui Wang*
Parkinson’s disease (PD) is the second most common neurodegenerative disease, characterized by the loss of dopaminegic
(DA) neurons in the substantia nigra pars compacta (SNpc) and
the presence of Lewy bodies in the surviving neurons, accompanied by the formation of dystrophic neurites (1–2). Although
the etiology of PD remains unknown, it is apparent that the interaction between environmental and genetic factors is critical.
Mitochondrial dysfunction plays an important role in PD pathogenesis; damaged mitochondria induce many cellular and pathological processes in the diseased brain, including the induction of
apoptosis and the production of reactive oxygen species (ROS).
ROS formation may result in oxidative stress and downstream
signaling that promotes neuroinflammation.
Mitophagy is a cellular process that clears damaged mitochondria. A number of PD-related gene products appear to be tightly
associated mitophagy, including PINK1 and parkin (3). Mutations
in these genes can result in a failure to clear damaged mitochon-
15. J. Zhang et al., J. Neurosci. 30, 5219 (2010).
16. G. A. Krafft, W. L. Klein, Neuropharmacology 59, 230 (2010).
Science 305, 1292 (2004).
19. P. Kurup et al., J. Neurosci. 30, 5948 (2010).
Laboratory of Molecular Neuropathology, Department of
Pharmacology, Soochow University College of Pharmaceutical
Sciences, Suzhou, Jiangsu, China
*
Corresponding Author: [email protected]
dria, leading to an accumulation of dysfunctional mitochondria
(4–5). Many factors involved in PD, including the genetic factors—parkin, DJ-1, Omi—and environmental factors—rotenone
and 1-methyl-4-phenyl-1,2,3,4-tetrahydropyridine (MPTP)—play
roles in the regulation of mitochondria. Loss of protective function or gain of toxic function in these factors is thought to induce
oxidative stress, apoptosis, and neuroinflammation in association with PD (6–9). This review summarizes and provides context
for these findings.
Autophagy and PD
There is growing evidence showing that autophagy is associated
with the pathogenesis of PD (10–12). The postmortem brains
of PD patients present with autophagic vacuoles in the SNpc,
indicating an involvement of autophagy (13). Additionally, the
loss of autophagy-related genes appears to result in neurodegeneration. For example, a conditional deletion of Atg7 in the
central nervous system or SNpc selectively impairs autophagy
and leads to a progressive loss of DA neurons, indicating the
protective role of basal levels of autophagy against neurodegeneration (14–15). Interestingly, many PD-related proteins such as
LRRK2, α-synuclein, DJ-1, PINK1, parkin, and Omi are involved
in the autophagy process (16–21). The protease Omi activates
7
NEURODEGENERATION RESEARCH
Produced by the Science/AAAS Custom Publishing Office
autophagy through its cleavage of Hax-1 as part of the Beclin
1-dependent autophagic pathway. In vitro data showed that
Omi-induced autophagy promotes the degradation of neurodegenerative proteins such as pathogenic A53T α-synuclein and
truncated polyglutamine (polyQ)-expanded huntingtin, whereas
knockdown of Omi decreases the basal level of autophagy and
increases the level of A53T α-synuclein and polyQ-expanded
huntingtin. Further, S276C Omi, a protease-defective mutant
present in the brains of mnd2 (motor neuron degeneration 2)
mice, is unable to regulate autophagy. These results indicate
that Omi is a regulator of autophagy and provide evidence that
Omi-associated autophagy may provide an essential means for
protein quality control in neurodegenerative diseases (21).
In addition, two other PD-related proteins, DJ-1 and parkin,
also regulate autophagy. Knockdown of DJ-1 induces autophagy
through activation of JNK and upregulation of Beclin 1 transcription. Inhibition of the JNK pathway was found to block autophagy
activation and p62 degradation induced by DJ-1 knockdown,
suggesting that autophagy regulation by DJ-1 is JNK-dependent
(18).
Parkin is an E3 ubiquitin ligase that has been shown to regulate mitophagy. Interestingly, in the cytosol, parkin represses
autophagy. Parkin, but not its E3 ligase-deficient mutants, monoubiquitinates Bcl-2, leading to a stabilization of this protein. The
resulting effective increase in Bcl-2 levels enhances the interaction between Bcl-2 and Beclin 1, thereby repressing autophagy.
In parkin-overexpressed cells, LC3 conversion is decreased,
whereas it is increased in parkin-knockdown cells, further suggesting that parkin represses autophagy. In contrast to cytosolic
parkin, mitochondrial parkin, after its recruitment to mitochondria
upon carbonyl cyanide m-chlorophenylhydrazone treatment, increases mitophagy, suggesting a differential role of cytosolic and
mitochondrial parkin in autophagy (19).
Mitochondria and Apoptosis
Recent studies have shown that the selective degradation of
damaged mitochondria by autophagy through the PINK1/parkin
pathway is an important quality control mechanism in PD pathogenesis (4). Loss of function of PINK1 or parkin causes an accumulation of abnormal mitochondria, which leads to increased
oxidative stress. Mitochondria have an integral role in the apoptotic cell death pathway by releasing Bax and cytochrome c into
the cytosol to induce apoptosis. Oxidative stress can lead to the
collapse of mitochondrial membrane potential, which may induce
a translocation of Bax to activate apoptotic pathways (22). Our
previous studies showed that DJ-1 has anti-apoptotic effects
that act through a mitochondria-associated pathway. DJ-1 exerts its cytoprotection through inhibition of the p53/Bax/caspase
pathway: it interacts with and represses p53 in the nucleus to
downregulate the expression of Bax and thereby inhibit caspase
activation. However, a K130R mutant of DJ-1 has been shown
8
to lose its inhibitory effects on p53 due to failed translocation of
the mutant protein to the nucleus (23–24). DJ-1 also functions
in the TNF-related apoptosis-inducing ligand (TRAIL) pathway
by blocking death-inducing signaling complex (DISC) formation
and inhibiting procaspase-8 activation (25). DJ-1 inhibits caspase-8 activation by its interaction with Fas-associated protein
death domain (FADD) to inhibit the formation of DISC. DJ-1, but
not its pathogenic mutant L166P, inhibits procaspase-8 activation by competing with it to bind to the death effector domain
of FADD. Decreased binding of procaspase-8 to DISC results
in lower activation of caspase-8 and the repression of apoptosis (25). Apart from the anti-apoptotic effect of DJ-1 exerted at
the cell membrane, it also has an effect on mitochondria. DJ-1
translocates to the mitochondria and binds to Bcl-XL following
oxidative stress. The binding of DJ-1 stabilizes Bcl-XL by inhibiting its ubiquitination and degradation through the ubiquitin-proteasome system. Knockdown of DJ-1 decreases Bcl-XL levels,
subsequently leading to mitochondrial Bax enrichment, caspase-3 activation, and cell death in response to oxidative stress
(26). Interestingly, the pathogenic DJ-1 mutant L166P shows
increased mitochondrial distribution and affinity for Bcl-XL. So,
unlike wild-type DJ-1, DJ-1(L166P) does not influence Bcl-XL
levels, but rather disrupts mitochondrial Bcl-XL/Bax heterodimerization, thereby releasing Bax from Bcl-XL/Bax heterodimers.
The released Bax oligomerizes in the outer mitochondrial membrane and induces cell death under oxidative stress. Thus, both
DJ-1 and its L166P mutant have direct effects on mitochondrial
Bcl-XL, but apparently induce mitochondria-related cell death in
different ways.
vironmental toxin activates microglia through the p38 MAPK
pathway (32). Most research has shown that those PD-related
genetic factors that have been studied cause DA neuronal cell
death directly or result in DA neurons sensitive to environmental
stimuli. It is well accepted that damage to DA neurons induces
neuroinflammation (30), but little is known about whether any
gene products related to PD pathogeneses are able to activate
microglia. Recently, we found that in mnd2 mice, which harbor
a protease-deficient Omi (S276C Omi) and show neurodegeneration, the activation of microglia may be a direct result of the
protease deficient Omi. Omi cleaves MEK1, which is upstream
of the extracellular signal-regulated kinase 1 and 2 (ERK1/2),
thereby repressing ERK1/2 activation. Loss of Omi protease
activity in mnd2 mice and knockdown of Omi in BV2 cells activates NF-κB and results in the production of inflammatory factors, including tumor necrosis factor-α and inducible nitric oxide
synthase. By contrast, inhibition of MEK1 blocks the loss of Omi
activity-induced NF-κB activation. These data suggest that the
PD genetic factor Omi is able to activate microglia directly (9),
hinting at a fundamental link between neuroinflammation, ROS,
and mitochondrial function in both environmentally and genetically induced PD.
In summary, mitochondrial dysfunction caused by genetic and
environmental factors—either by loss of protection or toxic insult—plays a crucial role in PD pathogenesis. Impairment of mitochondrial function triggers multiple cellular processes, including impaired autophagy, increased sensitivity to oxidative stress,
and induction of apoptosis associated with both neurodegeneration and neuroinflammation.
9. Q. Hu et al., Sci. Signal. 5, ra61 (2012).
10. B. Levine, G. Kroemer, Cell 132, 27 (2008).
11. Z. H. Cheung, N. Y. Ip, Mol. Brain 2, 29 (2009).
12. P. A. Jaeger, T. Wyss-Coray, Mol. Neurodegener. 4, 16 (2009).
13. P. Anglade et al., Histol. Histopathol. 12, 25 (1997).
14. M. Komatsu et al., Nature 441, 880 (2006).
15. I. Ahmed et al., J. Neurosci. 32, 16503 (2012).
16. E. D. Plowey, S. J. Cherra, 3rd, Y. J. Liu, C. T. Chu, J. Neurochem. REFERENCES
1 J. M. Savitt, V. L. Dawson, T. M. Dawson, J. Clin. Invest. 116, 1744 (2006).
105, 1048 (2008).
17. T. Vogiatzi, M. Xilouri, K. Vekrellis, L. Stefanis, J. Biol. Chem. 283, 23542 (2008).
18. H. Ren et al., Cancer Lett. 297, 101 (2010).
19. D. Chen et al., J. Biol. Chem. 285, 38214 (2010).
17. T. Vogiatzi, M. Xilouri, K. Vekrellis, L. Stefanis, J. Biol. Chem. 283, 23542 (2008).
18. H. Ren et al., Cancer Lett. 297, 101 (2010).
19. D. Chen et al., J. Biol. Chem. 285, 38214 (2010).
20. S. Michiorri et al., Cell Death Differ. 17, 962 (2010).
21. B. Li et al., Cell Death Differ. 17, 1773 (2010).
22. C. Henchcliffe, M. F. Beal, Nat. Clin. Pract. Neurol. 4, 600 (2008).
23. J. Fan et al., J. Biol. Chem. 283, 4022 (2008).
24. J. Fan et al., FEBS Lett. 582, 1151 (2008).
25. K. Fu et al., Oncogene 31, 1311 (2012).
26. H. Ren, K. Fu, D. Wang, C. Mu, G. Wang, J. Biol. Chem. 286, 35308 (2011).
27. H. M. Gao et al., J. Neurosci. 28 7687 (2008).
28. P. L. McGeer, S. Itagaki, B. E. Boyes, E. G. McGeer, Neurology 38, ROS, Neuroinflammation, and PD
Neuroinflammation presents in both animal models and the postmortem brains of PD patients (6, 27, 28). It has been proposed
that mitochondrial dysfunction may exert a crucial role in neuroinflammation and pathogenesis of PD (29). Injury of mitochondria induces the production of ROS in both neuronal and glial
cells, and causes DA neuronal cell death. Microglia, the main
cell type of the innate immune system in the brain, can be activated to produce inflammatory factors, promoting DA neuronal
cell death (30, 31). Rotenone, a mitochondrial complex I inhibitor,
is well known as an environmental factor that can induce ROS
production and oxidative damage in DA neurons. We found recently that rotenone induces ROS production in microglia. This
does not cause microglial cell death, but activates the nuclear
factor kappa B (NF-κB) signaling pathway, thereby inducing a
significantly increased expression of inflammatory cytokines, activation of caspase-1, and maturation of IL-1β (32). Treatment
of BV2 cells (a microglial cell line) with rotenone activates p38
mitogen-activated protein kinase (MAPK), whereas removal of
ROS with N-acetylcysteine (NAC), a ROS scavenger, reduces
p38 and NF-κB activation in BV2 cells, suggesting that this en-
(2008).
8. J. Waak et al., FASEB J. 23, 2478 (2009).
1285 (1988).
29. M. E. Witte, J. J. Geurts, H. E. de Vries, P. van der Valk, J. van Horssen, Mitochondrion 10, 411 (2010).
30. M. L. Block, J. S. Hong, Biochem. Soc. Trans. 35, 1127 (2007).
2. C. Liu et al., J. Biol. Chem. 282, 14558 (2007).
31. W. G. Kim et al., J. Neurosci. 20, 6309 (2000).
3. S. M. Jin, R. J. Youle, J. Cell Sci. 125, 795 (2012).
32. F. Gao, D. Chen, Q. Hu, G. Wang, PLOS ONE 8, e72046 (2013).
4. D. Narendra, A. Tanaka, D. F. Suen, R. J. Youle, J. Cell Biol. 183, 795 (2008).
Acknowledgments: This work was supported in part by the National
5. N. Matsuda et al., J. Cell Biol. 189, 211 (2010).
High-Tech Research and Development Program of China (“973” Program,
6. T. C. Frank-Cannon et al., J. Neurosci. 28, 10825 (2008).
2011CB504102), and the National Natural Sciences Foundation of China
7. C. Zhou, Y. Huang, S. Przedborski, Ann. N.Y. Acad. Sci. 1147, 93 (911327723).
9
NEURODEGENERATION RESEARCH
Produced by the Science/AAAS Custom Publishing Office
The Multiple Roles of Autophagy in Neural Function and Disease
Rui Sheng and Zheng-Hong Qin*
Autophagy is an evolutionarily conserved pathway that involves
the sequestration and delivery of cytoplasmic materials to lysosomes, where proteins, lipids, and organelles are degraded and
recycled. Generally, autophagy is divided into three categories:
macroautophagy, microautophagy, and chaperone-mediated autophagy (CMA). Autophagy is also involved in many fundamental
cellular activities and impacts numerous cellular regulatory pathways involved in diverse processes, including tumorigenesis,
longevity, immunity, organelle turnover, ER stress, and apoptosis. Important progress has been made in defining the role of
autophagy in human disease, enabling the development of better
treatments for cancer, neurological diseases, and cardiovascular
diseases.
Starting four decades ago, autophagy research expanded from
a relatively minor area to one of the most exciting and important topics in cell biology. The total number of autophagy papers
published in peer review journals worldwide has increased 15
fold in past 10 years (1). A number of autophagy-related genes
and pathways have been identified and innovative molecular
tools developed for genetic manipulation and chemical targeting
of pathway components (2). In recent decades, autophagy has
evolved into an active research field in China. Chinese scientists
have contributed a great deal to basic and clinical autophagy
research, their journal papers accounting for 10% of worldwide
publications on this topic.
In the Laboratory of Aging and Nervous Diseases at Soochow
University, one of the oldest domestic institutions carrying out
autophagy research, we have been working on the role of autophagy in the regulation of cell metabolism and cell survival in
the central nervous system (CNS), including neurodegenerative
diseases, neuronal excitotoxicity, cerebral ischemia, and ischemic preconditioning. In particular, our researches focus on the
degradation of pathogenic proteins, excitotoxicity, preconditioning, and the crosstalk between autophagy and apoptosis. This
work has elucidated some of the pathogenic mechanisms of nervous system diseases and suggested novel targets for therapies.
Below is a summary of some of the research recently performed
in our laboratory.
Department of Pharmacology and Laboratory of Aging
and Nervous Diseases, Soochow University School of
Pharmaceutical Science, Suzhou, China
*
Corresponding Author: [email protected]
10
Autophagy in Metabolism of Misfolded Proteins
Accumulation and aggregation of misfolded proteins is the hallmark of several neurodegenerative diseases. Examples include
the huntingtin (Htt) protein in Huntington’s disease (HD) and
α-synuclein in Parkinson’s disease (PD).
HD is caused by an expansion of the polyglutamine (polyQ)
tract in the HTT protein. This induces selective degeneration of
striatal projection neurons and cortical pyramidal neurons due to
the accumulation of N-terminal mutant HTT and intranulear inclusion formation in HD brains. Our studies suggest that autophagy
and lysosomal cathepsins play critical roles in the degradation of
N-terminal HTT. Altered processing of mutant HTT by autophagy
and cathepsin D contributes to HD pathogenesis (3–5). In macroautophagy, Beclin 1 has been shown to be involved in the degradation of HTT (6). Additionally, we have shown that HTT fragments can be degraded by CMA. The CMA components—heat
shock protein cognate 70 (Hsc70) and lysosome-associated
protein 2A (LAMP-2A)—play essential roles in the clearance of
HTT (7).
The pathological hallmark of PD is neuronal inclusions termed
Lewy bodies that are composed mainly of the misfolded protein
α-synuclein. Our studies have shown that autophagy plays a role
in the degradation of α-synuclein (8). Lysosomal enzymes cathepsin L and B are involved in the regulation of autophagy and
degradation of α-synuclein (9). Abnormal distribution of cathepsin L is thought to contribute to neuronal death; thus drugs inhibiting cathepsin L nuclear translocation might be neuroprotective
against PD (10, 11). Recent unpublished work has shown that
impaired autophagy flux is involved in PD, resulting in the accumulation and aggregation of α-synuclein protein and damage of
lysosomal function in dopamine neurons.
Autophagy in Neuronal Excitotoxicity
Excitotoxicity is thought to play an important role in the pathogenesis of a number of neurological disorders, including HD, PD,
and Alzheimer’s disease (AD) (12). In the late 1990s, we and
other laboratories provided morphological and biochemical evidence demonstrating that an apoptotic mechanism was involved
in excitotoxic neuronal death (13–15). In 2008, we reported that
excitatory amino acid receptor agonists activate autophagy in
animal models (16). Both kainic acid (KA) receptor agonist KA
and N-methyl-D-aspartate (NMDA) receptor agonist quinolinic
acid (QA) induced activation of autophagy, accompanied by
downregulation of Bcl-2, and upregulation of Bax, tumor sup-
pressor protein p53 (TP53), and TP53-upregulated modulator of
apoptosis (PUMA). Autophagy inhibitors and cathepsin inhibitors
markedly inhibited autophagy activation and the mitochondriamediated apoptotic pathway, suggesting that the autophagy-lysosmal pathway plays an important role in excitotoxic neuronal injury (17, 18). TP53 might be the key factor mediating autophagy
activation and mitochondrial dysfunction in neuronal excitotoxicity. TP53 inhibitor pifithrin-α not only inhibited excitotoxic neuronal injury, but also suppressed KA- or QA-induced autophagy
activation and mitochondrial dysfunction (17, 18). These studies
have advanced the understanding of the molecular mechanisms
of excitotoxicity, which involve necrosis, apoptosis, and autophagic cell death (19).
Different Roles of Autophagy in Cerebral
Ischemia and Ischemic Preconditioning
Stroke is a global health problem and one of the leading causes
of adult disability in developed countries, while the phenomenon
of ischemic preconditioning (IPC) has been recognized as one of
the most potent mechanisms to protect against ischemic stroke
(20). Morphological and biochemical evidence obtained from a
rodent permanent focal ischemia model suggests that ischemic
stroke induced autophagy activation. Using pharmacological approaches, we demonstrated that the autophagy-lysosomal pathway is involved in the neuronal death induced by permanent focal ischemia in adult rodents (21). Activation of autophagy and
lysosomal proteases also occurred in astrocytes and contributed
to the ischemia-induced glial death in both in vivo and in vitro
ischemia models (22). However, we found that IPC also activated autophagy and a blockade of autophagy activation during IPC
abolished the neuroprotective effects of IPC (23).
The studies above clearly demonstrated that autophagy could
play different roles under different pathological conditions. In an
effort to better elucidate the mechanisms involved, we showed
that both IPC and lethal cerebral ischemia induced activation of
autophagy, but the extent and persistence of this activation were
variable. The condition of permanent cerebral ischemia briefly
activated autophagy, while IPC itself mildly triggered autophagy
and this activation persisted longer during a second ischemic attack (24). The divergent effects of autophagy during ischemia
and preconditioning were also found to be related to endoplasmic reticulum (ER) stress (24, 25), which induced ER-associated
degradation (ERAD), a process that involves both the proteasomal and autophagic pathways. We found that pre-activation
of autophagy by IPC can upregulate molecular chaperones and
hence reduce excessive ER-stress–dependent apoptosis during
fatal ischemia. Our current studies have focused on the contribution of ER chaperone GRP78 to IPC-induced autophagy and
have found indications that GRP78 is a key mediator of autophagy activation during preconditioning (unpublished observations).
This work also suggests that induction of autophagy and ER
stress may provide a strong justification for the introduction of
IPC to treat cerebral ischemia (23).
Crosstalk Between Autophagy and Apoptosis
At the cellular level, the involvement of autophagy in the cell
death and cell survival processes appears to be complex. In QAand KA-induced apoptosis, the downregulation of pro-survival
protein Bcl-2, was partially blocked by the autophagy inhibitor
3-methyladenine (3-MA) (16, 17). Our studies also revealed
that mitochondrial inhibitor 3-nitropropionic acid (3-NP)-induced
neuronal death involved both apoptotic and autophagic mechanisms. TP53 was found to induce a number of downstream
genes, including Bax, PUMA, and damage-regulated autophagy
modulator 1 (DRAM1). Importantly, neuronal injury induced by
3-NP was inhibited by the TP53 specific inhibitor pifithrin-α, the
autophagy inhibitor 3-MA, and by knockdown of DRAM1 (26,
27). DRAM1 apparently affects autophagy through augmentation
of lysosomal acidification, fusion of lysosomes with autophagosomes, and clearance of autophagosomes (27). Also, 3-NP–induced autophagy seemed to contribute to the degradation of Bcl2 (28). We thus investigated the crosstalk between autophagy
and apoptosis. It is well defined that autophagy activation under
nutritional stress conditions benefits cell survival (29). We found
that autophagy activation induced by serum deprivation was accompanied by an upregulation of Bcl-2. Inhibition of Bcl-2 function or downregulation of Bcl-2 expression amplified autophagy
activation and resulted in apoptosis, while overexpression of
Bcl-2 limited autophagy activation and blocked cell death in response to serum deprivation (30). Bcl-2 thus plays an important
role in limiting autophagy activation and preventing initiation of
programmed cell death during serum starvation. These studies
have added significant information on a classical role of autophagy in nutritional stress.
Recently, we defined a novel pathway for crosstalk between
autophagy and apoptosis involving Bax and DRAM1. We found
that pro-apoptotic protein Bax was degraded through autophagy
under basal condition, keeping its levels low. However, DRAM1mediated activation of autophagy resulted in upregulation of Bax
due to protein-protein interaction between DRAM1 and Bax. Increased Bax was recruited to lysosomes by DRAM1, and Bax
lysosomal translocation permeabilized lysosomal membranes.
Apoptosis was then initiated by lysosomal protease cathepsin
B-mediated Bid cleavage, cytochrome c release, and caspase 3
activation (unpublished observations).
Autophagy Under Other Physiological and
Pathological Conditions
It is well known that regular endurance exercise is required
to maintain muscle mass and body fitness in mammals. Our
studies showed that long-term regular exercise increased
autophagy activity in both adult and aged animals (31).
11
NEURODEGENERATION RESEARCH
Produced by the Science/AAAS Custom Publishing Office
Exercise training increased autophagy activity, but reduced
apoptosis of muscle cells by modulating IGF-1 and its receptors,
the Akt/mTOR and Akt/FOXO3a signaling pathways in skeletal
muscles. Autophagy activation appeared to inhibit chloroquineinduced muscle damage through removal of abnormal mitochondria; it thus plays a positive role in muscle establishment and
fitness (unpublished data).
Traumatic brain injury (TBI) also activates autophagy and lysosomal proteases in neurons. Inhibition of this autophagy-lysosome axis may help attenuate traumatic damage and improve
functional recovery (32 ).
In summary, our research demonstrates that autophagy is coordinated with other cellular activities such as apoptosis, ER stress,
and mitochondrial function to maintain cell homeostasis. Autophagy plays different roles in neurodegenerative diseases, neuronal
excitotoxicity, cerebral ischemia, and ischemic preconditioning.
The multifunctional roles of autophagy are explained by its ability
to interact with distinct key components in various signaling pathways involving TP53, DRAM1, Bcl-2, and GRP78. We expect that
a better understanding of the roles of autophagy and its regulation
will bring deeper insight into pathogenic mechanisms of CNS diseases and provide useful information leading to the development
of novel therapies.
8. F. Yang et al., Neurosci. Lett. 454, 203 (2009).
9. B. Xiang et al., Brain Res. 1387, 29 (2011).
10. L. Li et al., Neurosci. Lett. 489, 62 (2011).
REFERENCES
12. X. X. Dong, Y. Wang, Z. H. Qin, Acta Pharmacol. Sin. 30, 379 (2009).
13. Z. H. Qin, Y. Wang, M. Nakai, T. N. Chase, Mol. Pharmacol. 53, 33
(1998).
14. Z. H. Qin, Y. Wang, T. N. Chase, Brain Res. 725, 166 (1996).
15. Z. H. Qin et al., J. Neurosci. 19, 4023 (1999).
16. Y. Wang et al., Autophagy 4, 214 (2008).
17. Y. Wang et al., Eur. J. Neurosci. 30, 2258 (2009).
18. X. X. Dong et al., Neuroscience 207, 52 (2012).
19. Y. Wang, Z. H. Qin, Apoptosis 15, 1382 (2010).
20. X. Q. Liu, R. Sheng, Z. H. Qin, Acta Pharmacol. Sin. 30, 1071 (2009).
21. Y. D. Wen et al., Autophagy 4, 762 (2008).
22. A. P. Qin et al., Autophagy 6, 738 (2010).
23. R. Sheng et al., Autophagy 6, 482 (2010).
24. R. Sheng et al., Autophagy 8, 310 (2012).
25. B. Gao et al., Acta Pharmacol. Sin. 34, 657 (2013).
26. X. D. Zhang et al., Autophagy 5, 339 (2009).
27. X. D. Zhang, L. Qi, J. C. Wu, Z. H. Qin, PLOS ONE 8, e63245 (2013).
28. X. D. Zhang et al., J. Neurosci. Res. 87, 3600 (2009).
1. W. D. Le, Z. H. Qin, Acta Pharmacol. Sin. 34, 583 (2013).
29. J. D. Rabinowitz, E. White, Science 330, 1344 (2010).
2. Y. P. Yang, Z. Q. Liang, Z. L. Gu, Z. H. Qin, Acta Pharmacol. Sin. 26, 30. H. D. Xu et al., PLOS ONE 8, e63232 (2013).
1421 (2005).
3. Z. H. Qin et al., Hum. Mol. Genet. 12, 3231 (2003).
31. L. Luo et al., Exp. Gerontol. 48, 427 (2013).
32. C. L. Luo et al., Neuroscience 184, 54 (2011).
4. Z. H. Qin et al., J. Neurosci. 24, 269 (2004).
12
Ya Ke1*, Ka-Chun Wu1, and Zhong-Ming Qian2*
11. X. F. Fei et al., Brain Res. 1264, 85 (2009).
Iron and Neurodegeneration
5. L. L. Chen et al., Acta Pharmacol. Sin. 33, 385 (2012).
Acknowledgments: This work was supported by grants from the Natural
6. J. C. Wu et al., Acta Pharmacol. Sin. 33, 743 (2012).
Science Foundation of China (30930035 and 81271459) and the Ministry
7. L. Qi et al., PLOS ONE 7, e46834 (2012).
of Science and Technology (“973”) Program (2011CB510003).
Iron is an essential element in many
physiological processes in the nervous system, including DNA synthesis, oxygen transport, mitochondrial
respiration, neurotransmitter synthesis and myelination (1). Brain iron levels increase with age, especially in
those regions associated with motor
functions, such as the globus pallidus
and substantia nigra. Age-related iron
accumulation in regions associated
with cognitive function, such as frontal
grey matter, has also been reported
(2). Recently, abnormal brain iron accumulation was reported in a wide
spectrum of neurodegenerative disorders, including Alzheimer’s disease
(AD) and Parkinson’s disease (PD),
suggesting a role for iron accumulation in neurodegeneration (1). In this
review, we discuss iron metabolism
in the brain and the toxic effects of
Figure 1. Brain iron metabolism and neurodegeneration. DMT1 and Lf/TfR, responsible for iron uptake, may
be upregulated (in Parkinson’s disease), while APP-MTP1 coupling underlying iron release may be reduced (in Aliron overloading in neurodegenerazheimer’s disease), facilitating iron accumulation. The resulting iron-induced increased Aβ and α-synuclein synthesis
tive disorders as well as briefly intromay facilitate pathological aggregation. Moreover, abnormally high iron levels can promote free radical production
via the Fenton reaction, leading to widespread oxidative damage. These processes likely contribute to the initiaduce some treatment strategies and
tion of neurodegeneration. Aβ, β-amyloid; MTP1, metal transporter protein-1 (ferroportin); APP, amyloid precursor
provide insights based on our recent
protein; Hp, hephaestin; CP, ceruloplasmin; ROS, reactive oxygen species; DMT1, divalent metal transporter 1; Lf,
lactoferrin; LfR, lactoferrin receptor; Tf, transferrin; TfR1, transferrin receptor 1.
findings.
cells act cooperatively to maintain appropriate iron levels to meet
Brain Iron Transport and Metabolism
The brain needs a stable environment in order to function the need of both neurons and the brain as a whole. Astrocytes
normally. Both a deficiency and an excess of iron can cause act as a bridge between neurons and BBB, and are responsible
severe damage, so it is important that iron levels in the brain are for regulating iron transport within the brain. Microglia serve as a
tightly regulated. The blood-brain barrier (BBB) forms a frontline brain iron capacitor by quickly accumulating and releasing iron,
boundary and maintains appropriate iron levels by regulating iron and thus contain variable iron content depending on conditions
influx and efflux (3). Additionally, iron exchange can take place (4, 5).
Similar to cells outside the brain, brain cells such as neurons,
in the blood-cerebrospinal fluid barrier that separates circulating
blood from the brain’s extracellular fluid (3). Inside the brain, glial microglia, and astrocytes express numerous iron-related proteins (2). We have previously described in detail the metabolic
pathways for cellular iron in the brain (3, 6). Briefly, transferrin
(Tf), the transferrin receptor (TfR), and the divalent metal trans1
School of Biomedical Sciences, Faculty of Medicine,
The Chinese University of Hong Kong, Hong Kong
porter 1 (DMT1) are involved in iron uptake. Tf carrying ferric iron
2
Laboratory of Neuropharmacology, Fudan University School
(Fe3+) binds to TfR on the cell membrane, inducing internalizaof Pharmacy, Shanghai, China
tion of the whole complex. With the help of endosomal metallore*
Corresponding Authors: [email protected] (Y. K.) and
ductase STEAP3, Fe3+ is converted to ferrous iron (Fe2+), which
[email protected] (Z. M. Q.)
13
NEURODEGENERATION RESEARCH
Produced by the Science/AAAS Custom Publishing Office
can then pass through the endosomal membrane and enter the
cytosol with the aid of DMT1. Alternatively, extracellular Fe2+ that
is not bound to transferrin may pass through the cell membrane
directly, via DMT1. After entering the cytosol, Fe2+ ions bind to
chaperone proteins that transfer it to target proteins, or it enters
the mitochondria, an iron-laden organelle, via mitoferrin 1 and 2.
Ferritin, a large, iron storage protein, is composed of two subunits: ferritin heavy chain (Ft-H)—a ferroxidase—and ferritin light
chain (Ft-L). Excess intracellular Fe2+ can be converted to Fe3+
by Ft-H and sequestered by Ft-L for storage. Iron regulatory protein 1 and 2 are responsible for regulating expression of proteins
that are sensitive to iron content, i.e., those that carry an ironresponsive element in their mRNA transcript. Finally, ferroportin
(or metal transporter protein-1, MTP1) functions to release cellular ferrous iron, which is then converted to Fe3+ by ceruloplasmin
(CP) or hephaestin (Hp). In the Fe3+ form, iron can then bind
to transferrin again for recirculation. However, the expression of
iron-related proteins is not ubiquitous in the brain (2), with certain iron-related proteins being differentially expressed, revealing
specialization of iron metabolism within different cells. For example, transferrin is most highly expressed in oligodendrocytes
while astrocytes specialize in synthesizing CP (1, 2), suggesting
that different cell types in the brain work cooperatively to maintain iron homeostasis.
Iron Accumulation in the Brain
Due to the pivotal role of iron, the brain has developed iron conservation mechanisms, as evidenced by the minimal reduction of
iron in the brain during systemic iron deficiency (7). Interestingly,
iron is found to accumulate with age (1), a process that some
suggest is passive and caused by a reduction in the efficiency of
the BBB and cells that control iron levels, while others argue that
a more active process is at work. A number of iron-related genes
have been found to be associated with neurodegenerative disorders including AD and PD (8), suggesting that abnormal iron accumulation may be a cause of accelerated neurodegeneration.
Iron Toxicity and Neurodegeneration
The toxicity of iron can be explained by the fact that it is highly
redoxactive (9). Under normal circumstances, most cellular iron
is sequestered by proteins. However, when the level of iron exceeds the capacity of iron-binding proteins, labile iron results.
Hydrogen peroxide, which is constantly generated by mitochondrial electron transport, can be converted to extremely toxic hydroxyl free radicals in the presence of labile iron via the Fenton
and Haber-Weiss reactions (9). These free radicals may elicit
an array of cellular damage, including protein carbonylation and
lipid peroxidation, and eventually evoke neuronal death. Since
neurons normally have a high metabolic demand, they tend to be
more susceptible to iron-induced oxidative damage. Moreover,
iron has also been found to facilitate abnormal protein aggrega-
14
tion, which contributes to the pathogenesis of many neurodegenerative disorders (10–12). However, the molecular mechanism
of iron accumulation in neurodegenerative diseases such as AD
and PD remains elusive.
Alzheimer’s Disease
AD is the most common form of dementia (13). Symptoms include
impairment of memory and disturbances in reasoning, planning,
and perception. This disease is characterized by the pathological
hallmarks of amyloid plaques and neurofibrillary tangles (NFTs)
that contain hyperphosphorylated tau aggregates, found in the
hippocampus, cortex, and other brain regions. The close relationship between iron and the pathogenesis of AD is supported
by several lines of evidence (14). First, iron accumulates in the
same brain regions that exhibit β-amyloid (Aβ) accumulation, including the parietal cortex and hippocampus, and most amyloid
plaques are found to contain iron. Second, amyloid precursor
protein (APP) mRNA carries an iron-responsive element (IRE)
in its 5’-untranslated region (5’-UTR). Third, iron can downregulate the endoprotease, furin. This can lead to the suppression of
α-secretase, and hence altered processing of APP and elevated
Aβ production. Fourth, iron accumulation is associated with NFT
formation, as iron binding to tau protein is found to precede the
aggregation of hyperphosphorylated tau and the subsequent
formation of NFTs. The cause of iron accumulation in AD still
remains largely unknown. The presence of the C2 allele of transferrin and the C282Y allele of the haemochromatosis gene have
been found to be associated with increased risk of developing
AD (15). Furthermore, deficiency of functional tau has also been
found to impair APP trafficking to the cell surface and hence iron
release mediated by APP-MTP1 coupling (16).
Parkinson’s Disease
PD is the second most common neurodegenerative disorder
(17). It manifests as resting tremor, bradykinesia, rigidity, and
postural instability, and is caused by the selective loss of the
dopaminergic neurons in the substantia nigra pars compacta
(SNpc). Abnormal accumulation of α-synuclein in the form of
Lewy bodies in SNpc is the major pathological hallmark (18).
There are numerous reports showing an increased level of iron
in the SNpc of PD patients (19). Studies indicate that accumulated iron in the SNpc is a critical factor in the degeneration of
these neurons in PD, and that the level of iron accumulation in
the SNpc is associated with the severity of PD symptoms (19–
22). Recent studies showed that both increased non-transferrin–
bound iron (NTBI) uptake via DMT1 and lactoferrin/lactoferrin receptor (Lf/LfR)-mediated processes, as well as transferrin-bound
iron (TBI) uptake into mitochondria mediated by Tf/TfR2, may
contribute to iron accumulation in PD (20, 23). The exact role
of iron in the pathogenic process of PD is not fully understood
(9). Apart from free radicals production, iron may also promote
auto-oxidation of dopamine in SNpc neurons and hence induce
oxidative stress. On the other hand, there is evidence suggesting that iron is related to α-synuclein toxicity and aggregation.
For example, α-synuclein aggregation can be facilitated by exposure to iron and iron-induced reactive oxygen species (ROS).
Additionally, the 5’-UTR of α-synuclein contains an IRE, implying that elevated iron levels could increase posttranscriptional
α-synuclein expression, facilitating its aggregation. Furthermore,
the toxicity induced by α-synuclein in the presence of iron results
in apoptosis, which is thought to be partly responsible for dopaminergic neuronal death. α-synuclein has also been reported to
be a ferrireductase that can generate highly active Fe2+, facilitating free radical generation. Since mitochondria are known to
use and store iron, sustained or enhanced iron import into mitochondria in the presence of defective complex I assembly has
been hypothesized to cause oxidative stress and contribute to
PD (23).
Conclusions
Despite the indispensable role of iron in normal brain function,
iron overload is known to cause oxidative damage and neurodegeneration. However, we currently lack specific and effective means of iron chelation that can slow down or reverse the
progression of the resulting neurodegeneration. Understanding
the molecular mechanisms of iron accumulation in the brain will
hopefully lead to the identification of disease-modifying strategies that target the source of iron abnormality and the treatment
of iron-related neurodegeneration.
Treatment Strategies
A number of iron chelators have been suggested as possible
treatments for iron-related neurodegeneration. Deferoxamine
(DFO), a natural iron scavenger isolated from Streptomyces
pilosus, was among the first-generation iron chelation drugs
for treating iron overload. In one clinical trial, DFO significantly
slowed AD progression (24), but its poor BBB penetration and
short half-life makes it a less than ideal therapy. Deferiprone
(DFP), which can pass through BBB, has been shown to be therapeutically effective in different animal and cell culture models
of neurodegeneration, including PD. Clinical trials are ongoing
to assess its treatment efficacy (25). Currently, most available
iron chelators lack cellular specificity and can lead to a number
of undesirable side-effects. The search for better iron restriction
methods is therefore ongoing.
Hepcidin is a newly discovered short peptide found to possess antimicrobial and iron-regulatory functions. It can bind to
MTP1 to regulate iron levels, and inflammation and elevated
iron levels can induce hepcidin expression (26). We have shown
that hepcidin is also expressed throughout the central nervous
system and its expression increases with age. The widespread
distribution of hepcidin in the brain and its presence in the peripheral organs implies that the peptide may also play a central role in brain iron homeostasis. We have also demonstrated
that hepcidin may have unique functions in the brain, not only
inhibiting iron release but also reducing uptake by downregulating DMT1 and TfR1 expression (27, 28). Our data also suggest
a reduced net iron uptake in response to hepcidin treatment,
which may help relieve iron overloading (27). It is not clear if
hepcidin levels are altered in neurodegeneration characterized
by brain iron accumulation. Further investigation is needed to
assess the potential role of hepcidin in modifying iron-induced
pathology.
REFERENCES
1. L. Zecca, M. B. Youdim, P. Riederer, J. R. Connor, R. R. Crichton, Nat. Rev. Neurosci. 5, 863 (2004).
2. T. A. Rouault, Nat. Rev. Neurosci. 14, 551 (2013).
3. Y. Ke, Z. M. Qian, Prog. Neurobiol. 83, 149 (2007).
4. P. Dusek, J. Jankovic, W. Le, Neurobiol. Dis. 46, 1 (2012).
5. R. Dringen, G. M. Bishop, M. Koeppe, T. N. Dang, S. R. Robinson, Neurochem. Res. 32, 1884 (2007).
6. Y. Ke, Z. M. Qian, Lancet Neurol. 2, 246 (2003).
7. T. A. Rouault, S. Cooperman, Semin. Pediatr. Neurol. 13, 142 (2006).
8. C. Borie et al., J. Neurol. 249, 801 (2002).
9. J. Sian-Hulsmann, S. Mandel, M. B. Youdim, P. Riederer, J. Neurochem. 118, 939 (2011).
10. M. Matsuzaki et al., Brain Res. 1004, 83 (2004).
11. A. Yamamoto et al., J. Neurochem. 82, 1137 (2002).
12. P. W. Mantyh et al., J. Neurochem. 61, 1171 (1993).
13. C. Qiu, M. Kivipelto, E. von Strauss, Dialogues Clin. Neurosci. 11, 111 (2009).
14. S. Altamura, M. U. Muckenthaler, J. Alzheimers Dis. 16, 879 (2009).
15. A. Lleo et al., J. Neurol. Neurosurg. Psychiatry 72, 820 (2002).
16. P. Lei et al., Nat. Med. 18, 291 (2012).
17. L. M. de Lau, M. M. Breteler, Lancet Neurol. 5, 525 (2006).
18. A. G. Kazantsev, A. M. Kolchinsky, Arch. Neurol. 65, 1577 (2008).
19. S. L. Rhodes, B. Ritz, Neurobiol. Dis. 32, 183 (2008).
20. J. Salazar et al., Proc. Natl. Acad. Sci. U.S.A. 105, 18578 (2008).
21. W. R. Martin, M. Wieler, M. Gee, Neurology 70, 1411 (2008).
22. G. Bartzokis et al., Magn. Reson. Imaging 17, 213 (1999).
23. P. G. Mastroberardino et al., Neurobiol. Dis. 34, 417 (2009).
24. M. P. Cuajungco, K. Y. Faget, X. Huang, R. E. Tanzi, A. I. Bush, Ann. N.Y. Acad. Sci. 920, 292 (2000).
25. R. B. Mounsey, P. Teismann, Int. J. Cell. Biol. 2012, 983245 (2012).
26. Q. Wang et al., Endocrinology 149, 3920 (2008).
27. F. Du et al., J. Nutr. Biochem. 23, 1694 (2012).
28. F. Du et al., Glia 59, 936 (2011).
Acknowledgments: This work was supported by the grants from the
National Basic Research “973” Program of China (2011CB510004), the
National Natural Science Foundation of China (31271132, 31371092, and
31330035), and the Hong Kong RGC (CUHK 466713).
15
NEURODEGENERATION RESEARCH
Produced by the Science/AAAS Custom Publishing Office
Novel Pathways Regulating Function and Metabolism of ß-Amyloid
Precursor Protein in Alzheimer’s Disease
Yun-Wu Zhang, Guojun Bu, and Huaxi Xu*
Alzheimer’s disease (AD) is the most common neurodegenerative disorder worldwide, defined by two classical hallmark pathologies: extracellular senile plaques and intraneuronal neurofibrillary tangles (NFTs) (1, 2). NFTs are composed of the
hyperphosphorylated microtubule-associated protein tau that is
abnormally phosphorylated primarily by glycogen synthase kinase-3 (GSK-3) and cyclin D kinase 5 (Cdk5) (2). Senile plaques
are composed of heterogeneous small peptides collectively
called β-amyloid (Aβ), derived from the β-amyloid precursor protein (APP) through sequential cleavage by β- and γ-secretases.
APP is synthesized in the endoplasmic reticulum (ER) and transported through the Golgi/trans-Golgi network (TGN) to the plasma
membrane, where it can be cleaved by α-secretase to produce
sAPPα. Non-cleaved APP is re-internalized and is subjected to
amyloidogenic processing for Aβ generation (1). Multiple lines of
evidence suggest that overproduction/aggregation of Aβ in the
brain is the primary cause of AD: Aβ is highly toxic to neurons
and can trigger a cascade of pathogenic events leading to cell
death. Therefore, detailed delineation of the function, processing, and regulated trafficking of APP is crucial for understanding
the mechanism underlying AD pathogenesis and for developing
AD therapeutic strategies.
APP Processing Towards Aß Generation
Full-length APP is a type I transmembrane protein transported
through the constitutive secretory pathway. During its endocytic
trafficking, APP can first be cleaved by β-secretase to release
a soluble APP extracellular domain called sAPPβ. The remaining membrane-associated APP carboxyl-terminal fragment-β (β
CTF) can then be cleaved by γ-secretase to generate Aβ and an
APP intracellular domain (AICD).
The type I transmembrane aspartyl protease beta-site APPcleaving enzyme (BACE1) is the primary β-secretase species.
Optimal enzymatic BACE1 activity requires acidic environments
such as those found in the TGN and endosomes where BACE1
is present in abundance (3). Mechanisms regulating BACE1 trafficking and activity have not been fully elucidated. Sorting nexin
(SNX) family members contain a conserved lipid-binding PX do-
Fujian Provincial Key Laboratory of Neurodegenerative Disease and Aging Research, Institute of Neuroscience, College
of Medicine, Xiamen University, Xiamen, Fujian, China
*Corresponding Author: [email protected]
16
main and play important roles in membrane trafficking and protein sorting (4). We recently found that a member of the SNX
family, SNX12, interacts with BACE1. Downregulation of SNX12
accelerates BACE1 endocytosis, thus increasing Aβ production,
whereas overexpression of SNX12 has the opposite effect. In addition, SNX12 levels are decreased in AD brains, suggesting that
changes in SNX12 levels may contribute to AD pathology (5). We
also found that the human CUTA protein, another novel protein
that interacts with BACE1, regulates its intracellular trafficking.
Downregulation of CUTA can decelerate intracellular trafficking
of BACE1 from the Golgi/TGN to the cell surface and increase
BACE1-mediated APP processing/Aβ generation (6).
In addition to its function as a β-secretase for APP, we recently
found that BACE1 may also contribute to memory and cognitive deficits associated with AD through an Aβ-independent
mechanism: BACE1 interacts with adenylate cyclases via its
transmembrane domain, resulting in a reduction in cellular cAMP
levels and thus decreased protein kinase A (PKA) activity and
CREB phosphorylation (7). Interestingly, during our search for
new genes that regulate Aβ generation, we identified a new gene
family, designated Rps23rg, whose encoded proteins also interact with adenylate cyclases via their transmembrane domains.
However, RPS23RG proteins increase cellular cAMP levels to
activate PKA, causing increased CREB phosphorylation and
GSK-3 phosphorylation. Phosphorylation of GSK-3 inhibits its
activity, resulting in reduced Aβ generation and tau phosphorylation (8, 9).
γ-cleavage is the last step in APP processing to generate Aβ
peptides. In addition to cleaving APP, the high molecular mass,
membrane-bound γ-secretase complex cleaves many other substrates such as Notch, Cadherin, and ErbB4. γ-secretase consists of four essential components: presenilin (PS, PS1, or PS2),
nicastrin, anterior pharynx-defective-1 (APH-1), and presenilin
enhancer-2 (PEN-2). We and others have shown that deficiency
in any one of these may dramatically affect the stability and intracellular trafficking of other components and impair γ-secretase
activity (1).
PS1 is the catalytic component of the γ-secretase complex.
In addition to cleaving γ-secretase substrates, PS1 has been
shown to have other functions, some of which are independent of
γ-enzymatic activity. For example, we and others show that PS1
can reciprocally regulate the intracellular trafficking of APP (see
next page) as well as several other membrane proteins (1, 10).
Functional Roles for APP and Its Metabolites
Since its identification as the precursor of Aβ, APP has been
studied extensively. However, the physiological function of APP
remains largely undetermined. APP is proteolyzed into various
fragments during intracellular trafficking and these APP metabolites mediate various and sometimes opposing functions. The net
effect of APP on cellular activity may be determined by the relative amounts of APP itself and its various metabolites.
In cells and brains deficient in APP, we observed an elevation
of Cdk5 activity where tau phosphorylation can be inhibited by
re-expressing APP or sAPPα. In addition, APP-deficient neurons
exhibit reduced metabolism and survival rates and are more susceptible to excitotoxic glutamate-induced apoptosis through a
mechanism involving Cdk5 activation. Our results define a novel
neuroprotective function for APP, specifically the extracellular
APPα domain, in preventing tau hyperphosphorylation through
the suppression of Cdk5 overactivation (11).
APP undergoes rapid anterograde transport in neurons. During its transport, APP interacts with kinesin-I and functions as
a membrane-associated kinesin-I receptor to mediate axonal
transport of β-secretase and PS1 (12, 13). We find that APP can
regulate cell surface delivery of the PS1/γ-secretase; APP deficiency accelerates transport of PS1 from the TGN to the cell
surface and increases cell surface levels of PS1, which can be
reversed by restoring APP levels (14). APP dosage also markedly decreases retrograde transport of nerve growth factor (15).
Moreover, APP interacts with the choline transporter and affects
its endocytosis (16). Together, these findings suggest that APP
plays a critical role in regulating protein trafficking.
Using AICD as bait, we identified a mitochondrial carrier protein
as an APP-interacting protein and designated it as appoptosin.
We found that elevated appoptosin-mediated heme biosynthesis
induced the release of reactive oxygen species and activated
intrinsic caspase-dependent apoptosis. Importantly, appopotosin
levels were upregulated in neurons treated with Aβ and in AD
brains, whereas downregulation of appoptosin prevents cell
death and caspase activation caused by Aβ insult, thereby
implicating a novel appoptosin-dependent mechanism underlying
Aβ neurotoxicity. Moreover, we found that although APP interacts
with appoptosin through the AICD domain, AICD was unable to
rescue appoptosin-induced cell death. These results suggest
that membrane-associated domains in the full-length APP and
APP CTFs are required to inhibit appoptosin-induced apoptosis.
Hence, membrane-anchored APP may interact with and retain
a certain amount of appoptosin in the cytosol, thus keeping the
level of appoptosin in mitochondria from being elevated for more
heme production in the presence of cellular insults or under
pathological conditions. Since membrane-dissociated AICD has
little effect on appoptosin-induced caspase activation, this could
imply that membrane-associated APP/appoptosin complexes
can be released and transported to mitochondria upon
γ-cleavage to increase heme synthesis and apoptosis. These
results demonstrate a function of APP in mediating trafficking of
nascent appoptosin from the cytosol to mitochondria (17).
Cytosolic AICD within the cell may translocate into the nucleus to regulate the transcription of several genes such as APP,
GSK-3β, BACE1, and low density lipoprotein receptor-related
protein 1 (LRP1) (18). We also find that AICD can bind to the
promoter region of the epidermal growth factor receptor (EGFR)
gene and regulate its expression. Consistent with the notion that
dysregulation of EGFR expression and activation is involved in
cancers, we found that PS1/γ-secretase deficient mice have increased EGFR levels and increased tumorigenesis, in particular
skin cancer. As AICD is generated concurrently with Aβ, which
is elevated in AD, our results imply that there is a negative correlation between AD and cancer incidence and that the strategy
for inhibiting PS1/γ-secretase activity to treat AD may increase
the risk of tumorigenesis. Both implications are supported by
recent findings that epidemiological studies have identified an
inverse association between cancer and AD (19), while AD clinical trials of the γ-secretase inhibitor semagacestat from Eli Lilly
have demonstrated that patients receiving the drug have an increased risk of skin cancer compared with those who received
a placebo.
Regulation of APP Intracellular Trafficking
Subcellular APP trafficking to divergent sAPPα or Aβ cleavage
pathways is critical to neurodegenerative onset, and mechanisms underlying APP trafficking are therefore integral to determining neuropathological AD outcome. We found that SNX17
can interact with APP in the early endosome and that downregulation of SNX17 leads to reduced APP levels with a concomitant
increase in Aβ (20). In addition, we and others have demonstrated that members of the low-density lipoprotein (LDL) receptor
family, including LRP1, LRP1B, SorLA/LR11, and apolipoprotein
E receptor 2, interact with APP and regulate its endocytic trafficking and Aβ generation. Moreover, we have shown that one of
the γ-secretase components, PS1, also regulates APP trafficking where loss of PS1 results in increased budding/generation of
APP-containing vesicles from both the ER and TGN, along with
a concomitant increase in cell surface localization of APP. Moreover, familial AD-linked PS1 variants are significantly impaired in
vesicle budding, thereby attenuating cell surface delivery of APP
(10). We also found that PS1 interacts with phospholipase D1
(PLD1), a phospholipid-modifying enzyme regulating membrane
trafficking events, and recruits PLD1 to the Golgi/TGN, thus potentially altering APP trafficking. Indeed, PLD1 overexpression
promotes budding of APP vesicles from the TGN, concomitantly increasing cell surface levels of APP (21, 22).
Perspective
Overproduction and aggregation of Aβ in the brain are key patho-
17
NEURODEGENERATION RESEARCH
Produced by the Science/AAAS Custom Publishing Office
genic events in AD. Studies from our group and others have revealed novel pathways by which APP function and processing are
regulated. Further studies investigating the function and regulation of APP in AD will not only help to elucidate the mechanism
underlying disease pathogenesis, but also to identify new targets
for disease treatment.
REFERENCES
28, 449 (2000).
13. A. Kamal, A. Almenar-Queralt, J. F. LeBlanc, E. A. Roberts, L. S. Goldstein, Nature 414, 643 (2001).
14. Y. Liu et al., J. Biol. Chem. 284, 12145 (2009).
15. A. Salehi et al., Neuron 51, 29 (2006).
16. B. Wang, L. Yang, Z. Wang, H. Zheng, Proc. Natl. Acad. Sci. U.S.A. 104, 14140 (2007).
1. Y. W. Zhang, R. Thompson, H. Zhang, H. Xu, Mol. Brain 4, 3 (2011).
2. V. M. Lee, M. Goedert, J. Q. Trojanowski, Annu. Rev. Neurosci. 24, 17. H. Zhang et al., J. Neurosci. 32, 15565 (2012).
1121 (2001).
3. R. Yan, P. Han, H. Miao, P. Greengard, H. Xu, J. Biol. Chem. 276, 11. P. Han et al., J. Neurosci. 25, 11542 (2005).
12. A. Kamal, G. B. Stokin, Z. Yang, C. H. Xia, L. S. Goldstein, Neuron 36788 (2001).
18. Q. Liu et al., Neuron 56, 66 (2007).
19. J. A. Driver et al., B.M.J. 344, e1442 (2012).
20. J. Lee et al., J. Biol. Chem. 283, 11501 (2008).
4. R. D. Teasdale, B. M. Collins, Biochem. J. 441, 39 (2012).
21. D. Cai et al., Proc. Natl. Acad. Sci. U.S.A. 103, 1941 (2006).
5. Y. Zhao et al., Mol. Neurodegener. 7, 30 (2012).
22. D. Cai et al., Proc. Natl. Acad. Sci. U.S.A. 103, 1936 (2006).
6. Y. Zhao, Y. Wang, J. Hu, X. Zhang, Y. W. Zhang, J. Biol. Chem. 287, 11141 (2012).
Acknowledgments: This work was supported by grants from National
7. Y. Chen et al., J. Neurosci. 32, 11390 (2012).
Natural Science Foundation of China (81225008 and 81161120496),
8. X. Huang et al., Hum. Mol. Genet. 19, 3835 (2010).
Natural Science Foundation of Fujian Province of China (2009J06022
9. Y. W. Zhang et al., Neuron 64, 328 (2009).
and 2010J01233), the Fundamental Research Funds for the Central
10. D. Cai et al., J. Biol. Chem. 278, 3446 (2003).
Universities of China, and the Fok Ying Tung Education Foundation.
Role of Tau Hyperphosphorylation in Alzheimer’s
Disease-Associated Neurodegeneration
Jian-Zhi Wang*, Qing Tian, and Di Gao
Tau proteins play an important role in maintaining the stability of
the neuronal cytoskeleton system. In Alzheimer’s disease (AD),
tau is abnormally hyperphosphorylated and aggregates into
paired helical filaments (PHFs) forming neurofibrillary tangles
(NFTs) in neurons (1). Clinical investigations have shown that the
intracellular accumulation of NFTs is positively correlated with
the severity of dementia. The transmission of abnormal tau or
NFTs from the entorhinal cortex to the hippocampus and cerebral cortex matches the clinical manifestation, which is the international gold standard at present for assessing AD progression
(Braak grading) (2). Recent studies suggest that the toxicity of
β-amyloid protein (Aβ, another pathogenic factor in AD) needs
Department of Pathology and Pathophysiology,
Key Laboratory of the Ministry of Education of China for
Neurological Disorders, Tongji Medical College, Huazhong
University of Science and Technology, Wuhan, China
*Corresponding Author: [email protected]
18
the tau protein as a mediator (3). These data together suggest
that abnormal tau plays an important role in the onset and evolution of neurodegeneration and the learning/memory deficits in
AD patients.
To date, no tau gene mutation has been found in the AD patients. The main focus of tau research therefore has been on
posttranslational modifications, of which hyperphosphorylation is
the most extensively studied. The imbalance of protein kinases
(generally upregulation) and protein phosphatases (generally
downregulation) is the direct cause for abnormal tau hyperphosphorylation. Among various protein kinases and protein phosphatases, glycogen synthase kinase-3β (GSK-3β) and protein
phosphatase-2A (PP2A) are most heavily involved in AD-like tau
hyperphosphorylation (4–6).
Role of GSK-3ß in Tau Hyperphosphorylation
GSK-3β was the first tau kinase to be identified; it can phosphorylate tau at multiple sites in vitro (7). In vivo activation of GSK-3β
causes tau hyperphosphorylation at
several sites related to impairment
of spatial memory in AD (8). Multiple factors can cause tau hyperphosphorylation through GSK-3β
activation, including peroxide nitrite
(9), advanced glycation end products (10, 11), persistent light illumination (12), endoplasmic reticulum
stress (13, 14), Aβ (8, 15), and proteasome dysfunction (16, 17). While
activating GSK-3β inhibits long-term
potentiation (LTP), inhibiting it promotes LTP through mechanisms
involving the presynaptic release of
neurotransmitter (18, 19). Inhibiting
GSK-3β significantly attenuates tau
hyperphosphorylation and improves
memory (20, 21), while overexpression of neuroprotective neuroglobin
decreases tau hyperphosphorylation
by inhibiting GSK-3β (22, 23).
In the brains of rats, inhibition of Figure 1. Schematic summarizing our current understanding of the various signaling pathways through which
phosphatidylinositol 3-kinase (PI3K) tau acts within the cell.
alone induces a transient activation
of GSK-3β, while simultaneous inhibition of PI3K and protein ki- to clearly define the conditions, but this work shows promise.
nase C (PKC) results in a sustained activation of GSK-3β, leading to prolonged tau hyperphosphorylation and spatial memory Role of PP2A in Tau Hyperphosphorylation
deficits with reduction of acetylcholine (8, 24, 25). Phosphoryla- Tau proteins are dephosphorylated by protein phosphatases
tion of GSK-3β at serine-9 is recognized to be negatively corre- such as PP2A; therefore, inactivation of phosphatases results
lated with GSK-3β activation, while cleavage of GSK-3β by cal- in tau hyperphosphorylation. PP2A is the most effective at depain counteracts the inhibitory effect of serine-9 phosphorylation phosphorylating abnormally hyperphosphorylated tau isolated
on GSK-3β activity induced by hydrogen peroxide (26).
from the AD brains (32, 33). In vitro, PP2A can dephosphoryPhosphorylation of tau by protein kinase A (PKA) primes tau, late abnormal tau at multiple sites and thus restore its biologimaking it a better substrate for GSK-3β, and at least partially cal activity (34). Inhibiting PP2A in vivo by injection of okadaic
explaining why the GSK-3β-preferred sites on tau can be hyper- acid or homocystine into the brain, or in vitro by incubating cells
phosphorylated even after transient activation by PKA (27, 28). with PP2A inhibitors, induces AD-like tau hyperphosphorylaInterestingly, we demonstrated that tau hyperphosphorylation by tion, intracellular accumulation, axoplasmic transport deficits,
GSK-3β seemed to be required for hippocampal neurogenesis and learning/memory dysfunction (35, 36). PP2A is activated
in the dentate gyrus. Further, tau phosphorylation and GSK-3β in the astrocytes of tg2567 mice—a widely used amyloidoactivation are essential for the adult neurogenesis in the subven- genic model of AD—and activation of PP2A stimulates the
tricular zone (SVZ), another niche of neurogenesis in the adult migration of astrocytes to the amyloid plaques through p38
brains (29, 30), suggesting that the neurogenesis in the brain MAPK inhibition, indicating that PP2A deficits observed in
may be tightly regulated by local microenvironments.
AD brains may cause Aβ accumulation by hindering astrocyte
Due to its role in tau phosphorylation, GSK-3β has been con- migration (37).
sidered as a drug target for inhibiting neurodegeneration. HowSince PP2A activity is significantly inhibited in AD brains, upregever, GSK-3β has other functions in the cell, so complete inhibi- ulating PP2A seems a promising therapeutic strategy. However,
tion would likely be detrimental. Recent studies that attempted currently no specific activator of PP2A exists, making the search
spatiotemporal targeting of abnormal GSK-3β activation found for an upstream regulator of PP2A a critical mission. We have
that pathologies and memory deficits in an AD mouse model demonstrated that GSK-3β activation can inhibit PP2A by upregcould be effectively attenuated (31). Further studies are needed ulating protein tyrosine phosphatase 1Β, which phosphorylates
19
NEURODEGENERATION RESEARCH
Produced by the Science/AAAS Custom Publishing Office
PP2A at tyrosine-307 (38, 39), suggesting that activation of GSK3β and inhibition of PP2A may create a positive feedback loop
that promotes tau hyperphosphorylation. Recent studies suggest
that PP2A activity is regulated by an endogenous nuclear protein inhibitor called inhibitor 2 of PP2A (I2PP2A) that is increased
in AD brains. By generating a phosphorylation site-specific antibody, we showed that I2PP2A is phosphorylated at serine-9,
resulting in cytoplasmic accumulation, as seen in AD brains (40 ).
We have also shown that zinc induces PP2A inactivation and
tau hyperphosphorylation through Src-dependent tyrosine-307
phosphorylation of PP2A (41). In addition, agricultural pesticides
can also inhibit PP2A and activate GSK-3β (42). Conversely, administration of betaine or nicotinamide mononucleotide adenylyltransferase 2 (the latter being a key enzyme involved in energy
metabolism that shows decreased expression in the AD brain),
can attenuate tau phosphorylation through activation of PP2A
and improve the memory deficits in AD mice (43, 44).
Tau Glycosylation and Phosphorylation
In the AD brain, abnormally phosphorylated tau protein is also
highly glycosylated through N-linked glycosidic bonds. Deglycosylation releases functional tau proteins from NFTs isolated from
the AD brains (45). O-glycosylation of tau has been found both
in normal and AD brains, and seems to inhibit phosphorylation
of tau proteins, as evidenced by the negative correlation seen
between tau O-glycosylation and phosphorylation in a rat model
of starvation (46).
Cellular Effects of Tau Hyperphosphorylation
To better understand how abnormal tau hyperphosphorylation
causes neurodegeneration, we expressed exogenous tau proteins in HEK293 (a human embryonic kidney cell line that does
not express endogenous tau) or N2a (a mouse neuroblastoma
cell line that expresses very low levels of endogenous tau), and
treated these cells, plus primary hippocampal neurons, with
different pro-apoptotic factors. We measured cell viability and
apoptotic markers after treatment as a proxy for neurodegeneration. We found that tau phosphorylation was significantly
increased upon treatment with pro-apoptotic factors and, unexpectedly, the intracellular accumulation of hyperphosphorylated
tau led the neurons to abort a chemically induced apoptosis program, both in vitro and in vivo, through pathways involving the
survival factor β-catenin (47). Further studies demonstrated that
overexpression of tau could antagonize apoptosis promoted by
Aβ and other pro-apoptotic factors, resulting in reduced p53 levels, decreased release of cytochrome c from mitochondria, and
inhibition of caspase-9/caspase-3 (48, 49), while dephosphorylation of tau induced apoptosis through failed dephosphorylation of
Bcl-2 (50). We also showed that elevation of I2PP2A or inhibition
of PP2A caused tau hyperphosphorylation with concomitant upregulation of p53 and Akt, a key serine/threonine protein kinase
20
(51). Since p53 and Akt serve as pro- and anti-apoptotic factors,
respectively, we speculate that the anti-apoptotic effects of Akt
may overwhelm the pro-apoptotic role of p53, leading the cells
to abort apoptosis and partially clarifying the role of hyperphosphorylated tau in blocking apoptosis. Taken together, our data
strongly suggest that tau hyperphosphorylation plays a dual role
in leading neurons to escape apoptosis while at the same time
promoting chronic neurodegeneration.
Potential Measurements for Clinical
Diagnosis of AD
The early clinical diagnosis of AD is still a challenge. Since cerebrospinal fluid (CSF) surrounds the brain tissue, biochemical
changes in the brain can be reflected in CSF. Therefore, quantitative measurement of the phosphorylated tau protein at specific sites in CSF may help to follow the progression of neurodegeneration. Since the level of phosphorylated tau in the form
of PHFs in CSF is too low to be detected by the conventional
enzyme-linked immunosorbent assay (ELISA), we have established an ultrasensitive bienzyme-substrate-recycle enzymelinked immunosorbent assay by linking the conventional ELISA
(with high specificity) with a bienzyme substrate cycle (with high
sensitivity). Using this method, we found that phosphorylation of
tau at the PHF-1 monoclonal antibody epitope was significantly
increased, as was the level of phosphorylated neurofilament proteins in CSF of AD patients (52, 53).
Using quantitative magnetic resonance phase-corrected imaging, we detected iron deposition in specific brain regions of
AD patients. The iron concentration in the parietal cortex was
positively correlated with the severity of cognitive impairment in
AD patients (54), which may represent a potential biomarker to
evaluate the progression of AD.
Blood oxygenation level dependent (BOLD) functional magnetic resonance imaging (fMRI) was used to analyze the relationship between the altered resting T2 fMRI signal and memory
capacity/tau phosphorylation in rats. We observed that tau hyperphosphorylation and memory deficit induced by activating
GSK-3β was positively correlated with a reduced BOLD signal in
rats (55), suggesting that the resting BOLD signal may serve as
a noninvasive quantitative marker for predicting AD-like spatial
memory deficits and tau hyperphosphorylation.
The level of acetylcholine, an important excitatory neurotransmitter, is decreased in the early stages of AD. Measurement of
acetylcholine in the brain may therefore help with the clinical diagnosis of AD. However, current methods for brain acetylcholine
measurement are invasive and not feasible to use in the clinic.
To address this problem, we employed both noninvasive proton
magnetic resonance spectroscopy (MRS), to measure cholinecontaining compounds, and high-performance liquid chromatography to measure the extracellular levels of acetylcholine by
microdialysis of samples from different brain regions. Analysis of
the data suggested that the choline signal intensity can be used
as a noninvasive indicator of the acetylcholine level in rat brains
(56), a procedure that could be applied in humans.
In summary (see Fig. 1), we have found that (1) GSK-3β and
PP2A can actively and reversely regulate tau phosphorylation;
(2) the activity of GSK-3β and PP2A can be modulated by multiple factors; (3) priming phosphorylation of tau by PKA makes
tau a better substrate for GSK-3β; (4) O-glycosylation inhibits
tau phosphorylation, while N-glycosylation and nitration may promote tau phosphorylation; and (5) hyperphosphorylation of tau
enables cells to block apoptosis through competitively reducing
the phosphorylation of β-catenin, and also inhibits proteasome
activity, promotes tau aggregation, damages axonal transport,
and impairs synaptic functions, all of which can trigger chronic
neurodegeneration.
24. G. G. Xu, Y. Q. Deng, S. J. Liu, H. L. Li, J. Z. Wang, Acta Bioch. 26. Y. Feng et al., J. Neurochem. 126, 234 (2013).
27. S. J. Liu et al., J. Biol. Chem. 279, 50078 (2004).
28. Y. Zhang, H. L. Li, D. L. Wang, S. J. Liu, J. Z. Wang, J. Neural REFERENCES
30. X. P. Hong et al., Neurochem. Res. 36, 288 (2011).
31. C. X. Peng et al., Neurobiol. Aging 34, 1555 (2013).
32. J. Z. Wang, C. X. Gong, T. Zaidi, I. Grundke-Iqbal, K. Iqbal, J. Biol. Chem. 270, 4854 (1995).
33. J. Z. Wang, I. Grundke-Iqbal, K. Iqbal, Brain Res. Mol. Brain Res. 38, 200 (1996).
34. J. Z. Wang, I. Grundke-Iqbal, K. Iqbal, Eur. J. Neurosci. 25, 59 (2007).
35. L. Sun et al., Neuroscience 118, 1175 (2003).
1. I. Grundke-Iqbal et al., Biol. Chem. 261, 6084 (1986).
36. Y. Yang et al., J. Neurochem. 102, 878 (2007).
2. I. Alafuzoff et al., Brain Pathol. 18, 484 (2008).
37. X. P. Liu et al., Glia 60, 1279 (2012).
3. L. M. Ittner, J. Götz, Nature Rev. Neurosci. 12, 65 (2011).
38. X. Q. Yao et al., Biochem. J. 437, 335 (2011).
4. J. Z. Wang, F. Liu, Prog. Neurobiol. 85, 148 (2008).
39. G. P. Liu et al., Neurobiol. Aging 29, 1348 (2008).
5. Q. Tian, J. Z. Wang, Neurosignals 11, 262 (2002).
40. G. Yu et al., Neurobiol. Aging 34, 1748 (2013).
6. J. Z. Wang, Y. Y. Xia, I. Grundke-Iqbal, K. Iqbal, J. Alzheimers Dis. 41. Y. Xiong et al., Neurobiol. Aging 34, 745 (2013).
33 Suppl. 1, S123 (2012).
7. J. Z. Wang, I. Grundke-Iqbal, K. Iqbal, Eur. J. Neurosci. 25, 59 Transm. 113, 1487 (2006).
29. X. P. Hong et al., Hippocampus 20, 1339 (2010).
Bioph. Sin. 37, 349 (2005).
25. Y. Wang et al., J. Neurochem. 106, 2364 (2008).
(2007).
42. N. N. Chen et al., J. Alzheimers Dis. 30, 585 (2012).
43. G. S. Chai et al., J. Neurochem. 124, 388 (2013).
44. X. S. Cheng et al., J. Alzheimers Dis. 36, 185 (2013).
8. S. J. Liu et al., J. Neurochem. 87, 1333 (2003).
45. J. Z. Wang, I. Grundke-Iqbal, K. Iqbal, Nat. Med. 2, 871 (1996).
9. Y. J. Zhang, Y. F. Xu, Y. H. Liu, J. Yin, J. Z. Wang, FEBS Lett. 579, 46. X. Li, F. Lu, J. Z. Wang, C. X. Gong, Eur. J. Neurosci. 23, 2078 6230 (2005).
(2006).
10. X. H. Li et al., Cell Death Dis. 4, e673 (2013).
47. H. L. Li et al., Proc. Natl. Acad. Sci. U.S.A. 104, 3591 (2007).
11. X. H. Li, et al., Neurobiol. Aging 33, 1400 (2012).
48. Z. F. Wang et al., J. Alzheimers Dis. 20, 145 (2010).
12. Z. Q. Ling et al., J. Alzheimers Dis. 16, 287 (2009).
49. H. H. Wang et al., J. Alzheimers Dis. 21, 167 (2010).
13. Z. C. Liu et al., J. Alzheimers Dis. 29, 727 (2012).
50. X. A. Liu et al., J. Alzheimers Dis. 19, 953 (2010).
14. Z. Q. Fu et al., J. Alzheimers Dis. 21, 1107 (2010).
51. G. P. Liu et al., Neurobiol. Aging 33, 254 (2012).
15. Z. F. Wang et al., J. Neurochem. 98, 1167 (2006).
52. Y. Y. Hu et al., Am. J. Pathol. 160, 1269 (2002).
16. Q. G. Ren, X. M. Liao, Z. F. Wang, Z. S. Qu, J. Z. Wang, FEBS Lett. 53. Y. Y. Hu et al., Neurosci. Lett. 320, 156 (2002).
580, 2503 (2006).
54. W. Z. Zhu et al., Radiology 253, 497 (2009).
17. Y. H. Liu et al., Neurobiol. Aging 30, 1949 (2009).
55. Z. H. Hu et al., Acta Bioch. Bioph. Sin. 36, 803 (2004).
18. L. Q. Zhu et al., J. Neurosci. 27, 12211 (2007).
56. X. C. Wang, X. X. Du, Q. Tian, J. Z. Wang, Neurochem. Res. 33, 814 19. L. Q. Zhu et al., J. Neurosci. 30, 3624 (2010).
(2008).
20. Y. S. Xiong et al., CNS Neurol. Disord.-DR 12, 436 (2013).
21. J. Yin et al., J. Pineal Res. 41, 124 (2006).
Acknowledgments: This work was supported by the Natural Science
22. P. Zhou et al., Rejuv. Res. 14, 669 (2011).
Foundation of China (81171195 and 91132305) and the Ministry of
23. L. M. Chen et al., J. Neurochem. 120, 157 (2012).
Science and Technology (2013DFG32670).
21
NEURODEGENERATION RESEARCH
Produced by the Science/AAAS Custom Publishing Office
Dopaminergic Modulation of Astrocyte Functions in the
Pathogenesis of Parkinson’s Disease
Yingjun Liu and Jiawei Zhou*
Dopamine is one of the most versatile neurotransmitters in the
central nervous system (CNS). It
is involved in multiple brain functions including movement coordination, reward, learning, and memory. Dopamine functions through
five G protein-coupled dopamine
receptors (D1-5) that are highly
expressed in many subtypes of
neurons throughout the CNS. Interestingly, accumulating evidence
has demonstrated that dopamine
receptors are also expressed in
glial cells, although expression is
much lower than in neuronal cells
(1–4). Exactly how dopamine regulates glial cell functions in health
and disease through glial dopaFigure 1. Schematic showing signal transduction pathways involved in dopamine D1, D2 receptors, or PI-linked
mine receptors is still not fully unD1-like receptor-modulated biosynthesis and release of FGF-2 in striatal astrocytes. Enhanced release of high
derstood. Glial cells are the most
molecule weight (HMW) and low molecule weight (LMW) FGF-2 leads to increased survival of DA neurons. DA, dopamine; D1, dopamine D1 receptor; D2, dopamine D2 receptor; PI-linked D1LR, phosphatidylinositol-linked D1-like
abundant components of the CNS,
receptor; GluR, glutamate receptor; cAMP, cyclic adenosine monophosphate; p-PKC, phosphorylated protein kinase
constituting 65% to 90% of the total
C; IP3, inositol 1,4,5-trisphosphate; p-PKA, phosphorylated protein kinase A; p-MAPK, phosphorylated mitogen-activated protein kinase; Ca, calcium; CaM, calmodulin; MZF-1, myeloid zinc finger-1; FGF-2, fibroblast growth factor-2.
cells in the mammalian CNS. They
are associated with many aspects
of normal brain function and are important players in neurode- terventions for symptomatic relief of PD, although there are still
generative disease (5–8). Here, we summarize our recent data no available treatments that stop or reverse the progression of
on the dopaminergic modulation of astrocyte functions, especial- the disease (9). Receptor agonists can mimic dopamine funcly in the context of Parkinson’s disease (PD).
tions, directly restoring the dopamine signaling in postsynaptic
neurons in vivo. They may also exert neuroprotective effects by
enhancing
the expression and secretion of neurotrophic factors
Dopamine Signaling Controls Astrocytic FGF-2
such as brain-derived neurotrophic factor (BDNF) and glial cell
Biosynthesis and Secretion
PD is a movement disorders in which dopaminergic neurons in line-derived neurotrophic factor (GDNF) (10), which are known to
the substantia nigra pars compacta selectively degenerate. It is be produced primarily in neuronal cells. GDNF is predominantly
believed that gradual dopamine loss in the striatal brain tissues expressed in parvalbumin-positive neostriatal interneurons in
of PD patients occurs with disease progression and is the princi- normal and injured rodents (11), even though it has been previpal mechanism underlying the cardinal clinical symptoms of PD. ously thought to be largely produced in glial cells (12). ConverseDopamine receptor agonists are among the main therapeutic in- ly, fibroblast growth factor-2 (FGF-2), an important regulator of
cell growth and differentiation for nigral DA neurons, is predominantly synthesized by astrocytes, while other members of the
Institute of Neuroscience, State Key Laboratory of
FGF family are primarily synthesized by neurons (13). Whether
Neuroscience, Shanghai Institutes for Biological Sciences,
the production of astrocyte-derived FGF-2 is influenced by dopaChinese Academy of Sciences, Shanghai, China
*
Corresponding Author: [email protected]
mine receptors, and the biological significance of its regulation in
22
the context of PD, remain elusive. We demonstrated that classical dopamine D1 and D2 receptors potently regulated astrocytic FGF-2 biosynthesis and secretion (14). FGF-2 synthesis
and release were markedly upregulated in primary cultured astrocytes exposed to apomorphine (APO), a non-selective dopamine receptor agonist, in a dose- and time-dependent manner.
Conditioned medium from APO-treated astrocytes (APO-CM)
dramatically enhanced the survival of dopaminergic (DA) neurons in vitro. Interestingly, neutralization of FGF-2 in APO-CM by
using FGF-specific antibodies abrogated the survival-promoting
effect of the conditioned medium upon DA neurons in culture,
indicating that FGF-2 is the major effector component. Further
studies showed that activation of the D1 receptor preferentially
increased protein kinase A (PKA) activity, whereas activation
of the D2 receptor only promoted phosphorylation of mitogenactivated protein kinase (MAPK). The D1 receptor-associated
pathway had an inhibitory effect on the D2 receptor pathway via
PKA and MAPK, but not on cyclic adenosine monophosphate
(cAMP), suggesting a delicate system of regulation for control of
FGF-2 expression in astrocytes (Fig. 1). Importantly, APO-modulated FGF-2 expression was independent of Akt/phosphoinositide 3-kinase signaling (14), which is known to mediate classical
dopaminergic signaling in neurons. Furthermore, APO-induced
enhancement of the biosynthesis and release of FGF-2 was mediated by the activation of transcription factor myeloid zinc finger
protein 1 (MZF-1) in striatal astrocytes (15).
Besides dopamine receptors D1 and D2, the phosphatidylinositol (PI)-linked D1-like receptor is also involved in modulation of
FGF-2 expression in astrocytes (16). SKF83959, a selective PIlinked D1-like receptor agonist, induced a remarkable increase
in the levels of FGF-2 in cultured striatal astrocytes in a classical dopamine D1 and D2 receptor-independent manner. In vivo
administration of SKF83959 reversed neurotoxin 1-methyl-4phenyl-1,2,3,6-tetrahydropypridine (MPTP)-induced reduction in
FGF-2 expression in the nigrostriatum and resulted in marked
protection of dopaminergic neurons from MPTP-induced neurotoxicity. Interestingly, inositol 1,4,5-trisphosphate (IP3)-dependent calcium oscillations were essential for SKF83959-mediated
FGF-2 induction in astrocytes. Collectively, our data suggest
that dopamine signaling plays crucial roles in the regulation of
biosynthesis and secretion of growth factor FGF-2 in astrocytes
via both well-characterized and uncharted intracellular signaling
pathways (Fig. 1). However, we should bear in mind that most of
these pharmacological studies were performed using dopamine
receptor agonists or antagonists. Although these compounds
have been widely used in the field, the possibility that they may
have multiple targets and complex biological consequences cannot be entirely excluded. For example, recent studies showed
that SKF83959 and some of its analogs are also potent allosteric
modulators of the sigma-1 receptor, which modulates calcium
signaling through IP3 receptors (17, 18). Future investigation is
Figure 2. Astrocytic Drd2 functions to suppress inflammatory responses
specifically in astrocytes. Model for communications among astrocytes, microglial, and dopaminergic neurons.
required to determine the underlying mechanisms of SKF83959induced FGF-2 expression.
Astrocytic Dopamine D2 Receptor Activation
Normally Suppresses Neuroinflammation in
the CNS
It has been widely recognized that neuroinflammation primarily mediated by microglia and astrocytes is a common feature
of neurodegenerative diseases including PD and may play important roles in the initiation and progression of neurodegenerative diseases. It has thus been proposed that anti-inflammation
treatment is a promising therapeutic approach in patients with
neurodegenerative disease. However, the molecular properties
and regulation of neuroinflammation in the CNS is poorly understood, hampering attempts at anti-inflammatory treatments. Previous results have demonstrated that dopamine receptors were
expressed in immune-relevant cells, such as macrophages, and
modulated their immune responses in the periphery tissue and
organs, suggesting the possible involvement of dopamine in immune regulation (19, 20). Our own recent studies have shown
that dopamine D2 receptor (Drd2) negatively modulated the expression of inflammatory cytokines in astrocytes, but not microglia (21). Drd2-deficient astrocytes showed hyper-responsiveness
23
NEURODEGENERATION RESEARCH
Produced by the Science/AAAS Custom Publishing Office
both in primary astrocytic cultures and specific astrocytic Drd2
conditional knockout (cKO) mice in response to immune stimuli.
Moreover, global Drd2 KO mice and astrocytic Drd2 cKO mice
were more vulnerable to MPTP-induced neurotoxicity, suggesting a potentially clinically relevant role for Drd2. Furthermore,
we identified alpha-B crystalline (CRYAB), a small heat shock
protein with anti-inflammatory and neuroprotective activities,
that was responsible for astrocytic Drd2-mediated suppression
of neuroinflammation. Deletion of Drd2 resulted in a drastic decrease in the levels of CRYAB expression in cultured astrocytes
and brain tissues, while over-expression of CRYAB specifically
targeting astrocytes dramatically inhibited the upregulation of
inflammatory cytokines in Drd2-null mice. These results indicate that CRYAB levels are tightly controlled by Drd2 and are
responsible for dopamine-mediated immune inhibition (Fig. 2).
However, how dopamine regulates CRYAB expression and how
CRYAB controls immune responses of astrocytes are still open
questions.
Concluding Remarks
Traditionally, as a neurotransmitter, dopamine is deemed to be a
critical mediator of neuron-neuron communication through synapses. However, new evidence elucidating the functional roles of
glial dopamine receptors under both physiological and pathological conditions is challenging the old dogma. Based on observations from our group and others, it is very likely that dopamine
is also involved in neuron-glia interactions. Our studies clearly
demonstrate that dopamine signaling is critical for the regulation
of astrocyte function under both normal and pathologic conditions. These observations also highlight the important roles of
astrocytes in neurodegenerative diseases. Further investigation
is required to unravel the complex and intricate biological functions of dopamine in CNS cell-cell communication.
15. X. Luo, X. Zhang, W. Shao, Y. Yin, J. Zhou, J. Neurochem. 108, 952 Rakic, Proc. Natl. Acad. Sci. U.S.A. 98, 1964 (2001).
2. K. M. Lucin, T. Wyss-Coray, Neuron 64, 110 (2009).
3. I. Miyazaki, M. Asanuma, F. J. Diaz-Corrales, K. Miyoshi, N. Ogawa, Brain Res. 1029, 120 (2004).
REFERENCES
1. Z. U. Khan, P. Koulen, M. Rubinstein, D. K. Grandy, P. S. Goldman-
24
Chengyuan Tang, Tongmei Zhang, and Zhuohua Zhang*
4. M. Mladinov et al., Translat. Neurosci. 1, 238 (2010).
5. N. J. Maragakis, J. D. Rothstein, Nat. Clin. Pract. Neurol. 2, 679 (2006).
6. B. Liu, J. S. Hong, J. Pharmacol. Exp. Ther. 304, 1 (2003).
7. W. J. Streit, Glia 40, 133 (2002).
8. C. S. Lobsiger, D. W. Cleveland, Nat. Neurosci. 10, 1355 (2007).
9. O. Rascol, C. Goetz, W. Koller, W. Poewe, C. Sampaio, Lancet 359, 1589 (2002).
10. H. Guo et al., Eur. J. Neurosci. 16, 1861 (2002).
11. M. Hidalgo-Figueroa, S. Bonilla, F. Gutierrez, A. Pascual, J. Lopez-
Barneo, J. Neurosci. 32, 864 (2012).
12. M. Ohta et al., Biochem. Biophys. Res. Commun. 272, 18 (2000).
13. B. Reuss, O. von Bohlen und Halbach, Cell Tissue Res. 313, 139 (2003).
14. A. Li et al., FASEB J. 20, 1263 (2006).
(2009).
16. X. Zhang et al., J. Neurosci. 29, 7766 (2009).
17. L. Guo et al., Mol. Pharmacol. 83, 577 (2013).
18. T. Hayashi, T. P. Su, Cell 131, 596 (2007).
19. P. J. Gaskill, L. Carvallo, E. A. Eugenin, J. W. Berman, J. Neuroinflamm. 9, 203 (2012).
20. G. Hasko, C. Szabo, Z. H. Nemeth, E. A. Deitch, J. Neuroimmunol. 122, 34 (2002).
21. W. Shao et al., Nature 494, 90 (2013).
Acknowledgments: This work was supported by grants from the Chinese
Academy of Sciences, National Basic Research Program of China
(2011CBA00408 and 2011CB504102), the Natural Science Foundation of
China (31021063 and 31123002), and the Shanghai Metropolitan Fund for
Proteolytic Pathways in Parkinson’s Disease
Research and Development.
Parkinson’s disease (PD) is the most common neurodegenerative movement disorder and is currently incurable. The pathological hallmarks of the disease include loss of dopaminergic (DA)
neurons in the substantia nigra pars compacta and the formation
of intracellular Lewy bodies in surviving neurons (1). Lewy bodies are formed by the accumulation of a number of unfolded/
misfolded proteins, including synaptic protein α-synuclein. The
role of α-synuclein in selective neurodegeneration in PD remains
unknown. Human studies have demonstrated the critical role of
genetic background in PD pathogenesis, identifying mutations
in at least 14 genes associated with the disease (2–4). Neurons, as postmitotic cells, rely heavily on the
cellular quality control system to remove newly
synthesized or stress-induced proteins that are
misfolded. The ubiquitin-proteasome system
(UPS) and the autophagy-lysosomal pathway
(ALP) are two major proteolytic quality control
pathways. UPS degrades mostly short-lived,
soluble proteins while ALP removes damaged
intracellular components, such as proteins and
organelles, through lysosomes. Recent studies
strongly indicate that impairment of both UPS
and ALP contributes to PD pathogenesis.
Figure 1. Schematic illustration of PPD E3 ligase complex. In the
complex, PINK1 and DJ-1 function as regulatory units (Reg) while parkin
serves as a catalytic domain (Cat). In the presence of ubiquitin conjugating enzyme (E2) and ubiquitin (Ub), PPD promotes poly-ubiquitination of
substrate (S).
Ubiquitin-Proteasome System
UPS plays critical role in cellular quality conFigure 2. Neddylation stabilizes 55 kDa PINK1 cleavage product in cells. (A) Representative imtrol by degrading soluble intracellular proteins. munoblot for FLAG-tagged PINK1 and NEDD8 upon cycloheximide treatment. Duration of treatment
The first step in the UPS degradation process is shown above gel. (B) Quantitation of PINK1 and its cleavage product after cycloheximide treatment.
Results are average from two independent experiments. Values represent mean ± standard error.
is ATP-dependent covalent linkage of a chain of 55kDa PINK1, the 55kDa PINK1 fragment; Control, empty vector. *, P=0.0314 at 6hr treatment time
activated ubiquitin to lysine residues on the tar- point between 55kDa PINK1+NEDD8 and 55kDa PINK1+Control using unpaired, two-tailed t test.
get substrates. This process is accomplished by Adapted from (14).
the sequential action of ubiquitin-activating enzymes (E1), ubiq- genes associated with PD, the products of parkin, PINK1, DJ-1,
uitin-conjugating enzymes (E2), and ubiquitin ligases (E3). The uchL1, and FBXO7 are involved in the ubiquitin pathway (6–11),
ubiquitin chain serves as a signal to target the substrate to the at least three of which have E3 ligase activity: uchl1, FBXO7
proteasome for degradation. In this process, E3 ubiquitin ligases as part of the SKP1/cullin/F-box (SCF) complex, and the parkin/
are critical for the specificity of substrates to be degraded. Many PINK1/DJ-1 (PPD) complex. UchL1 functions as both a ubiquitin
neurodegenerative diseases, including PD, are the consequence E3 ligase and ubiquitin hydrolase (10). FBXO7 is known to reguof dysfunctional folding and trafficking of proteins (5). Compo- late cyclin D via its E3 ligase activity (12). The PPD E3 ligase
nents of the ubiquitin proteasomal pathway as well as chap- complex uses parkin as a catalytic unit and PINK1 and DJ-1 as
erones have been detected in Lewy bodies. Of the 14 known the regulatory units (Fig. 1) (8). Nevertheless, how impairment
of these E3 ligases affects PD pathogenesis remains unknown.
The substrates, particularly those related to PD pathogenState Key Laboratory of Medical Genetics, Xiangya Medical
esis,
have yet to be defined for these previously characterized
School, Central South University, Changsha, Hunan, China
*
PD-related
E3 ligases. However, it has been shown that UchL1
Corresponding Author: [email protected]
25
NEURODEGENERATION RESEARCH
Produced by the Science/AAAS Custom Publishing Office
promotes K63 ubiquitination of α-synuclein in vitro and that
two PD-associated mutations have been shown to reduce E3
ligase activity (10). FBXO7 interacts with a number of cell cycle
regulatory proteins, but does not promote UPS-mediated degradation of these proteins. A recent study suggests that FBXO7
regulates ubiquitination of mitofusin-1, and thereby mitophagy,
via direct interaction with parkin and PINK1 (13). Likewise, the
substrates of the PPD E3 ligase complex include those of previously identified parkin substrates. Overexpression of parkin
substrates or heat shock treatment resulted in parkin substrate
accumulation in PINK1-deficient cells (8). One explanation for
this observation is that the PPD complex functions as a quality control system to degrade misfolded parkin substrates induced by oxidative stress in neurons. Pathogenic mutations
that reduce PPD activity likely disrupt UPS, leading to increased
susceptibility to oxidative stress and accumulation of misfolded proteins in cells, and eventual Lewy body formation and
neuronal death.
The regulation of the PD-associated E3 ligases is not yet well
understood. FBXO7 is reported to regulate parkin E3 ligase activity (13). Nedd8 modification of both parkin and PINK1 is likely
another way to modulate PPD E3 ligase activity. In this case,
PINK1 neddylation results in increased stability of its 55KDa proteolytic fragment, which forms a complex with parkin in the cytosol (Fig. 2). Furthermore, PD associated neurotoxin 1-methyl-4phenylpyridinium (MPP+) suppresses the PPD E3 ligase activity
via, at least partially, inhibiting neddylation of parkin and PINK1
(14). These results suggest a PD pathogenic mechanism of genetic and environmental interaction.
Autophagy-Lysosomal Pathway
Autophagy maintains cellular homeostasis and integrity by recycling metabolic precursors and clearing subcellular debris. In response to environmental cues, the autophagic machinery—products of the autophagy-related genes (Atg)—initiates a series of
membrane inclusion processes that packages the unnecessary
or dysfunctional cellular components and delivers them to the
lysosome for degradation (15). PD-associated genes have only
recently been shown to be involved in regulating autophagy, despite the fact that autophagy has been observed in DA neurons
of the substantia nigra of PD patients for many years.
Parkin and PINK1 have been shown to mediate mitophagic
degradation of depolarized mitochondria (16, 17). Upon depolarization, PINK1 recruits parkin and promotes parkin’s E3
ubiquitin ligase activity resulting in the ubiquitination of outer
membrane proteins and ultimately leading to mitochondrial autophagy. FBXO7 regulates parkin translocation to mitochondria
and parkin-mediated ubiquitination of mitofusin-1, resulting in
mitophagy (13). A protein encoded by the PD-related leucinerich repeat kinase 2 (LRRK2) gene likely plays a role in the late
stages of autophagy by altering overall cellular autophagy flux
26
through modulation of lysosome acidification (18). These studies
linked abnormalities in autophagy to PD pathogenesis, but did
not show a causal relationship. However, two independent studies recently demonstrated that inactivation of autophagy-related
protein 7 in mouse brain resulted in loss of dopaminergic neurons, formation of ubiquitinated protein aggregates, and accumulation of α-synuclein and LRRK2 proteins (19, 20), strongly suggesting an important role for autophagy in PD pathogenesis. It
remains to be seen whether mitophagy impairment results in the
accumulation of damaged mitochondria in DA neurons in these
mouse models. It will also be important to demonstrate in vivo
that PD-gene–regulated autophagy is involved in the pathological changes seen in the disease.
Evidence has accumulated from in vitro studies indicating that
both UPS and ALP protein degradation mechanisms play critical
roles in PD pathogenesis. In vivo evidence shows that UPS and
ALP components are associated with PD pathology, providing
additional strong support. The two pathways likely functionally
interact and are under tight regulatory control in neurons. The
actual mode of action of UPS and ALP in PD pathogenesis in
vivo remains to be elucidated.
REFERENCES
1. D. W. Dickson, Cold Spring Harb. Perspect. Med. 2, 1 (2012).
2. C. E. Krebs et al., Hum. Mutat. 34, 1200 (2013).
3. M. Quadri et al., Hum. Mutat. 34, 1208 (2013).
4. A. B. Singleton, M. J. Farrer, V. Bonifati, Mov. Disord. 28, 14 (2013).
5. M. Jucker, L. C. Walker, Ann. Neurol. 70, 532 (2011).
6. Y. Imai, M. Soda, R. Takahashi, J. Biol. Chem. 275, 35661 (2000).
7. H. Shimura et al., Nat. Genet. 25, 302 (2000).
8. H. Xiong et al., J. Clin. Invest. 119, 650 (2009).
9. Y. Zhang et al., Proc. Natl. Acad. Sci. U.S.A. 97, 13354 (2000).
10. Y. Liu, L. Fallon, H. A. Lashuel, Z. Liu, P. T. Lansbury, Jr., Cell 111, 209 (2002).
11. S. Shojaee et al., Am. J. Hum. Genet. 82, 1375 (2008).
12. H. Laman et al., EMBO J. 24, 3104 (2005).
13. V. S. Burchell et al., Nat. Neurosci. 16, 1257 (2013).
14. Y. S. Choo et al., Hum. Mol. Genet. 21, 2514 (2012).
15. M. A. Lynch-Day, K. Mao, K. Wang, M. Zhao, D. J. Klionsky, Cold Spring Harb. Perspect. Med. 2, a009357 (2012).
16. D. Narendra, A. Tanaka, D. F. Suen, R. J. Youle, J. Cell Biol. 183, 795 (2008).
17. D. P. Narendra et al., PLOS Biol. 8, e1000298 (2010).
18. P. Gomez-Suaga et al., Hum. Mol. Genet. 21, 511 (2012).
19. I. Ahmed et al., J. Neurosci. 32, 16503 (2012).
Animal Models of Huntington’s Disease
Xiao-Jiang Li1,2* and Shihua Li1
A variety of transgenic animal models have been established for
uncovering the pathogenesis and therapeutic targets of Huntington’s disease (HD). However, most genetic mouse models lack
the striking neurodegeneration found in the brains of HD victims.
Large animal models of HD may demonstrate pathological features that are more similar to those in patients and thereby help
uncover more effective therapeutic targets. This review focuses
on the differential neuropathology seen in mouse and large animal models of HD and discusses the use of large animal models
to investigate the pathogenesis of the disease.
Introduction
Alzheimer’s disease (AD), Parkinson’s disease (PD), Huntington’s disease (HD), and several other neurodegenerative diseases share the common pathological feature of age-dependent,
selective neurodegeneration. These diseases are all caused by
misfolded proteins that form aggregates in the brains of affected
patients. Consistently, the age-dependent formation of these aggregates in patient brains is correlated with the progression of
the neurological symptoms of AD, PD, and HD.
HD shares common pathological changes with other neurodegenerative diseases, but is caused by an autosomal dominant
genetic mutation, making it an ideal model to study how protein
misfolding leads to selective neurodegeneration. The mechanistic insight obtained from studying HD should be very helpful for
understanding other neurodegenerative diseases resulting from
protein misfolding.
HD and Polyglutamine Expansion
HD is characterized by motor dysfunction, cognitive decline, and
psychological dysfunction. HD displays selective neurodegeneration that occurs preferentially in the brain striatum. The genetic
cause of HD is the expansion of a CAG repeat (>36 CAGs) in
exon 1 of the HD gene (HTT). The CAG repeat expansion results
in an expanded polyglutamine (polyQ) tract in the N-terminal region of the large (3,144 amino acid) huntingtin protein (Htt) (1, 2).
While the primary function of Htt has yet to be determined, it is
20. L. G. Friedman et al., J. Neurosci. 32, 7585 (2012).
Acknowledgments: The study was supported by grants from the National
Basic Research Program of China (2011CB510002), the Natural Science
Foundation of China (81161120498 and 31330031), and the Program of
Introducing Talents of Discipline to Universities in China (B13036).
1
Department of Human Genetics, Emory University School of
Medicine, Atlanta, GA
2
Institute of Genetics and Developmental Biology, Chinese
Academy of Sciences, Beijing, China
*
Corresponding Author: [email protected]
known to be essential for early development and probably plays
a role in cellular trafficking as a scaffold protein (1, 2). It is evident
that only N-terminal mutant fragments of Htt are able to form
aggregates and are in fact more toxic than full-length mutant Htt
(mHtt) (3, 4). This has prompted a search for proteolytic cleavage
sites mediated by various proteases, including calpains, aspartyl
proteases, and caspases capable of cleaving mHtt to generate
N-terminal mHtt fragments (5). These fragments can accumulate
in the nucleus, whereas the majority of full-length mHtt remains
in the cytoplasm (3, 6). The nuclear localization of N-terminal
mHtt can lead to abnormal binding of the fragments to various
transcription factors, affecting gene expression (1, 2).
Mouse Models of HD
Various HD mouse models have been established, including
transgenic mice (R6/2, N171-82Q) expressing N-terminal mHtt
(7, 8), full-length mHtt transgenic mice (using YACs and BACs)
(9, 10), and HD repeat knock-in (KI) mice (11–13). Characterization of various HD mouse models provides clear evidence that
small N-terminal Htt fragments with expanded polyQ tracts become misfolded and form aggregates. Despite the milder phenotypes of HD mice that express full-length mHtt, such as YAC128
and HD KI mice, these mice show accumulation of mHtt in the
striatum, consistent with the preferential loss of the medium spiny
neurons in the striatum of HD patients. Thus, the accumulation
of mHtt in neuronal cells is clearly a prerequisite for neuronal
dysfunction and degeneration.
Since mHtt is ubiquitously expressed in various types of cells
including neurons and male germ cells, and localized in different
subcellular regions including the nucleus and synapses, it seems
important to investigate how mHtt localization might contribute
to disease progression. Expression of N-terminal mHtt in astrocytes is sufficient to cause age-dependent neurological symptoms (14). Recently, transgenic mice that selectively expressed
N-terminal mHtt in presynaptic terminals were established and
also showed severe neurological phenotypes, providing convincing evidence for the critical role of synaptic mHtt in the disease
(15).
Although HD mouse models have been used widely to uncover the pathogenesis of HD and to develop treatments, most
of these mouse models do not show overt neurodegeneration.
Similarly, in other polyQ expansion mouse models, the lack of
striking neurodegeneration is a noteworthy phenomenon. These
facts highlight the fact that, although mouse models are used
27
NEURODEGENERATION RESEARCH
Produced by the Science/AAAS Custom Publishing Office
widely to investigate the pathogenesis of HD and other neurological diseases, they have their limitations and do not replicate
the full range of neurological phenotypes seen in related human
diseases. Such limitations reflect the importance of species differences in the etiology of neurodegeneration.
Neuropathology of Large Transgenic Animal
Models of HD
The insertion of a foreign gene into the monkey genome in 2001
(16) demonstrated that nonhuman primate models expressing
transgenes was possible, including a rhesus monkey expressing exon 1 mHtt with an 84 unit polyQ tract under the control of
the human ubiquitin promoter (17–19). Unlike transgenic mice,
which can survive after birth when expressing the same exon
1 mHtt with an even longer polyQ repeat (150Q) (7, 20), HD
transgenic monkeys with 84Q died postnatally and this early
death was directly related to the mHtt transgene (17). Despite
their early death, some transgenic monkeys developed key clinical HD features including dystonia, chorea, and seizure (17).
Like the brains of HD mouse models and patients, the HD monkey brains also show abundant Htt aggregates in neuronal nuclei and processes. More importantly, transgenic HD monkeys
display the degeneration of axons and neuronal processes
in the absence of obvious cell body degeneration, suggesting
that neuronal degeneration in HD may initiate at the neuronal
processes.
In collaboration with Dr. Liangxue Lai at the Chinese Academy of Sciences’ Guangzhou Institutes of Biomedicine and
Health, we created transgenic HD pigs that expressed N-terminal mHtt consisting of the first 208 amino acids and including 105Q (N208-105Q) (21). Like transgenic monkey models of
HD, most of these transgenic piglets died postnatally, with some
showing a severe chorea phenotype before death, suggesting
that mHtt is more toxic to larger animals than to mice. More importantly, in all transgenic pig brains examined, apoptotic cells
were seen (21), something that had not been reported in any HD
mouse models.
In contrast to the above results, transgenic HD sheep expressing full-length mHtt with a 73Q tract live normally and show
only a decrease in the expression of the medium spiny neuron
marker DARPP-32 (22). Thus, as with HD mouse models, the
expression of N-terminal mHtt can result in robust neurological
phenotypes and pathological changes in large animals. These
studies also suggest that the protein context and the length of
Htt fragments may determine the nature of the neuropathology.
It is possible that neurodegeneration in large animals only occurs when a sufficient level of degraded N-terminal fragments
have accumulated in old animals. Thus, expressing N-terminal
mHtt fragments alone can speed the progression of the disease, resulting in the early postnatal death in transgenic HD pigs
and monkeys.
28
Insights from Large Transgenic Animal
Models of HD
It is clear that species differences play a critical role in the neurological phenotype of HD in small and large animal models. There
are considerable differences in development, life span, physiology, genetics, and anatomy between small and large animals.
Of greater interest is elucidating the mechanisms behind these
differences. The short lifespan of mice is often believed to be
responsible for the failure of HD mouse models to develop overt
neurodegeneration. It is also possible that the misfolded form of
N-terminal mHtt is more toxic to the neuronal cells of pigs and
monkeys than to rodent neurons. Also, because the brain circuitry in pigs and monkeys is more complex than in mice, this
may render neurons in large animals more vulnerable to misfolded mHtt. Finally, the cellular ability to cope with misfolded
proteins during development and adulthood may differ between
species and the rapid maturation of rodent neurons during early brain development may reduce their sensitivity to misfolded
proteins, potentially explaining why mouse models survive to
adulthood even when they express the same mHtt N-terminal
fragment.
Future Studies with Large Animal Models
The generation of genetic models using large animals is significantly more challenging than with rodents. Most work done
on transgenic monkeys involved the use of lentiviral vector infection of fertilized oocytes followed by embryo transplantation
(17–19), requiring a considerable number of donor and surrogate monkeys. Successful generation of transgenic pigs can
be achieved via nuclear transfer, a cloning strategy that has a
low success rate (<2%) for transferred pig embryos developing
to term (23). In addition, the costs of maintaining and breeding
large animals, as well as the ethical concerns and strict regulations for use, also make biomedical research with these animals
more difficult.
Recent advances in gene targeting have opened new avenues
for generating large animal models. New technology using bacterially derived transcription activator-like effector nucleases
(TALENs) and clusters of regularly interspaced short palindromic
repeats (CRISPR)/cas9 enable very specific gene targeting,
even eliminating the need for using embryonic stem cells (24,
25). This approach can be applied to one-cell fertilized embryos
to generate null mutations in specific genes.
Although rodent models of neurodegenerative diseases will
remain an important modeling system for investigating a variety of diseases, in large part because of the difficulty and expense of generating and characterizing large animals, large
animal models provide a more rigorous system for validating
the relevance of critical findings from small animal models.
Transgenic models using higher mammalian species to model
important neurodegenerative diseases will give us deeper in-
sight into the pathogenesis of neurodegenerative diseases.
In addition, given the frequent failures when it comes to clinical trials of drugs based on successful testing in small animal models, large transgenic animals could yield a more reliable system for verifying therapeutic efficacy before moving to
clinical trials.
11. V. C. Wheeler et al., Hum. Mol. Genet. 9, 503 (2000).
12. C. H. Lin et al., Hum. Mol. Genet. 10, 137 (2001).
13. L. B. Menalled et al., J. Neurosci. 22, 8266 (2002).
14. J. Bradford et al., Proc. Natl. Acad. Sci. U.S.A. 106, 22480 (2009).
15. Q. Q. Xu et al., J. Cell Biol. 202, 1123 (2003).
16. A. W. Chan et al., Science 291, 309 (2001).
REFERENCES
1. P. Harjes, E. E. Wanker, Trends. Biochem. Sci. 28, 425 (2003).
17. S. H. Yang et al., Nature 453, 921 (2008).
18. E. Sasaki et al., Nature 459, 523 (2009).
19. Y. Niu et al., Proc. Natl. Acad. Sci. U.S.A. 107, 17663 (2010).
2. S. H. Li, X. J. Li. Trends Genet. 20, 146 (2004).
20. P. H. Cheng et al., Am. J. Hum. Genet. 93, 306 (2013).
3. C. A. Gutekunst et al., J. Neurosci. 19, 2522 (1999).
21. D. Yang et al., Hum. Mol. Genet. 19, 3983 (2010).
4. H. Zhou et al., J. Cell. Biol. 163, 109 (2003).
22. J. C. Jacobsen et al., Hum. Mol. Genet. 19, 1873 (2010).
5. Z. H. Qin, Z. L. Gu, Acta Pharmacol. Sin. 25, 1243 23. L. Lai, R. S. Prather, Cloning Stem Cells 5, 233 (2003).
(2004).
6. M. DiFiglia et al., Science 277, 1990 (1997).
24. B. Shen et al., Cell Res. 23, 720 (2013).
25. H. Wang et al., Cell 153, 910 (2013).
7. S. W. Davies et al., Cell 90, 537 (1997).
8. G. Schilling et al., Hum. Mol. Genet. 8, 397 (1999).
Acknowledgments: This work was supported by the U.S. National
9. E. J. Slow et al., Hum. Mol. Genet. 12, 1555 (2003).
Institutes of Health grants NS036232, AG019206, and NS041669 for
10. M. Gray et al., J. Neurosci. 28, 6182 (2008).
X.J.L. and AG031153 for S.H.L.
FMRP Regulates Microtubule Network
Formation, Neurogenesis, and DNA Damage Response
Aiyu Yao and Yong Q. Zhang*
Fragile X syndrome (FXS) is the most common form of inherited mental retardation worldwide (1) and is caused by the loss
of function of the Fragile X mental retardation protein (FMRP;
encoded by the FMR1 gene), a multi-domain RNA-binding protein believed to regulate synaptic development and plasticity by
controlling the localization and translation of specific mRNA targets. Though the gene mutated in FXS was identified over two
decades ago, exciting new discoveries on the functions of FMRP
continue to emerge from in vivo studies. In this review, important
advances in our current understanding of FMRP’s role in microtubule regulation, neurogenesis, and DNA damage response are
discussed. The reader is referred to two excellent reviews for
general information on the neuronal functions of FMRP (1) and
the pathophysiology of FXS (2).
FMRP Regulates Microtubule Formation
The first line of evidence implicating FMRP in microtubule regulation came from the discovery that FMRP binds and suppresses
the translation of mRNA for the microtubule-stabilizing microtubule-associated protein 1B (MAP1B) (3, 4) and its homolog Futsch in Drosophila (5). In the neonatal mouse brain, FMRP suppresses the expression of MAP1B during synaptogenesis, and
the elevated expression of MAP1B in FMR1 knockout mice leads
to increased microtubule stability (6). Together, these data establish a key role for FMRP in microtubule regulation. However,
direct in vivo evidence linking the loss of FMRP to microtubule
abnormalities has been scarce.
Spastin was identified through a genetic screen in Drosophila
as a dominant suppressor of the rough eye phenotype caused
by dfmr1 (the Drosophila homolog of FMR1) over-expression (7).
Spastin encodes a microtubule-severing protein, and a mutation
in this gene can lead to neurodegenerative hereditary spastic
paraplegia. Muscle cells mutant for dfmr1 have more perinuclear
but fewer distal microtubules, and the distal microtubules have
short, bent, and tangled fibers. Conversely, dfmr1 overexpression produces thick, parallel microtubule bundles, along with in-
29
NEURODEGENERATION RESEARCH
Produced by the Science/AAAS Custom Publishing Office
creased levels of α-tubulin protein and acetylated
microtubules (7). Genetic analysis revealed that
dfmr1 acts upstream of or in parallel with spastin in multiple processes, including neuromuscular junction growth and microtubule formation.
The mechanism by which dFMRP regulates the
microtubule network remains unclear. However,
based on the observed genetic interaction between
dfmr1 and spastin (7), the two proteins may regulate microtubules through a similar mechanism
(Fig. 1).
A recent study confirmed the regulation of microtubules by dFMRP at the neuromuscular junction,
and further demonstrated that the FMRP-MAP1B/
Futsch microtubule pathway is regulated by bone
morphogenetic protein signaling (8).
Figure 1. Fragile X mental retardation protein (FMRP) regulates microtubules to control synapse
growth and axonal transport by suppressing the expression of the microtubule-associated protein,
MAP1B/Futsch. FMRP also regulates synapse growth in parallel with the microtubule-severing protein, spastin, mutations of which lead to hereditary spastic paraplegia.
FMRP Regulates Neurogenesis
Figure 2. In mammals (Mamm.), FMRP regulates proliferation and fate specification of an adult
A significant recent discovery is the involvement neural progenitor cell (aNPC) by inhibiting the expression of cyclin D1 (CycD1) and the Wnt signaling pathway component glycogen synthase kinase 3 β (GSK3β), respectively. In Drosophila (Dros.),
of FMRP in adult neurogenesis. In FMRP-deficient dFMRP controls proliferation of larval neuroblasts and the female germline by suppressing the exmice, both proliferation and fate specification of pression of cyclin E (CycE). dFMRP also suppresses cyclin B (CycB) expression to regulate cell
cycle progression in embryos and to control DNA damage response in other stages of the life cycle.
adult neural progenitor cells (aNPCs) were altered
in the subgranular zone of the dentate gyrus of the
hippocampus through dysregulation of the Wnt/β-catenin signal- eration during larval brain development (11). Loss of dfmr1 leads
ing pathway. A reduction in Wnt signaling activity changed the to a significant increase in the number of mitotic neuroblasts
expression of neurogenin 1, and consequently, aNPC fate speci- (NBs) in the early larval brain. dFMRP also inhibits neuroblast
fication (9). It was later shown that targeted ablation of FMR1 exit from quiescence in larval brains, as evidenced by the aberin aNPCs via inducible gene recombination decreased hippo- rant expression of cyclin E, a critical regulator of the G1/S trancampal neurogenesis, and significantly impaired hippocampus- sition of the cell cycle. Clonal analyses in the developing brain
dependent learning in mice, while restoring FMR1 expression further revealed that dfmr1-deficient NBs generated significantly
rescued the learning deficits (10). Based on these results, the more neuronal progeny than wild-type NBs (11). Consistent with
authors proposed that defective adult neurogenesis contributes an essential role in a wide range of developmental processes,
to the learning impairment exhibited by FXS patients, and that dFMRP also regulates germline proliferation and maintenance,
these can be alleviated by a restoration of FMRP function spe- as well as cell cycle progression during oogenesis (14, 15). It is
cifically in aNPCs, suggesting new therapeutic strategies for the unclear whether FMRP has similar functions in the mammalian
germline.
treatment of FXS.
In addition to adult neurogenesis, FMRP also regulates embryonic and post-embryonic neurogenesis in mammals and Dro- FMRP Regulates DNA Damage Response by
sophila (11–13). Intriguingly, studies using mammalian embry- Suppressing Cyclin B Expression
onic NPCs versus aNPCs reveal opposite roles for FMRP. In the The changes in differentiation and proliferation of germline cells
embryonic brain, loss of FMRP leads to the overproduction of and NPCs in both mammals and flies resulting from (d)FMRP
neurons in the cortex (12, 13), whereas in aNPCs, fewer neurons deficiency point to an important role for FMRP in cell cycle regand more glia are generated in the hippocampus (9, 10), imply- ulation. The molecular pathways for cell cycle control, together
ing that the role of FMRP in mammalian neurogenesis is context- with those involved in DNA repair and apoptosis, are activated
specific. In Drosophila, dFMRP is required to control NPC prolif- in response to DNA damage (16). It was observed that dfmr1
mutants were hypersensitive to DNA damage (17). Loss of
dfmr1 compromised the G2/M DNA damage checkpoint, leading to increased mitotic activity accompanied by elevated levels
Key Laboratory of Molecular and Developmental Biology,
of cyclin B (CycB), a major regulator of the G2/M transition. In
Institute of Genetics and Developmental Biology, Chinese
addition, dfmr1 mutants had excessive apoptosis, more DNA
Academy of Sciences, Beijing, China
*
Corresponding Author: [email protected]
strand breaks, and greater genomic instability. Taken together,
30
these data strongly argue that dFMRP regulates the DNA damage response by suppressing CycB expression and inhibiting
apoptosis (17).
At present, it is unknown whether FXS patients and FMR1
knockout mice also have a defective DNA repair pathway. In
Drosophila, there is a single FMRP homolog, compared to
three in mammals. It is thus possible that functional redundancy
between these three family members can mask the phenotype of
a single mutant. It would be interesting to test whether double or
triple mutants of FMR1 exhibit DNA damage hypersensitivity, as
in the case of dfmr1 mutants. If confirmed, it would suggest that
the DNA repair pathway could be exploited as a new target for
intervention in FXS pathogenesis.
Translational Implications
FXS patients display a range of neurobehavioral abnormalities
such as learning and memory deficits, attention deficit, hyperactivity, sleep disturbance, and autistic features. FMRP participates
in many processes, including microtubule network regulation,
neurogenesis, and DNA damage response, which are cellular
context-dependent and developmental stage-specific (Fig. 2). At
present, there is no unifying mechanism that connects the diverse in vivo functions of FMRP. Drugs targeting specific FMRPmediated pathways are currently in clinical trials, but effective
interventions for FXS may require a combinatorial approach that
targets multiple pathways, rather than a single pathway regulated by FMRP.
13. T. A. Tervonen et al., Neurobiol. Dis. 33, 250 (2009).
REFERENCES
1. O. Penagarikano, J. G. Mulle, S. T. Warren, Annu. Rev. Genomics Hum. Genet. 8, 109 (2007).
2. C. L. Gatto, K. Broadie, Curr. Opin. Neurobiol. 21, 834 (2011).
3. V. Brown et al., Cell 107, 477 (2001).
4. J. C. Darnell et al., Cell 107, 489 (2001).
5. Y. Q. Zhang et al., Cell 107, 591 (2001).
6. R. Lu et al., Proc. Natl. Acad. Sci. U.S.A. 101, 15201 (2004).
7. A. Yao et al., Hum. Mol. Genet. 20, 51 (2011).
8. M. Nahm et al., Neuron 77, 680 (2013).
9. Y. Luo et al., PLOS Genet. 6, e1000898 (2010).
10. W. Guo et al., Nat. Med. 17, 559 (2011).
11. M. A. Callan et al., Hum. Mol. Genet. 19, 3068 (2010).
12. M. Castren et al., Proc. Natl. Acad. Sci. U.S.A. 102, 17834 (2005).
14. A. M. Epstein, C. R. Bauer, A. Ho, G. Bosco, D. C. Zarnescu, Dev. Biol. 330, 83 (2009).
15. Y. Yang et al., PLOS Genet. 5, e1000444 (2009).
16. J. W. Harper, S. J. Elledge, Mol. Cell. 28, 739 (2007).
17. W. Liu, F. Jiang, X. Bi, Y. Q. Zhang, Hum. Mol. Genet. 21, 4655 (2012).
Acknowledgments: This work was supported by grants from the Chinese
Academy of Sciences Strategic Project B (XDB02020400 and KSCX1YW-R-69) and the National Science Foundation of China (NSFC,
30430250 and 30525015) to Y. Q. Z. and NSFC (31271121) to A. Y.
31
GENETIC AND CLINICAL STUDIES
Produced by the Science/AAAS Custom Publishing Office
Progress in Treating Hereditary Ataxia in Mainland China
Hong Jiang1,2,3, Junling Wang1, Juan Du1, Ranhui Duan2, Jiada Li2, and Beisha Tang1,2,3,*
Hereditary ataxia (HA) encompasses a group
of high heterogeneous neurodegenerative
disorders. Genetically, HA can be classified
based on the mode of inheritance: autosomally
dominant, such as autosomally dominant
cerebellar ataxia (ADCA) or spinocerebellar
ataxia (SCA); autosomally recessive, including autosomally recessive cerebellar ataxia
(ARCA); X-linked ataxia; or mitochondrial
ataxia. Through our research, we have developed new technologies and strategies for diagnosis and treatment of HA, identified a novel
causative gene for SCA, and tested an experimental treatment for spinocerebellar ataxia type
3/Machado-Joseph disease (SCA3/MJD). Here
we provide insight into some of these findings.
Figure 2. (A) Partial pedigree of families CS and LY, brain magnetic resonance imaging of proband of family CS, and mapped region on chromosome 20.
(B) Heterozygous missense mutation, genomic organization of the human transglutaminase 6 (TGM6) gene, and domain structure of the TGM6 protein.
Figure 1. Frequencies of SCAs in mainland China. SCA3/MJD (62.09%) is the most common subtype
in patients with ADCA and sporadic forms, followed by SCA2 (6.28%), SCA1 (5.81%), SCA6 (1.86%),
SCA7 (1.86%), SCA35 (0.47%), SCA12 (0.23%), and SCA17 (0.23%). The remainder are the genetically
unidentified cases (21.4%).
Diagnosis of Hereditary Ataxia
Genetic testing is the gold standard for a definitive diagnosis
of HA and next generation sequencing (NGS), which is highthroughput, functional, and cost-effective, offers a commercially available diagnostic strategy (1). To date, classical genetic
studies have revealed 34 distinct genetic forms of ADCAs (including SCAs), but only 22 causative genes have been identified.
In mainland China, SCA3/MJD is by far the most common
subtype in patients with ADCA and sporadic (non-genetically inherited) forms, followed by SCA2, SCA1, SCA6, and SCA7. Subtypes SCA12, SCA17, SCA35, and dentatorubral-pallidoluysian
atrophy are seldom found, and SCA5, SCA8, SCA10, SCA11,
SCA13, SCA27, and SCA31 are very rare (Fig. 1). Genetically uncharacterized cases suggest that other disease-causing
genes are involved in negative pedigrees. Additionally, etiological
factors other than genetics may be involved in sporadic cases
(2–5). Although SCAs caused by abnormal stretches of repetitive
sequence cannot be accurately diagnosed by NGS, it is still a
useful strategy for gaining new insights into these disorders (6).
For example, in 2006 we identified two novel ARCA mutations in
the ataxia-telangiectasia (AT) gene, designated as ATM, by direct
Department of Neurology, Xiangya Hospital, Central South
University, Changsha, Hunan, China
2
State Key Laboratory of Medical Genetics, Changsha, Hunan, China
3
Hunan Province Key Laboratory of Neurodegenerative Disorders,
Changsha, Hunan, China
*
Corresponding Author: [email protected]
1
32
DNA sequencing (7). Further, the recent discovery of six new
mutations in the ATM gene and two in the SACS gene by NGS
provides evidence that this technology shows great promise for
improving the diagnosis of inherited neurological disorders (1).
Although X-linked ataxia and mitochondrial ataxia are rare in
mainland China, the use of conventional sequencing and NGS
is beneficial for obtaining a definitive genetic diagnosis. Hence,
we proposed a genetic diagnostic strategy relying on the NGS
platform. For diseases with relatively low genetic complexity and
clear clinical manifestations, disease-specific NGS assays are
attractive because good sequencing depth can be achieved at
a relatively low cost (1). For diseases with genetic and phenotypic heterogeneity and/or complexity, making diagnosis difficult,
whole exome sequencing (WES) is a convenient approach to
confirm known causative variations or uncover novel candidate
mutations, providing a clearer genetic diagnosis and enabling
disease-risk counseling (1).
Novel Causative Gene for Autosomal Dominant
Cerebellar Ataxia
SCA type 35 (SCA35) is an autosomal dominant neurodegenerative disorder. We identified transglutaminase 6
(TGM6) as the gene mutated in SCA35, using a combination
of techniques including linkage analysis, exome sequencing,
and copy number variation analysis. WES analysis of a family diagnosed with SCA35 identified a mutation located in the
same region as that validated by linkage analysis (8). We further
confirmed the finding by identifying a second novel mutation in
Figure 3. Overexpression of TGM6 increased the formation of polyglutamine
(polyQ) aggregates. HEK293 cells were co-transfected with plasmids that expressed enhanced green fluorescent proten-tagged polyQ and Myc-tagged
TGM6, including wild-type (WT) TGM6, and mutant TGM6 (D327G, L517W). The
percentages of aggregate-positive cells indicated statistically significant differences compared with empty vector controls.
a different SCA35 family that cosegregated with the phenotype.
Both mutations, predicted to be functionally damaging, were
in a highly conserved domain of the TGM6 gene and absent
in 500 normal, unaffected individuals. The finding of TGM6 as
a novel causative gene of spinocerebellar ataxia illustrates a
promising, highly accurate, and affordable method that will provide new insights into the discovery of novel causative genes.
The precise role of TGM6 in SCA35 remains unclear. Recently,
we demonstrated that TGM6 protein was abundantly expressed
in the septal region, basal ganglia, hypothalamus, and brainstem
(9). We found that both mutants of TGM6 exhibited decreased
transglutaminase activity and decreased stability (10).
Furthermore, overexpression of TGM6 mutants sensitized cells
to staurosporine-induced apoptosis by increasing the activity of
the caspases (10). A common feature of polyglutamine (polyQ)
diseases is the presence of polyQ protein aggregates in neuronal
cells caused by expanded tracts glutamine residues (11). PolyQ
proteins are substrates of transglutaminase 2 (TGM2), and the
increased activity of TGM2 in polyQ diseases suggests that it
may be directly involved in the formation of the aggregates (12).
We recently found that TGM6 interacts and colocalizes with
both normal and expanded polyQ proteins in HEK293 cells (13).
Moreover, the overexpression of TGM6 promotes the formation
of polyQ aggregates and the conversion of soluble polyQ into
insoluble polyQ aggregates (Fig. 3). However, the mutations
associated with SCA35 should not affect their binding to polyQ
proteins (13). Thus, our study suggest that TGM6 could be
involved in polyQ diseases, but whether there exists a common
pathological link between polyQ-associated SCA and SCA35 will
require further investigation.
Experimental Treatment of SCA
In previous studies, we used several SCA3/MJD Drosophila
models through which we determined that heat shock protein 22
(Hsp22), lithium chloride, and valproic acid (VPA) are potential
therapeutic agents for the treatment of SCA3/MJD. Hsp22, a
member of the small heat shock protein (sHsp) family, plays a
significant role as chaperone. In the SCA3 Drosophila model
33
GENETIC AND CLINICAL STUDIES
Produced by the Science/AAAS Custom Publishing Office
Genetic Etiology of Parkinson’s Disease in China
Figure 4. Gmr-SCA3tr-78Q is a strain of transgenic Drosophila that expresses expanded polyQ tracts in fly compound eyes. Overexpression of Hsp22, lithium chloride, and
valproic acid (VPA) prevents polyQ-induced eye depigmentation in an SCA3 Drosophila model. Paired dissecting microscope (Magnification: A, ×65; B, ×115; C, ×80) and scanning
electron microscope (×1000) images of adult fly eyes are
shown. (A) Hsp22, (B) VPA, (C) lithium chloride.
(gmr-GAL4/+; elav-GAL4/+; UAS-MJDtr-Q78/+;
UAS-HSP22/+), expression of the MJDtr-Q78
transgene—containing an expanded polyglutamine tract—showed a greater loss of cell integrity, and the pigmentation of adult flies was faded
and showed black, point-like necrosis. Drosophila co-expressing an expanded polyQ protein
and either one or two copies of the HSP22 gene
underwent heat shock, resulting in induction of
expression of the HSP22 gene to differing levels
depending on the number of copies introduced. The strains carrying the HSP22 gene were identical, except for the number of
gene copies. Findings showed that Hsp22 expression impacted
eye depigmentation (Fig. 4), growth restriction, ability for eclosion, and average lifespan (14). Previous studies have suggested that an imbalance in histone acetylation may play a key role
in transcriptional dysregulation in polyQ diseases. We examined
the effect of VPA, a promising therapeutic HDAC inhibitor, in a
Drosophila SCA3 model (elav-MJDtr-Q78; gmr-MJDtr-Q78). We
expressed the MJDtr-Q78 transgene both in the developing eyes
and in neurons. Expression of the MJDtr-Q78 protein produced
deleterious phenotypes including faded eye pigmentation, impaired climbing ability, and decreased mean lifespan, similar to
the characteristics of human SCA3. To test the therapeutic potential of VPA in vivo, a series of daily doses of VPA were administered to SCA3 flies before cross-breeding. Results showed that
long-term use of VPA at an optimal dose partly prevented eye depigmentation (Fig. 4), alleviated climbing disability, and extended
the average lifespan of SCA3/MJD transgenic flies (15).
Lithium (Li) is known to be neuroprotective in various models
of neurodegenerative disease and can reduce mutant protein aggregation by inducing autophagy. In an SCA3 Drosophila model
(gmr-MJDtr-Q78; nrv2-MJDtr-Q78), expression of MJDtr-78Q
caused faded eye pigmentation, decreased locomotor ability,
and reduced lifespans in adult flies. These flies also showed
a late-onset, progressive, neurodegenerative phenotype with
similar features to SCA3/MJD patients. To study the effects of Li
treatment on SCA3 pathogenesis in vivo, we administered daily
doses of Li chloride (LiCl) at different levels to SCA3 flies prior
to cross-breeding. Results confirmed that chronic treatment with
LiCl prevented eye depigmentation (Fig. 4), alleviated locomotor
disability, and extended the average lifespan (16).
In summary, DNA-based technologies have provided new diagnostic strategies, the ability to identify causative genes, and
34
Jifeng Guo1,2,3, Xinxiang Yan2,3, Danling Wang1, Qian Xu2,3, Beisha Tang1,2,3, and Zhuohua Zhang1*
the opportunity to find new therapeutic interventions for HA, as
well as provided new insights in pathogenic mechanisms. In
the future, individuals suspected of suffering from HA can be
genetically evaluated to gain a quicker and more complete diagnosis as well as a personalized disease-risk profile and improved
genetic counseling.
REFERENCES
1. Z. Chen et al., Neurobiol. Aging 34, 2411 (2013).
2. J. Wang et al., Zhong Nan Da Xue Xue Bao Yi Xue Ban 36, 482 (2011).
3. B. Tang et al., Arch. Neurol. 57, 540 (2000).
4. H. Jiang et al., Chin. Med. J. (Engl.) 118, 837 (2005).
5. H. Jiang et al., J. Neurol. Sci. 236, 25 (2005).
6. A. Sailer et al., Neurology 79, 127 (2012).
7. H. Jiang et al., J. Neurol. Sci. 241, 1 (2006).
8. J. L. Wang et al., Brain 133, 3510 (2010).
9. Y. T. Liu et al., Anat. Rec. 296, 1576 (2013).
10. W. J. Guan et al., Biochem. Biophys. Res. Commun. 430, 780 (2013).
11. A. Michalik, C. Van Broeckhoven, Hum. Mol. Genet. 12 Spec. No. 2, R173 (2003).
12. C. A. Ross, M. A. Poirier, Nat. Med. 10 Suppl., S10 (2004).
13. W. J. Guan et al., Biochem. Biophys. Res. Commun. 437, 94 (2013).
14. Q.H. Li et al., Prog. Biochem. Biophys. 35, 1430 (2008).
15. J. Yi et al., PLOS ONE 8, e54792 (2013).
16. D. D. Jia et al., Cerebellum 12, 892 (2013).
Acknowledgments: These studies were supported by grants from the
Parkinson’s disease (PD) is generally considered to be a multifactorial neurodegenerative movement disorder commonly characterized by bradykinesia, resting tremor, rigidity, and postural
instability. Approximately 10% to 15% of patients with PD have
a family history of the disease, but the majority of PD is sporadic
in origin. Although the pathogenesis of PD remains unclear, it
is generally accepted that genetic factors, environmental insult,
and gene-environment interactions contribute to PD etiology. Exploration the genetics of PD may provide significant insights into
the disease mechanism, particularly since the SNCA gene, coding for α-synuclein, was reported to be a causative factor in PD.
However, the genetic heterogeneity inherent in PD means that
there are still much work to be done understanding the genetic
basis of this disease.
Autosomal Dominant PD
The SNCA (PARK1/4), LRRK2 (PARK8), UCHL1 (PARK5),
VPS35, and EIF4G1 genes have been identified as being associated with autosomal dominant PD (ADPD). However, only
rarely have mutations in these genes been found in Chinese familial ADPD cases, suggesting that they are not the main genetic
cause of PD in China (1–3).
Wang et al. (4, 5) reported identification of two novel variants
in Chinese familial PD: c.A1847G (p.K616R) in a conserved
residue within a domain of unknown function in the leucinerich repeat kinase 2 gene (LRRK2), and c.G77A (p.G26E) in
a non-conserved position in the high temperature requirement
protein A (HtrA) serine peptidase 2 (HtrA2) gene. However,
other family members have not yet been tested and cosegregation analysis was not performed. Further genetic and functional studies are therefore required to verify whether these
two variants are causative. Similarly, no pathogenic mutations
have been found in Chinese families with ADPD in the growth
factor receptor-bound protein 10 (GRB10) interacting GYF protein 2 (GIGYF2) gene. However, eight heterozygous and one
homozygous sequence variants were identified in sporadic PD
that do not present in 300 control individuals (6). Nevertheless,
polymorphic alleles p.Ala99Ala and p.Pro460Thr in GIGYF2
Major State Basic Research Development Program of China (“973”
Program) (2012CB944601, 2012CB517902, and 2011CB510002), a New
1
Century Excellent Talents in University grant from the Department of
2
Education (NCET-10-0836), and the National Natural Science Foundation
of China (81130021, 81271260, 30971585, 30871354, 30710303061, and
30400262).
State Key Laboratory of Medical Genetics, Changsha, Hunan, China
Department of Neurology, Xiangya Hospital, Central South University, Changsha, Hunan, China
3
Hunan Province Key Laboratory of Neurodegenerative Disorders,
Changsha, Hunan, China
*
Corresponding Author: [email protected]
have been suggested to be associated with increased risk of
sporadic PD (7).
CAG triplet repeat expansions in the SCA3 gene have been
found in patients with PD. In China, mutations in SCA2, SCA3,
and SCA17 are likely linked to both familial and sporadic forms
of PD (8, 9). Patients carrying these mutations show typical PD
features with a bilateral reduction in uptake of a dopamine transporter tracer in PET scans. Moreover, some had mild symptoms
of ataxia. The mutation frequency in SCA2 and SCA3 is 1.5%
and 3%, respectively, in familial cases, and 0.5% and 0.8%,
respectively, in sporadic cases (8). Thus, it is suggested that
SCA2, SCA3, and SCA17 be screened for mutations in order to
diagnose ADPD in the ethnic Chinese population.
Autosomal Recessive PD
Mutations in parkin (PARK2), PINK1 (PARK6), DJ-1 (PARK7),
FBXO7 (PARK15), ATP13A2 (PARK9), and PLA2G6 (PARK14)
are known to be involved in autosomal recessive PD (ARPD).
In contrast to those genes known to be associated with ADPD,
mutations in these ARPD-associated genes are well-represented
in cases of both familial autosomal recessive early-onset Parkinsonism (AREP) (10, 11) and sporadic early onset Parkinsonism
(EOP) in Chinese families (12–14). Parkin mutations are the most
commonly identified genetic variations in Chinese patients with
AREP (48.3%) or EOP (12.6%). Consistent with findings in other
ethnic groups, exonic rearrangements within parkin are the most
common types of mutations in Chinese PD cases with AREP. Mutations in the PINK1, DJ-1, and PLA2G6 genes were also found
in cases of AREP (6.9%, 3.4%, 3.4%, respectively) and sporadic
EOP (3.1%, 2.4%, 0.79%, respectively). Thus far, no pathogenic
mutations have been detected in the ATP13A2 and FBXO7 genes
in the Chinese population. A list of PD-related genes is shown in
Figure 1.
Mutations in PLA2G6 at the PARK14 locus have previously been shown to be linked to infantile neuroaxonal dystrophy
(INAD), neurodegeneration with brain iron accumulation (NBIA),
and dystonia-parkinsonism. In the mainland Chinese population,
a homozygous PLA2G6 missense mutation c.G991T (p.D331Y)
was identified in PD patients in a consanguineous family (11)
(Fig. 2). In this family, the proband showed typical PD clinical
features with no extrapyramidal symptoms, while a PET study
indicated a substantial reduction in dopamine transporter (DAT)
binding in the brain. A slight reduction of DAT binding was also
found in his younger sister who showed no PD symptoms. Further studies indicated that the mutant protein had only 30% phos-
35
GENETIC AND CLINICAL STUDIES
Produced by the Science/AAAS Custom Publishing Office
Figure 1. PD-related Genes in Chinese
Population. (A) Frequency of mutations
in familial autosomal recessive early-onset Parkinsonism (AREP). (B) Frequency
of mutations in sporadic early onset Parkinsonism (EOP).
Figure 2. Family Carrying PLA2G6 Mutation (A) Pedigree of
consanguineous family with PLA2G6 p.D331Y mutation. [PET],
PET scan performed; m/m, homozygous mutation carrier; m/
wt, heterozygous mutation carrier; wt/wt, homozygous wildtype. (B) Representative results of enzyme activity assay from
three independent experiments. The mock group represented
endogenous phospholipases enzyme activity of 0.37nmol/
min/mg and was subtracted from each of the other groups. (C)
PET images of the family members with the p.D331Y PLA2G6
mutation. Proband (II-2) shows significiant reduction in dopamine transporter (DAT) binding. The homozygous mutation
carrier (II-4) shows slight reduction in DAT binding in the right
posterior putamen. The heterozygous D331Y mutation carriers
(I-2 and II-3) and the normal genotype member (II-1) show no
abnormalities in this assay.
pholipase enzyme activity comparing to its wild-type counterpart.
Based on this data, we proposed AREP without complex phenotypes as the fourth subtype of PLA2G6-related neurodegenerative disorders. Partial loss of PLA2G6 enzyme activity likely contributes to the PD pathogenesis in the Han Chinese population.
This is further supported by a novel PLA2G6 mutation found in a
case of sporadic EOP (14).
In contrast to monogenic diseases in which Mendelian inheritance can adequately explain pathogenesis, PD is a complex disorder, likely involving a combination of multiple genetic factors.
Identification of a family with EOP harboring heterozygous missense mutations in both PINK1 and DJ-1 genes indicates a digenic inheritance of PD, suggesting a new genetic mechanism for PD
pathogenesis (15).
Other Loci and SNPs in PD
Susceptibility genes play important roles in the pathogenesis of
sporadic PD in many different populations. The following genes
have been studied in the Han Chinese population.
36
SNCA
SNCA, which codes for α-synuclein, has been shown to modulate susceptibility of sporadic PD using both a candidate gene
approach and genome wide association studies. The single
nucleotide polymorphism (SNP) rs11931074, located in 3’ region
of SNCA, is hypothesized to modify expression of α-synuclein
in certain cases of sporadic PD in China (16); other SNPs in
SNCA have also been reported to contribute to sporadic
PD (17).
GBA
Mutations in the GBA (glucosidase, beta, acid) gene was first reported in consanguineous relatives with Gaucher disease (GD),
a lysosomal storage disorder. Heterozygous mutations GBA
have been linked to PD pathogenesis: in the Han Chinese population, the L444P mutation in GBA is linked with susceptibility to
sporadic PD, with an odds ratio of 24.29 (18). By contrast, the
F213I, R353W, and N370S mutations—the most frequent mutations in non-Jewish populations after L444P—are rare in Han
Chinese (18).
PITX3
The gene for pituitary homeobox 3 (PITX3) is reported to be associated with PD in several populations (19). The rs3758549
SNP within PITX3 has been associated with PD in China, especially late onset PD. Further meta-analysis of PITX3 has
provided strong evidence to support its role in susceptibility in
Chinese populations (19). However, evidence of the involvement
of rs2281983, rs4919621, and other SNPs in PITX3 in Chinese
patients has not been found.
Other Genes
A large number of susceptibility genes/loci related to sporadic
PD have been identified in various populations by GWAS, including LRRK2, GAK, BST1, and PARK16. Recently, largescale studies revealed that PARK16 (rs823128, rs823156, and
rs16856139), BST1 (rs11724635), GAK (rs1564282), and LRRK2
(G2385R and R1268P) were associated with increased risk of
developing sporadic PD in the Chinese population. Considering that genetic heterogeneity is common in complex diseases
such as PD and the genetic basis for the majority of sporadic PD
cases remains unclear, a comprehensive and large-scale study
focusing on genome level analysis of sporadic PD in China is
recommended.
10. J. F. Guo et al., Mov. Disord. 23, 2074 (2008).
REFERENCES
1. J. Y. Tian et al., Neurosci. Lett. 516, 207 (2012).
2. J. F. Guo et al., Parkinsonism Relat. Disord. 18, 983 (2012).
3. K. Li et al., BMC Neurol. 13, 38 (2013).
4. L. Wang et al., Neurosci. Lett. 468, 198 (2010).
5. C. Y. Wang et al., Brain Res. 1385, 293 (2011).
6. L. Wang et al., Neurosci Lett. 473, 131 (2010).
7. L. Wang et al., J. Clin. Neurosci. 18, 1699 (2011).
8. J. L. Wang et al., Mov. Disord. 24, 2007 (2009).
9. Q. Xu et al., Parkinsonism Relat. Disord. 16, 700 (2010).
11. C. H. Shi et al., Neurology 77, 75 (2011).
12. J. F. Guo et al., J. Neurol. 257, 1170 (2010).
13. L. Z. Luo et al., Neurosci. Lett. 482, 86 (2010).
14. J. Y. Tian et al., Neurosci. Lett. 514, 156 (2012).
15. B. Tang et al., Hum. Mol. Genet. 15, 1816 (2006).
16. Y. Hu et al., J. Neurol. 259, 497 (2012).
17. C. Wang et al., Parkinsonism Relat. Disord. 18, 958 (2012).
18. Q. Y. Sun et al., Mov. Disord. 25, 1005 (2010).
19. J. Liu et al., Brain Res. 1392, 116 (2011).
Acknowledgments: This work was supported by grants from the Major
State Basic Research Development Program of China (“973” Program)
(2011CB510001) and the National Natural Science Foundation of China
(81130021, 81171198, and 81371405).
PKD and PRRT2 -Related Paroxysmal Diseases in China
Jun-Ling Wang1,2,3, Nan Li1,2,3, Hong Jiang1,2,3, Lu Shen1,2,3,
Kun Xia2, and Beisha Tang1,2,3*
Paroxysmal kinesigenic dyskinesia (PKD) (OMIM# 128200), also
known as episodic kinesigenic dyskinesia (EKD), is a rare and
remarkable hereditary neurological disorder that was first described in 1892 and more precisely defined during the 20th cen-
Department of Neurology, Xiangya Hospital, Central South
University, Changsha, Hunan, China
2
State Key Laboratory of Medical Genetics, Changsha, Hunan, China
3
Human Province Key Laboratory of Neurodegenerative Disorders,
Changsha, Hunan, China
*
Corresponding Author: [email protected]
1
tury. PKD is characterized by childhood or adolescent onset of
brief but recurrent attacks of the involuntary movements triggered
by sudden voluntary movements. The episodes usually present
with dystonia, chorea, athetosis, ballism, or their combination,
with no loss or alteration of consciousness. The attacks usually
last for under 15 seconds and the frequency of attacks can range
from more than 100 per day to less than one per month. Other
characteristic features include excellent control of attacks with
carbamazepine or phenytoin, a history of benign infantile convulsions in some patients, and a normal neurological examination.
Most idiopathic PKD patients have a family history, usually in an
37
GENETIC AND CLINICAL STUDIES
Produced by the Science/AAAS Custom Publishing Office
Figure 1. Schematic diagram of PRRT2 gene and wild type and mutant PRRT2 protein. The
three mutations identified are indicated with red arrows (top). Schematic representation of chromosome 16, shows the overlapping loci for PKD (EKD2), ICCA, and BFIS, and the position of the
PPRT2 gene within the 16p11.2 region (bottom).
autosomal dominant mode, although sporadic
cases are also observed.
Linkage studies with a variety of ethnicities
confirmed a critical region at 16p11.2–q12.1
that was named PKD1, was about 12.4cM
long, and encompassed over 157 candidate
genes. Two other regions that did not overlap
with PKD1 were also identified and denoted
PKD2 (16q12.1–q21) and PKD3 (precise location not known). Over the last decade, a considerable amount of work has been done on the
identification of the genetic basis for PKD using
traditional positional cloning strategies. Nevertheless, no causative gene could be identified.
Using a cloning strategy as a primary means for
disease gene identification is limited, especially
in circumstances such as small pedigrees, locus
heterogeneity, substantially reduced reproductive fitness, and an abundance of candidate
genes present in the mapped region. Recently,
whole exome sequencing has been shown to be
a powerful approach to identify causative genes
underlying a number of Mendelian disorders. For
example, in 2010, we used a combined strategy
of exome sequencing and linkage analysis to
identify a novel spinocerebellar ataxia causative
gene, TGM6 (1). This strategy considerably reduces the number of variants potentially implicated, making it a highly accurate and affordable
method for identifying a causative gene using
only a very small number of samples.
In 2011, our group utilized the above strategy
to identify proline-rich transmembrane protein
2 (PRRT2) as the first causative gene of PKD
(2). We performed genetic linkage mapping with
eleven markers that encompassed the pericentromeric region of chromosome 16 in 27 members of two families with autosomal dominant
PKD. We then performed whole exome sequencing in three patients from these two families. By combining the sequencing and linkage
data (region 16p12.1–q12.1), we identified a
c.649dupC (p.P217fsX7) mutation in one family
and a nonsense c.487C>T (p.Q163X) mutation
in the other. Both mutations created premature
38
termination codons, indicating either a truncated protein or degradation of the mutated messenger RNA by nonsense-mediated
decay. To confirm these findings, we sequenced the exons and
flanking introns of PRRT2 in another three PKD families. The
c.649dupC (p.P217fsX7) mutation was found in two families,
and a missense mutation, c.796C>T (R266W), was identified in
the other (2). This finding was rapidly followed by many similar reports of mutations in the same gene in other families with
PKD in China, and PRRT2 was identified as the major causative
gene for PKD by independent groups in China. Studies of different Chinese PKD cohorts have identified PRRT2 mutations in
37% of cases, including 40 out of 58 familial cases (69%) and
26 out of 120 sporadic cases (22%) (3–10). The fact that the
remaining 63% of typical PKD patients do not harbor mutations
in PRRT2 suggests at least one other causative gene, or other
types of PRRT2 gene disruptions that are difficult to detect by
gene sequencing. In total, 26 mutations have been reported,
with the c.649dupC (p.P217fsX7) mutation being the most common (~80% of PKD sufferers), making a good case for priority
screening for this mutation in Chinese PKD patients. Notably, the
c.649dupC mutation is also observed as a de novo mutation in
some sporadic PKD cases and is adjacent to nine consecutive
cytosine (C) bases, indicating that the region is potentially genetically unstable (5, 11).
oxysmal dyskinesia. Family studies of PKD and ICCA showed
linkage to the same pericentromeric region on chromosome
16 (Fig. 1). It has long been proposed that ICCA and PKD are
two different expressions of the same disorder. In our study, although most patients with a PRRT2 mutation exhibited the pure
forms of PKD, there were two members separately diagnosed
with ICCA syndrome in infancy from two unrelated families.
This suggested that PRRT2 may also be the causative gene for
ICCA (12).
BFIS without PKD was also linked to the same locus on chromosome 16p12-q12 (Fig. 1). Despite distinct clinical features,
we hypothesized that mutations in PRRT2 might cause multiple
clinical phenotypes including pure forms of PKD, pure forms
of BFIS, or a combination of both (ICCA). To test this, we collected samples from two BFIS families of Han Chinese origin.
Linkage analysis showed that the BFIS-causing locus mapped to
16p12.1–q12.2, the location of PRRT2. We then performed mutational analysis of PRRT2 by direct sequencing and identified a
c.649dupC (p.P217fsX7) mutation in all BFIS patients (12). Interestingly, in the subsequent studies in China, PRRT2 gene mutations were also found be implicated in other paroxysmal disorders, including paroxysmal exercise-induced dyskinesia (PED),
paroxysmal non-kinesigenic dyskinesia (PNKD), febrile convulsions (FC), infantile non-convulsive seizures (INCS), nocturnal
Disease
Mutation Type
China
World
PKD
c.649dupC, c.649 delC, and others
+
+
ICCA
c.649dupC, c.649 delC, and others
+
+
BFIS
c.649dupC, and others
+
+
PNKD
c.649dupC
+
+
PED
c.649dupC
+
-
FC
c.649dupC, and others
+
-
NC
c.649dupC
+
-
INCS
c.649dupC,c.510dupT, and others
+
-
Episodic ataxia (EA)
c.649dupC
-
+
Hemiplegic migraine (HM)
c.649dupC, and others
-
+
Infantile focal epilepsy
c.649dupC
-
+
Table 1. Summary of PRRT2 screening in paroxysmal disease patient. “+” = observed; “-”= not observed
Some PKD patients, or their first and second degree relatives,
presented with infantile convulsions (IC) or infantile convulsion
and choreoathetosis (ICCA). ICCA is inherited in an autosomal
dominant fashion and is characterized by benign familial infantile
seizures (BFIS) that start within the first year of life and usually stop by three years of age, variably associated with par-
convulsions (NC), and one case representing a phenotypic overlap among different subtypes of paroxysmal dyskinesis (PxDs)
that did not fit into the traditional classification groups (Table 1)
(4, 9, 13).
We also noticed that diseases such as BFIS, PKD, and ICCA
with PRRT2 mutations share some characteristics, such as
39
GENETIC AND CLINICAL STUDIES
Produced by the Science/AAAS Custom Publishing Office
responding well to anti-epileptic treatment and attacks being
paroxysmal. Therefore, considering the fact that this group of
PRRT2-related diseases share the same causative gene (possibly even the same mutation), that these disorders show overlapping occurrence, and that they respond to the same treatment
method, we suggest naming this spectrum of disorders “PRRT2related paroxysmal diseases,” or PRPDs, to simplify clinical classification. PRPDs should be regarded as the general terminology
currently associated with BFIS, PKD, PNKD, PED, ICCA, and
certain types of epilepsies, and in which paroxysmal attacks and
mutations in PRRT2 are characteristic features. This nosological criterion is meant to serve several purposes. It highlights the
use of anticonvulsants to treat this spectrum of diseases and enlarges the disease spectrum of PxDs, especially for those PxDs
cases that are difficult to fit into current the classifications based
on triggers.
Although the PKD causative gene has been identified, the
pathophysiology of PRRT2 mutations remains largely unknown.
Whether PKD has a subcortical or cortical origin is controversial. A resting-state functional magnetic resonance imaging
study performed in ten Chinese PKD patients suggested that increased spontaneous brain activity in the cortical-basal ganglia
circuitry, especially in the motor preparation areas, is a common
pathology of PKD, indicating a lack of readiness for movement
in PKD. It was also found that the differences in the spatial patterns of increased amplitude of low-frequency fluctuation (ALFF)
between PKD patients with and without the p.P217fsX7 mutation
might reflect the distinct pathological mechanism resulting from
the PRRT2 mutation (14). In addition, a genotype/phenotype relationship study in China revealed that PRRT2-PKD has different phenotypes and drug response compared with non-PRRT2–
PKD. These findings provide new insights into the neural system
effects of the PRRT2 gene in PKD (15).
In conclusion, we used a strategy combining exome sequencing and classical genome-wide linkage studies to successfully
identify PRRT2 as the causative gene for PKD. This gene also
appears to be responsible for other similar disorders including
BFIS, ICCA, PNKD, PED, FC, INCS, NC, and a series of paroxysmal disorders. For convenience in both clinical practice and
research, we suggest naming this group of paroxysmal diseases
harboring PRRT2 mutations, PRPDs. It is conceivable that PRPDs will provide a new research area in the field of neuroscience
and provide new insights into the causes of epilepsy and movement disorders.
REFERENCES
1. J. L. Wang et al., Brain 133, 3510 (2010).
2. J. L. Wang et al., Brain 134, 3493 (2011).
3. W. J. Chen et al., Nat. Genet. 43, 1252 (2011).
4. J. Li et al., J. Med. Genet. 49, 76 (2012).
5. L. Cao et al., Parkinsonism Relat. Disord. 18, 704 (2012).
6. Q. Liu et al., J. Med. Genet. 49, 79 (2012).
7. Y. P. Chen et al., Eur. J. Neurol., Epub ahead of print (2013) doi: 10.1111/ene.12122.
8. C. H. Shi et al., Mov. Disord. 28, 1313 (2013).
9. X. R. Liu et al., Genes Brain Behav. 12, 234 (2013).
10. X. Y. Jing et al., Parkinsonism Relat. Disord. 19, 639 (2013).
11. H. F. Li, W. Ni, Z. Q. Xiong, J. Xu, Z. Y. Wu, CNS Neurosci. Ther. 19, 61 (2013).
12. J. L. Wang et al., Neurosci. Lett. 552, 40 (2013).
13. K. Wang et al., Brain Dev. 35, 664 (2013).
14. C. Luo et al., Neurol. Sci., Epub ahead of print (2013) doi: 10.1007/
s10072-013-1408-7.
15. H. F. Li et al., Neurology 80, 1534 (2013).
Acknowledgments: This work was funded by projects in the Major
State Basic Research Development Program of China (“973” Program)
(2011CB510001), the State Key Program of the National Natural Science
Foundation of China (81130021), and the National Natural Science
Foundation of China (81300981).
Clinical and Molecular Biological Studies of Amyotrophic
Lateral Sclerosis in China
Zhangyu Zou1 and Liying Cui2*
Amyotrophic lateral sclerosis (ALS) is a syndrome characterized
by progressive muscle weakness and atrophy resulting from loss
of motor neurons in the corticospinal tract, brainstem, and anterior cells of the spinal cord. More than 60% of patients die within
three years of onset (1). In the past fifteen years, epidemiological,
genetic, and electrophysiological studies, as well as research
into the treatment of ALS, have been performed at the Peking
Union Medical College Hospital of China. Here we summarize
some of the findings.
Epidemiological studies
A registry-based study of ALS was conducted from March 1,
2009 to August 31, 2009 at 10 centers overseen by the Chinese
ALS Association. Data from 455 patients yielded the following
statistics: the male to female ratio was 1.63:1; sites of onset included the bulbar region (30.9%), the upper limbs (54.1%), the
thoracic region (0.7%), the lower limbs (20.7%), and unspecified
regions (3.7%); and the mean age at onset in China was 52.4
years, which was earlier than that reported in many other countries, including Japan, Uruguay, Germany, France, and the
United States (mean age at onset range 55.7–66.0 years) (2).
The data showed an earlier age at onset for manual (blue collar)
workers (50.7±11.4 years) than that of professional (white collar)
workers (56.2±12.9 years) (P < 0.001). The most common age at
onset (45 to 49 years for females and 55 to 59 years for males)
was also earlier than that of patients from Europe and the United States. The earlier mean and peak age at onset for Chinese
ALS patients suggests a potential role for genetic susceptibility
and exogenous factors, such as environmental pollution, in the
pathogenesis of ALS (2).
Genetic Studies
Approximately 5%–10% of patients have familial ALS (FALS) and
show a Mendelian pattern of inheritance; the remaining 90%–
95% carry sporadic ALS (SALS) mutations (3). Sixteen genes
and loci have been identified in FALS: SOD1, ALS2, SETX,
1
Department of Neurology, Affiliated Union Hospital of Fujian Medical
University, Fuzhou, China
2
Department of Neurological, Peking Union Medical College Hospital,
Beijing, China
*
Corresponding Author: [email protected]
40
SPG11, FUS, VAPB, ANG, TARDBP, FIG4, OPTN, ATXN2, VCP,
UBQLN2, C9ORF72, and PFN1 (1, 4). Mutations in SOD1, ANG,
TARDBP, FUS, VCP, C9orf72, and PFN1 have been screened in
ALS patients of Chinese origin.
In a Chinese cohort of 142 SALS patients and seven FALS
pedigrees, three missense mutations in the SOD1 gene—p.
H46R, p.G72C, and p.E133V—were identified in three FALS
patients; another three missense mutations—p.V29A, p.V47A,
and p.N86I—were identified in three SALS cases. Phenotypic
analysis revealed that all four patients within the p.H46R
mutation FALS pedigree showed lower limb onset and very
slow progression, with a survival of more than ten years (5,
6). The clinical features of a p.H46R mutation pedigree in our
study were similar to those previously reported in p.H46R
pedigrees analyzed in studies of Japanese patients (7, 8).
These findings suggest that the p.H46R mutation is more common in the Asian population, and associated with a specific
phenotype of lower limb onset, slow progression, and long
survival (5, 6).
In 202 SALS patients and 10 FALS pedigrees of Chinese origin,
a novel missense mutation—ANG c.379G>A (p.V103I)—was
identified in one SALS patient. This was the first ANG mutation
identified in an Asian population. The prevalence of ANG mutations
in SOD1-, FUS-, and TARDBP-mutation–negative ALS patients
in Chinese population was 0.5% (9). Although an association of
ALS with the G variant of the ANG rs11701 (T/G) single nucleotide
polymorphism (SNP) was reported in Italian, Irish, and Scottish
populations, no such association was demonstrated in the Chinese
population (9).
In a Chinese cohort of 13 FALS pedigrees and 321 SALS patients,
two heterozygous missense mutations in the TARDBP gene—
c.875G>A (p.S292N) and c.1043G>T (p.G348V)—were identified
in three SALS patients. In the Chinese population, the estimated
frequency of TARDBP mutations in SALS patients (0.73%) is higher than Japanese and lower than Caucasian populations, whereas the estimated mutation frequency in SOD1-negative FALS
patients (15.2%) is higher than both Japanese and Caucasian
populations (10).
Screening for mutations in the FUS gene was performed in 320
ALS patients and 16 FALS pedigrees identified two missense
mutations—c.1562G>T (p.R521L) and c.1561C>G (p.R521G)—
in one FALS proband. Two additional missense mutations—
c.1562G>A (p.R521H) and c.1574C>T (p.P525L)—as well as a
41
GENETIC AND CLINICAL STUDIES
Produced by the Science/AAAS Custom Publishing Office
two-base pair deletion c.1509_1510delAG (p.G504Wfs*12) and
a nonsense mutation c.1483C>T (p.R495X), were identified in
five SALS patients (11, 12). In the Chinese population, the frequency of FUS mutation in FALS is 10.0%, similar to the Japanese (10%), but higher than in Caucasians (4.9%). The frequency of FUS mutation in SALS patients is 1.5%, which is similar to
Korean sufferers (1.6%), but higher than in Caucasians (0.6%)
(11, 12).
No mutations in the VCP or PFN1 genes (13, 14) and no abnormal GGGGCC hexanucleotide repeat expansions in the noncoding region of the C9orf72 gene (15) were identified in either
FALS or SALS patients of Chinese origin, which suggests that
mutations in VCP, PFN1, and C9orf72 are rare in this population.
The SNP rs10260404 in the DPP6 gene is strongly associated
with susceptibility to ALS across different populations of European and American ancestry. However, it was not associated with
ALS susceptibility in Chinese SALS patients (16).
Our studies suggest that mutations in SOD1 and FUS are the
most common mutations in ALS patients of Chinese origin (5,
11, 12), whereas mutations in C9orf72 are rare (15), highlighting
another difference between Chinese and Caucasian populations
in the genetic background of ALS.
Electrophysiological Studies
ALS affects both upper motor neurons (UMN) and lower motor
neurons (LMN), however, in the early stages of the disease
patients may have focal involvement of only UMN or LMN.
Therefore, it can be challenging to diagnose ALS in the early
stages by clinical examination alone (17). A registry-based study
showed that the mean duration from onset to diagnosis of ALS
was 13.8 months in a Chinese cohort, longer than that reported
in western countries (9 to 12 months) (2, 18). Electrodiagnostic evaluation is valuable in the differential diagnosis of ALS.
It extends a physical examination and is able to detect LMN
abnormalities early in the disease. It is also key to enabling doctors to exclude other diseases that mimic ALS and to establish
some diagnostic certainty once other disease processes have
been excluded (17).
Nerve conduction studies are an essential part of the electrodiagnostic evaluation of patients with suspected ALS. A study
of nerve conduction in 205 ALS patients identified common
abnormalities including prolonged distal motor latency (20.6%),
absent compound muscle action potentials (CMAPs) or CMAPs
with decreased amplitude (52.4%), decreased F-wave frequency
(68.9%), and slowed conduction velocity consistent with axon loss
(19). A strong association was found between decreased CMAP
amplitudes and Medical Research Council (MRC) Scale strength
and disease duration. The amplitudes of CMAP decreased as
muscle strength decreased and the disease progressed. F-wave
persistence was abnormal in 122 out of 177 patients with ALS
and decreased as MRC grade strength decreased (19).
42
Needle electromyography (EMG) is the most important component of the electrodiagnostic evaluation in ALS. Since the clinical
and electrophysiological identification of LMN abnormalities in
the same body region have equal diagnostic weigh, needle EMG
can enable earlier diagnosis and treatment before clinically obvious symptoms arise (17). Needle EMG in ALS patients should
reveal decreased motor unit recruitment and/or large-amplitude,
long-duration motor unit potentials in combination with abnormal
spontaneous activity, including positive sharp waves, fibrillations,
and/or fasciculation potentials (17, 20).
For a definitive diagnosis of ALS, LMN abnormalities must
be documented in at least three of the four anatomical regions
usually examined. Needle EMG evaluation should therefore
include the cervical and lumbar regions, as well as the muscles
innervated by the cranial nerves and the thoracic region (17). The
tongue is traditionally sampled to evaluate LMN involvement, but
it is difficult to achieve relaxation and obtain accurate readings.
The sternocleidomastoid muscle (in the neck) is innervated by
the cranial nerve and upper cervical roots, so EMG changes in
this muscle could be regarded as evidence of LMN involvement
in the bulbar region (21). Paraspinal muscles between the T8T10 root levels are not innervated by the lower cervical or upper
lumbar segments and could therefore be regarded as thoracic
region muscles for the purposes of EMG testing (22).
UMN involvement in patients with ALS can be physiologically assessed using transcranial magnetic stimulation (TMS).
An ALS study performed with 40 patients, eight with pure LMN
involvement and 34 health controls, suggested that TMS-motor
evoked potential (MEP) is not a good measure for early diagnosis of ALS. However, in patients with so-called clinically probable, laboratory-supported ALS or possible ALS (measured
based on the number of body regions involved), TMS-MEP may
help to demonstrate UMN involvement in these patients and
facilitate early diagnosis (23). Fifty ALS patients and 22 healthy
controls were examined using conventional TMS and the triple
stimulation technique (TST) at the foot muscle, abductor digiti
minimi. Central motor conduction time (CMCT), MEP, and rest
motor threshold (RMT) were assessed. The TST amplitude ratio
was significantly decreased in ALS patients with symptoms of
UMN involvement compared to those without involvement or
the controls (both P<0.001). Results using TST correlated well
with those assessed on the Modified Ashworth Scale (r=-0.772,
P<0.001). These findings suggest that TST may be a more accurate and sensitive measure of detecting UMN abnormality in
ALS patients than previously used techniques, and could reveal
subclinical UMN impairment in ALS patients. It may also help to
assess the UMN loss in ALS and serve as an objective parameter
for monitoring disease progression (24).
Research into Treatments
Currently, there are no effective treatments or drugs for this fatal
disease. In vivo studies have demonstrated neuroprotective effects of L-3-n-butylphthalide (L-NBP), extracted from the seeds
of Apium graveolens var. Linn. (also known as Chinese celery),
in ischemic, vascular dementia, and β-amyloid–infused animal
models (25). DL-3-n-butylphthalide (DL-NBP), an analog of LNBP, has been chemically synthesized and was approved in
2002 by the Chinese State Food and Drug Administration (SFDA)
for clinical use in stroke patients. In vivo studies have shown that,
following the presentation of ALS-like symptoms in transgenic
SOD1-G93A mice, oral administration of DL-NBP significantly
decreased the progression of motor deficits, suppressed body
weight loss, attenuated motor neuron loss, delayed motor unit
reduction, and extended the lifespan, compared with vehicle
controls. DL-NBP treatment also caused a significant reduction
in immune reactivity of CD11b and glial fibrillary acidic protein,
markers for microglia and astrocytes, respectively. Additionally,
downregulation of nuclear factor кB (NF-кB), p65, and tumor
necrosis factor-α protein levels, and a slight upregulation of NFE2–related factor 2 (Nrf2) and heme oxygenase-1 (HO-1) were
found in the spinal cords of SOD1-G93A mice treated with DLNBP. These results suggest that DL-NBP could be a promising
therapeutic agent for ALS (26, 27).
7. K. Abe et al., J. Neurol. Sci. 136, 108 (1996).
8. T. Arisato et al., Acta Neuropathol. 106, 561 (2003).
9. Z. Y. Zou et al., Amyotroph Lateral Scler. 13, 270 (2012).
10. Z. Y. Zou et al., Neurobiol. Aging 33, 2229.e11 (2012).
11. Z. Y. Zou et al., Neurobiol. Aging 34, 1312.e1 (2013).
12. Z. Y. Zou et al., Eur. J. Neurol. 19, 977 (2012).
13. Z. Y. Zou, M. S. Liu, X. G. Li, L. Y. Cui, Neurobiol. Aging 34, 1519.
REFERENCES
1. O. Hardiman, L. H. van den Berg, M. C. Kiernan, Nature Rev. Neurol. 7, 639 (2011).
2. M. S. Liu, L. Y. Cui, D. S. Fan, Acta Neurol. Scand. Epub ahead of print (2013) doi: 10.1111/ane.12157.
1713.e5 (2013).
15. Z. Y. Zou, X. G. Li, M. S. Liu, L. Y. Cui, Neurobiol. Aging 34, 1710.
e5 (2013).
16. X. G. Li et al., Chinese Med. J. 122, 2989 (2009).
17. L. Y. Cui, Zhong Guo Sheng Jin Mian Yi Xue He Sheng Jin Bing Xue Zha Zhi 19, 247 (2012).
18. A. Al-Chalabi, O. Hardiman, Nature Rev. Neurol. Epub ahead of print (2013) doi: 10.1038/nrneurol.2013.203.
19. X. H. Feng et al., Zhong Hua Sheng Jin Ke Zha Zhi 44, 178 (2011).
20. M. S. Liu et al., Zhong Hua Sheng Jin Ke Zha Zhi 43, 204 (2010).
21. M. S. Liu, L. Y. Cui, X. F. Tang, B. H. Li, H. Du, Zhong Hua Sheng Jin Ke Zha Zhi 35, 361 (2002).
22. X. F. Tang, H. Pan, B. H. Li, H. Du, Zhong Hua Sheng Jin Ke Zha Zhi e3 (2013).
14. Z. Y. Zou, Q. Sun, M. S. Liu, X. G. Li, L. Y. Cui, Neurobiol. Aging 34, 36, 176 (2003).
23. F. Jian et al., Nao Yu Shen Jin Ji Bin Zha Zhi 14, 345 (2006).
24. Y. Wang, L. Y. Cui, H. Wang, Zhong Hua Sheng Jin Ke Zha Zhi 43, 562 (2010).
25. Y. Peng et al., J. Alzheimers Dis. 29, 379 (2012).
26. X. Feng, Y. Peng, M. Liu, L. Cui, Neuropharmacology 62, 1004 3. A. Chio et al., Neurology 70, 533 (2008).
4. C. H. Wu et al., Nature 488, 499 (2012).
27. X. H. Feng, W. Yuan, Y. Peng, M. S. Liu, L. Y. Cui, Chin. Med. J. 5. X. G. Li et al., Zhonghua Shen Jing Ke Za Zhi 43, 686 (2010).
(2012).
125, 1760 (2012).
6. X. G. Li et al., Xie He Yi Xue Zha Zhi 3, 337 (2012).
Brainnetome Studies of Alzheimer’s Disease Using Neuroimaging
Tianzi Jiang1,2,3,4*, Yong Liu1,2, and Bing Liu1,2
The human brain has been modeled as a hierarchy of complex
networks on different temporal and spatial scales. The brainnetome (brain-net-ome), a new “ome” in which the brain network is the basic research unit used to investigate the structural
and functional organization of the human brain, from genes and
neuronal circuits to behaviors (1). Alzheimer’s disease (AD), the
most common form of dementia, is a neurodegenerative disease
with typically but not exclusively memory deficits that affect daily
cognitive function. The symptoms of AD are not thought to be
due to a single, regionally specific pathophysiology, but rather result from abnormal interactions between different brain regions.
Therefore, the malfunctioning of connections and brain networks
43
GENETIC AND CLINICAL STUDIES
Produced by the Science/AAAS Custom Publishing Office
may underlie some of the manifestations of AD. Here we review
advances in the brainnetome studies of AD using neuroimaging
techniques including magnetic resonance imaging (MRI), functional MRI (fMRI), and diffusion MRI. We have demonstrated that
the patients with AD have brain atrophy (2–5), and lower amplitude and reduced regional homogeneity (4, 6, 7) in brain activity.
Emerging evidence has shown that AD is also associated with
impaired functional integration among spatially distinct brain regions (4, 8, 9).
Disrupted Brain Networks in AD
The human brain can generate and integrate information with
high efficiency and maintain a perfect balance between local and
global interactions. Studies have demonstrated that the symptoms of AD result from distortions of functional and anatomical
networks.
Altered Brain Networks Based on Region-ofInterest Studies
Region-of-interest analysis is the most common method for investigating the functional connectivity pattern of a specific region. A region is selected as a ‘seed’ and the correlation map
between this region and every other voxel of the brain is then
evaluated. Applying this method to AD patients, disrupted functional connectivity between the hippocampus and a set of brain
regions in the default mode network was found, as well as increased functional connectivity between the left hippocampus
and the right lateral prefrontal cortex (10). Zhou and colleagues,
selecting the thalamus as the region of interest, found that alterations in connectivity between the thalamo-default mode
network and thalamo-cortical network were related to the severity of AD symptoms (11). Additionally, using fMRI on AD
sufferers, researchers found impaired functional connectivity
(compared to normal controls) to be located mainly between
the amygdala and regions that are encompassed by the default mode (i.e., medial posterior parietal cortex and dorsal
medial prefrontal cortex), context conditioning, and extinction
networks (12).
These disease-related decreases in functional connectivity support the hypothesis of network disconnection reflected
1
Brainnetome Center, Institute of Automation, Chinese Academy of
Sciences, Beijing, China
2
National Laboratory of Pattern Recognition, Institute of Automation,
Chinese Academy of Sciences, Beijing, China
3
The Queensland Brain Institute, The University of Queensland,
Brisbane, Australia
4
Key Laboratory for NeuroInformation of Ministry of Education,
School of Life Science and Technology, University of Electronic
Science and Technology of China, Chengdu, China
*
Corresponding Author: [email protected]
44
in cognitive function impairment, while the increased connectivity could be interpreted as a compensatory recruitment of cognitive resources to maintain task performance in
AD patients.
Abnormalities of Specific Brain Networks in AD
Using independent component analysis, Song and colleagues
found severity-related decreased functional connectivity in the
several cognitive-related brain networks in AD, including the bilateral precuneus of the precuneus network, the posterior cingulate cortex and left precuneus of the posterior default mode
network, and the left superior parietal lobule of the left frontoparietal network (3). In addition, we found that the functional connectivity pattern of the default mode network and its anticorrelation network are disrupted, and that these connections can be
used as features to distinguish AD patients from healthy subjects
(13). These results suggest that the brain networks supporting
complex cognitive processes are specifically and progressively
impaired over the course of AD.
Abnormalities of Whole Brain Networks in AD
We carried out a study in both AD patients and healthy elderly
subjects in which we virtually divided the entire brain into regions
using an anatomically labeled template and paired every region
with every other region, calculating the correlation coefficients
between each pair of regions (14). Our results demonstrated
that most of the decreased functional connectivity occurred between frontal and parietal lobes. This abnormal brain network
was confirmed by another independent study in which we found
that patients with AD had weaker functional connectivity between
regions that were separated by a greater physical distance, especially in several regions previously described as components
of the default mode network. More importantly, the impaired
functional connectivity in severe AD was consistent with results
from studies of mild AD and amnestic mild cognitive impairment
(4). The more severe the impairment, as measured by the minimental state examination (MMSE), the greater the attenuation
of functional connectivity, especially in brain regions functionally
connected over a long distance (4).
Abnormal Topological Characteristics of
the Brain Network in AD
The human brain is a complex network that reflects a balance
between local processing and global integration of information.
Two of our independent studies demonstrated that the topological characteristics of the brain network are disrupted in AD (4, 8).
We found that the altered brain regions were mainly located in
the default mode network, the temporal lobe, and certain subcortical regions that are closely associated with neuropathological
changes in AD (4, 8). In patients with severe AD, as mentioned
above, functional connectivity was particularly attenuated be-
tween regions separated by a greater physical distance, and loss
of long distance connectivity was associated with a less efficient
global and nodal network topology. Another independent structural network study also confirmed that the AD brain network is
distorted, showing a longer path length (9) and suggesting a loss
of network efficiency in AD. The results also showed greater attenuation in those patients with more severe cognitive impairment, hinting at the potential for brainnetome measures to be
used as biomarkers or predictors of disease progression in AD.
Brain Networks are Modulated by AD Risk Genes
Recent pedigree studies have demonstrated that brain networks
disrupted in AD are affected by genetic factors. For example,
data from a resting state fMRI study showed that the heritability of default network functional connectivity was about 42%
(15). Also, the information transfer efficiency of functional brain
networks has been reported to be under genetic control. Previous genetic and bioinformatics studies have demonstrated that
AD was affected by interactions between multiple genetic variants and environments (16). Therefore, genetic imaging studies,
which investigate the influence of specific genetic variations on
disruptions in brain networks in AD, are attracting increased interest from Alzheimer’s researchers.
There is particular interest in the genetic aspects of the default
mode network, which is consistently reported to be disrupted in
AD. The KIBRA gene, which is associated with memory performance and AD, has been found to impact synchronization within
the default-mode in healthy, young populations (17). Studies
have also reported that distinct patterns of resting brain activity within the default network in young and older carriers of the
apolipoprotein (Apo) E-ε4 allele can be detected before any
neurophysiological expression of neurodegenerative processes.
An additional resting state fMRI study reported that the ApoE
genotype affects brain network properties, especially in AD patients (8). Moreover, Liu et al. (18) reported that the catechol-Omethyltransferase (COMT) val158met polymorphism, which
plays a unique role in regulating the prefrontal dopamine level,
significantly modulates prefrontal-related functional connectivity within the default mode network. All of these studies suggest
that default network activity and connectivity are affected by ADrelated genetic variants.
Besides functional imaging, structural brain connectivity and
neural integrity markers are also attracting increasing attention in
imaging the genetics of AD. Studies have reported that the specific polymorphism and haplotypes of the COMT gene modulate
the integrity of important white matter tracts (19, 20). A diffusion
MRI study found that the global information transfer efficiency of
the human brain anatomical network was affected by a disruption
in the schizophrenia 1 gene (21). Studies also reported that ADrelated cortical thickness and gray matter volume was modulated
by ApoE genetic variants (22) and specific long noncoding RNAs
(23). However, recent studies found that certain specific brain
connectivity or brain networks were modulated by risk genes for
AD (8, 17, 18, 20, 24), and thus systematically investigating the
genetic control of neural connectivity within AD-related brain systems will likely be an important research direction in the near
future.
Conclusions
As discussed, impairment of the functional and anatomical
network architecture of the brain in AD has been verified using neuroimaging. Moreover, studies have demonstrated that
genes associated with AD risk play a role. These findings indicate that abnormalities in brain networks may be biomarkers of disease progression in AD. The brainnetome research
may therefore open up new avenues for understanding and
treating AD.
REFERENCES
1. T. Jiang, Neuroimage 80C, 263 (2013).
2. S. Li et al., Acta Radiol. 49, 84 (2008).
3. J. Song et al., PLOS ONE 8, e63727 (2013).
4. Y. Liu et al., Cereb. Cortex. Epub ahead of print (2013) doi: 10.1093/
cercor/bhs410.
5. S. Li et al., Am. J. Neuroradiol. 28, 1339 (2007).
6. Y. He et al., Neuroimage 35, 488 (2007).
7. Z. Zhang et al., Neuroimage 59, 1429 (2012).
8. X. Zhao et al., PLOS ONE 7, e33540 (2012).
9. Z. Yao et al., PLOS Comput Biol. 6, e1001006 (2010).
10. L. Wang et al., Neuroimage 31, 496 (2006).
11. B. Zhou et al., Curr. Alzheimer Res. 10, 754 (2013).
12. H. Yao et al., Eur. J. Radiol. 82, 1531 (2013).
13. K. Wang et al., Med. Image Comput. Comput. Assist. Interv. 9, 340 (2006).
14. K. Wang et al., Hum. Brain Mapp. 28, 967 (2007).
15. D. C. Glahn et al., Proc. Natl. Acad. Sci. U.S.A. 107, 1223 (2010).
16. B. Liu et al., Biochem. Biophys. Res. Commun. 349, 1308 (2006).
17. D. Wang et al., Neuroimage 69, 213 (2013).
18. B. Liu et al., J. Neurosci. 30, 64 (2010).
19. J. Li et al., Am. J. Med. Genet. B Neuropsychiatr. Genet. 150B, 375 (2009).
20. B. Liu et al., Neuroimage 50, 243 (2010).
21. Y. Li et al., Cereb. Cortex. 23, 1715 (2013).
22. M. Fan et al., Neurosci. Lett. 479, 332 (2010).
23. G. Chen et al., Hum. Mutat. 34, 338 (2013).
24. A. J. Trachtenberg et al., Neuroimage 59, 565 (2012).
Acknowledgments: This work was partially supported by the National
Key Basic Research and Development “973” Program (2011CB707800),
the Strategic Priority Research Program of the Chinese Academy of
Sciences (XDB02030300), and the Natural Science Foundation of China
(91132301 and 81270020).
45
GENETIC AND CLINICAL STUDIES
Produced by the Science/AAAS Custom Publishing Office
Fragile X Syndrome in China: A Clinical Review
Ranhui Duan
Fragile X syndrome (FXS) (MIM#300624) is the most common
cause of inherited intellectual disability, affecting 1 in 4,000 to
1 in 6,000 males and 1 in 8,000 females (1). Since expansion
of a trinucleotide repeat in the fragile X mental retardation gene
(FMR1) was first identified to cause FXS in 1991, basic science
and clinical researchers have been working to find ways to improve the disease outcome in FXS families. Chinese scientists
and health care professionals have made significant contributions to this work. This review describes advances in the diagnosis, genetic testing, and counseling of FXS sufferers in China.
Introduction
J. Purdon Martin and Julia Bell firstly described a pedigree of
X-linked mental retardation in 1943 that we now know as FXS.
The vast majority of cases are caused by expansion of the trinucleotide CGG repeat in the 5ʹ-untranslated region of FMR1.
Normal alleles contain between six and 44 repeats, intermediate alleles between 45 and 54 repeats, and premutation alleles
in the range of 55 to 200 unmethylated repeats. During either
oogenesis or early post-zygotic events, premutations may undergo further expansion, and methylation, to generate over 200
repeats. Subsequent hypermethylation and epigenetic silencing
of FMR1, resulting in the loss of its protein product—fragile X
mental retardation protein (FMRP)—causes FXS (2). Moreover,
40% of males (over 50 years old) with premutations will develop
fragile X-associated tremor/ataxia (FXTAS), a progressive neurodegenerative disorder, and approximately 20% of females with
premutations are at increased risk for premature ovarian insufficiency (3, 4).
A survey of the parents of children with special needs demonstrated that 95% of them were not aware of FXS (5). An investigation among medical school students showed more than
two-thirds had never heard of FXS before attending the medical
genetic classes (6). The first cases of FXS in China were reported 30 years ago (by Xu Bizhen and Xue Jinglun) (7), but, despite
the basic and clinical research undertaken since then, there is
still much to be done to improve diagnosis and treatment of FXS
patients. Here, we briefly review the clinical aspects of FXS in
China and provide recommendations to assist Chinese health
care professionals in the future.
State Key Laboratory of Medical Genetics, Central South University,
Changsha, Hunan, China.
Corresponding Author: [email protected]
46
Clinical Diagnosis
Epidemiological studies indicated a lower premutation frequency
among Chinese and other east Asians compared with Europeans and Africans; the prevalence of the full fragile X mutation in
China remains unknown (8).
Both boys and girls with FXS have certain physical, behavioral,
and intellectual phenotypes. Screening checklists have been applied to evaluate these physical characteristics for appropriate
referrals to genetic testing, while some of them have also been
validated to reduce unnecessary testing (9). A six-item screening
checklist for children that includes characteristics such as mental retardation, family history of mental retardation, protruding
ears, elongated face, attention deficit hyperactivity disorder, and
autistic-like behavior has proved to be a reliable and effective
diagnostic tool in two studies in China of groups with suspected
fragile X or developmental delays (10, 11). Similar characteristics
were in fact noted in a large Chinese population survey in 1986
(12). The checklist designed by Nolin and Laing has been demonstrated to not be applicable for use in China due to the age and
ethnic differences of patients used to develop the checklist (13).
Based on our clinical records, individuals affected by FXS are
most likely to visit a doctor under the following circumstances:
(1) around one to two years old when obvious developmental delay is observed, (2) between seven and nine years old,
when learning difficulties are noticed at school, (3) as teenagers who have been misdiagnosed or under-diagnosed, (4)
ages 16 and over, when family members or relatives identify a
problem.
Molecular Testing
The FMR1 gene is located on the X chromosome at Xq27.3. Approximately 95–99% of fragile X cases are caused by an increase
in the number of CGG repeats of the gene. A cytogenetic method
to count the number of fragile sites in cultured folate-deficient
cells was the first assay to be developed for diagnosis of FXS.
It has been gradually replaced by a combination of molecular
methods: CG-rich polymerase chain reaction (PCR) for accurate
sizing, together with Southern blotting to determine methylation
status (14). If standard molecular tests indicate a normal CGG
tract, sequencing of FMR1 to assess deletion and point mutation
may be performed upon request. Additional detection methods
have been developed that may be adopted as clinical tests in the
future, including methylation-sensitive PCR, methylation-specific
melting curve analysis, and methylation-specific multiplex liga-
tion-dependent probe amplification (15, 16). Certain large hospitals in China provide genetic testing services for FXS, but not all
can perform high-CG–content PCR and southern blot analysis
in parallel due to difficulties amplifying CG-rich sequences. As
a result, female carriers and those showing genetic mosaicism
may be misdiagnosed. Some commercial FXS tests have been
promoted in China such as the AmplideX™ FMR1 PCR Kit from
Asuragen, a fragile X PCR test from Abbott, and a trial product
from PerkinElmer. The cost, instrument requirements, and training needs impede the widespread use of these technologies in
the laboratory or clinic for adult and prenatal diagnosis.
Physician referrals and networks formed by FXS families are
the main channels for FXS patients and their families to advocate for testing. As of the end of 2012, most counties and cities
in China had joined the national medical insurance program,
but the cost of genetic testing for FXS is still not covered by
the major health care insurance plans. Furthermore, the cost
of the test is $300 to $600, unaffordable for some low-income
families (17).
We propose that a committee of multidisciplinary experts
should be established in China to formulate standards and
guidelines regarding diagnosis, genetic testing, and counseling for FXS. Doctors of obstetrics and gynecology specialized
in clinical genetics should be provided with periodic training in
the diagnosis and care of individuals with FXS. In addition, we
suggest that a publicly funded network should be provided to
educate and assist patients as well as provide resources to support clinics carrying out genetic testing.
REFERENCES
1. F. J. Song, P. Barton, V. Sleightholme, G. L. Yao, A. Fry-Smith, Health. Technol. Assess. 7, 1 (2003).
2. T. Wang, S. M. Bray, S. T. Warren, Curr. Opin. Genet. Dev. 22, 256 (2012).
3. S. Jacquemont et al., JAMA 291, 460 (2004).
4. D. J. Allingham-Hawkins et al., Am. J. Med. Genet. 83, 322 (1999).
5. J. L. Sun, C. Y. Ji, N. N. Xiong, Y. Zhang, J. Wang, Chin. J. Public Health 23, 140 (2007).
6. J. Li, W. Huang, S. Y. Luo, Y. T. Lin, R. H. Duan, J. Genet. Couns. Genetic Counseling
Unlike in Western countries, genetic counseling guidelines in
China have only recently been developed. In China, genetic
counselors must have completed a medical genetics specialty
and have a Master’s or doctoral degree in medicine. Physicians mainly focus on the diagnostic and clinical aspects of
the disease, but may overlook ethical issues and essential
emotional support for the patients, areas covered by genetic
counselors (18). China has some unique challenges that require careful handling, including newborn screening, carrier
screening, confidentiality issues, and reproductive options for
FXS patients.
In China, children with mild intellectual disability are eligible
for the standard nine-year compulsory education, while children
with moderate to severe intellectual disability must make use
of special mental health services. Since drug development for
FXS is still in the research phase, Chinese physicians currently
follow western conventions by treating the symptoms of the
disease; a diagnosis of FXS generally doesn’t influence the
therapies prescribed. The most common symptoms seen include
attention deficit problems, anxiety, and autism (19). Traditional
Chinese medicine offers some options for improving various
developmental and behavioral aspects in children with autism
(20), which may be exploited for use with FXS sufferers.
Epub ahead of print (2013) doi: 10.1007/s10897-013-9634-y.
7. J. L. Xue et al., Journal of Fudan University 23, 319 (1984).
8. M. K. Hill, A. D. Archibald, J. Cohen, S. A. Metcalfe, Genet. Med. 12, 396 (2010).
9. V. A. Johnson, J. Cult. Divers. 15, 117 (2008).
10. Y. Z. Guo et al., Acta Acad. Med. Sin. 22, 85 (2000).
11. Z. W. Zhu et al., paper presented at the JiangZheHu Pediatric Academic Conference, Jiaxing, China, 1 November 2012.
12. X. T. Zhou et al., Acta Genet. Sin. 13, 310 (1986).
13. Y. H. Yi, X. F. Lu, W. P. Liao, Z. Chen, X. F. Xia, J. Clin. Neurol. 15, 167 (2002).
14. A. McConkie-Rosell et al., J. Genet. Couns. 14, 249 (2005).
15. X. Y. Guo, J. Liao, F. H. Lan, Chin. J. Med. Genet. 29, 296 (2012).
16. L. Q. Wu et al., Hereditas (Beijing) 25, 123 (2003).
17. T. J. Musci, A. B. Caughey, Am. J. Obstet. Gynecol. 192, 1905 (2005).
18. S. L. Sui, Chin. Med. Ethics 26, 252 (2013).
19. J. H. Hersh, R. A. Saul, Pediatrics 127, 994 (2011).
20. V. C. Wong, J. G. Sun, J. Altern. Complement. Med. 16, 545 (2010).
Acknowledgments: This work was funded in part by grants from the
National Natural Science Foundation of China (81071028 and 81172513),
the National Basic Research “973” Program (2011CB510000 and
2012CB944600), and the Program for New Century Excellent Talents
(7603230006).
47
THERAPEUTIC STRATEGIES
Produced by the Science/AAAS Custom Publishing Office
Deep-Brain Stimulation Research in China
Bomin Sun* and Wei Liu
China has the largest single population in the world, at 1.3 billion,
accounting for 20% of the total global population. In the past 10
years, China has been advancing universal health care coverage, and a basic social medical insurance system has been developed and is expanding rapidly (1). However, due to inequity
between the urban and rural areas, bad health service utilization,
and escalation of the costs, the financial burden of health care
has rapidly increased and affordability of care for the rural population remains a challenge (2, 3).
Following the pioneering work of Benabid et al., (4) deep-brain
stimulation (DBS) has been shown to be an effective treatment
for a variety of neurological and psychotic disorders refractory to
normal treatments, including Parkinson’s disease (PD), dystonia,
tremors, and obsessive-compulsive disorder (OCD) (5–8). In the
United States, the application of DBS in essential tremor (ET),
PD, primary generalized and segmental dystonia, and OCD was
approved by the U.S. Food and Drug Administration in 1997,
2002, 2003, and 2009, respectively, while in Europe, DBS has
been approved to treat epilepsy and OCD. Many researchers in
other countries such as Japan, Korea, Brazil, and China, have
also studied the application of DBS in neurological and psychotic
disorders (9–13). In China, the first article introducing DBS was
published in Chinese in 1995 (14), and the first DBS surgery was
performed on two PD patients in 1998 (15). Since then, DBS has
rapidly gained acceptance in China as an effective treatment for
patients with certain refractory movement disorders. However,
due to the high cost of implantable hardware, this technology is
still not affordable for the majority of patients, especially those
in the rural areas. Here we review the development of DBS in
China over the past decade, and discuss the ongoing challenges
and possible solutions for the future.
The Development of DBS in China
In the decade after the first DBS surgery was performed, its use
in China underwent a period of rapid development. During this
time, more than 200 articles on DBS were published, in Chinese
as well as English, and more than 2,000 patients underwent
DBS implantations. Most of the medical centers providing DBS
treatment were located in the more developed eastern part of
China, such as Tiantan and Xuanwu Hospitals in Beijing, Ruijin
and Changhai Hospitals in Shanghai, Anhui Provincial Hospital
Department of Functional Neurosurgery, Shanghai Jiao Tong
University School of Medicine, Rui Jin Hospital, Shanghai, China
*Corresponding Author: [email protected]
48
in Anhui, and Zhujiang Hospital in Guangdong. Similar to other
countries, PD patients account for the largest DBS treatment
population, followed by dystonia and ET patients. Although China lagged behind western countries in DBS treatment, Chinese
neurosurgeons make important contributions to the development of DBS technology and application. For example, Sun et
al. introduced subthalamic nucleus (STN) stimulation in primary
dystonia, a new target of DBS, at the 2003 quadrennial meeting
hosted by the American Society for Stereotactic and Functional
Neurosurgery (16). In 2006, Zhang et al. reported on six secondary dystonia patients who received STN DBS treatment (17).
Besides movement disorders, new indications such as anorexia
nervosa (18–20) and drug psychological dependence (21, 22)
have been investigated by Chinese researchers.
Application of DBS in China
Parkinson’s Disease (PD)
As mentioned above, the largest treatment group for DBS is
PD sufferers. The targets chosen for the first two patients were
the ventrolateral thalamus (Vim) and STN. Both of the patients
showed clear improvement relative to before the surgery (15).
Although Vim DBS has been effective in controlling parkinsonian
tremors, the lack of reliable effects on other motor symptoms has
limited Vim DBS for the treatment of PD. Therefore, STN as well
as the internal globus pallidus (GPi) have been the most commonly used targets in PD patients in China. Since publication
of the first DBS study, many neurosurgeons in medical centers
across China have performed the DBS procedure in PD patients,
with satisfactory outcomes. Authors have not only explored the
mechanisms of DBS treatment, but have also shared their experiences including complications encountered, changes seen
using positron emission tomography imaging, and postoperative
management of DBS (23–25).
Dystonia
In the last century, stereotactic GPi ablation was the most commonly used method for the treatment of dystonia in China (26).
Successful application of DBS in PD patients led Chinese neurosurgeons to start using DBS for the treatment of dystonia. STN,
rather than the more commonly used GPi, is the most frequent
stimulation target for dystonia in China and outcomes are reported to be satisfactory. In 2006, Sun et al. described 15 primary
dystonia patients who received STN DBS treatment; over 76%
of their symptoms improved by the six month follow-up (27). In
the same year, Zhang et al. published on six cases of secondary
dystonia treated with STN DBS, obtaining similar results (17). In
2009, another group in China reported on 17 dystonia patients
treated with STN DBS in which the Burke-Fahn-Marsden scale
improved from 22% to 95.8% (28). More recently, Cao et al. (29)
published a long-term follow-up study that proposed the subthalamus as an alternative target to GPi for DBS to treat primary
dystonia.
Essential Tremor
Only a few journal papers have been published from China
involving DBS treatment in ET patients. Hu et al. reported
that two ET patients demonstrated significant tremor control
after STN DBS (30), while Meng et al. treated five ET patients with Vim DBS and demonstrated satisfactory tremor
control (31). Despite some success, the outcomes in ET patients treated with DBS are relatively poor compare to those
with PD.
Other Disorders
Apart from movement disorders, Chinese neurosurgeons have
been investigating new indications for DBS. For example, in a
long-term follow-up study Sun et al. reported an average of 65%
increase in body weight in four severe and refractory patients
with anorexia nervosa following a DBS procedure. To the best of
our knowledge, this is the first such report of DBS treatment in
anorexia patients.
Before the introduction of DBS, capsulotomy was the most
frequently used method for the treatment of medically refractory
OCD in China (32). DBS has now mostly replaced capsulotomy
because it has fewer complications and is reversibility. Chen et
al. reported that of six refractory OCD patients who underwent
nucleus accumbens (NAc) DBS on the right side of their brain
and anterior capsulotomy on the left side, five patients showed
obvious improved while one patient exhibited mild improvement
(33). Zhou et al. investigated white matter abnormalities in patients with treatment-resistant OCD by comparing diffusion tensor imaging (DTI) data before and after DBS procedures, and
concluded that the mechanism of DBS action in refractory OCD
may be related to ablation of the function of fibers in the frontal
lobes and cingulates fasciculus (34).
Xu et al. presented the first case report in China of NAc
DBS for the treatment of psychological opiate drug dependence (21). At the end of a three month follow-up, there is
no evidence of relapse. We obtained similar results in our
medical center and believe that a bright future exists for DBS
treatment of addiction in patients (35). NAc is now considered a promising DBS target for the treatment of addiction
(22).
The application of DBS in chronic pain and epilepsy are still in
the early stages, with only a few studies published in China (36,
37).
Problems and Strategies
DBS has been actively used in China for more than 10 years, but
there are still some challenges associated with its broader application. First, the high cost of DBS might be the chief obstacle
in China. Of the over 1.7 million people aged 55 years or older
suffering from PD in mainland China (38), many will become eligible for the surgical intervention because of intense levodopa
usage. However, only very few will receive DBS treatment due
to high out-of-pocket expenses. With the advent of new, cheaper
devices (39), it is hoped that more Chinese patients will be able
to afford this treatment in the future.
Second, inclusion criteria for patients need to be strictly enforced. We were pleased to see an expert consensus statement
on DBS treatment in PD patients in China published in 2012 (40),
which has greatly enhanced the development of DBS for PD patients. However, expert consensus is also needed for other disorders such as dystonia, OCD, and depression. The potentially
positive effects of DBS in the treatment of psychiatric disease
have not yet been recognized by most psychiatrists in China.
Therefore, greater communication and collaboration should be
encouraged between neurosurgeons and psychiatrists in the future. Finally, strict inclusion criteria for psychotic patients should
be required to ensure that excessive and unnecessary operations are not performed.
REFERENCES
1. C. Li, X. Yu, J. R. Butler, V. Yiengprugsawan, M. Yu, Soc. Sci. Med. 73, 359 (2011).
2. S. Hu et al., Lancet 372, 1846 (2008).
3. Q. Long, L. Xu, H. Bekedam, S. Tang, Int. J. Equity Health 12, 40 (2013).
4. A. L. Benabid, P. Pollak, A. Louveau, S. Henry, J. de Rougemont, Appl. Neurophysiol. 50, 344 (1987).
5. G. Deuschl et al., N. Engl. J. Med. 355, 896 (2006).
6. J. Mueller et al., Mov. Disord. 23, 131 (2008).
7. D. O’Sullivan, M. Pell, Brain Res. Bull. 78, 119 (2009).
8. W. K. Goodman, R. L. Alterman, Annu. Rev. Med. 63, 511 (2011).
9. P. R. Arantes et al., Mov. Disord. 21, 1154 (2006).
10. T. Kaido et al., Neuromodulation 14, 123 (2011).
11. S. H. Paek et al., J. Neurol. Sci. 327, 25 (2013).
12. B. Sun, S. Chen, S. Zhan, W. Le, S. E. Krahl, Acta Neurochir. Suppl. 97, 207 (2007).
13. J. Zhang et al., Acta Neurochir. Suppl. 99, 43 (2006).
14. P. Wei, Foreign Med. Sci. (Section on Neurology & Neurosurgery) (Chin.) 22, 299 (1995).
15. X. Guan, G. Luan, B, Zhang, Mod. Rehabilit. (Chin.) 5, 33 (2001).
16. B. Sun et al., paper presented at the 2003 Quadrennial Meeting held by the American Society for Stereotactic and Functional Neurosurgery, New York City, May 18–21, 2003.
17. J. Zhang, K. Zhang, Z. Wang, M. Ge, Y. Ma, Chin. Med. J. (Engl.) 119, 2069 (2006).
49
THERAPEUTIC STRATEGIES
Produced by the Science/AAAS Custom Publishing Office
18. C. Wu, Y. Pan, F. Li, B. Sun, Med. Recapi. (Chin.) 16, 2290 (2010).
33. B. Sun et al., CMINSJ 13, 58 (2008).
19. C. Wu et al., World Neurosurg. 80, 29 (2012).
34. Y. Zhou, J. Zhu, L. Li, J. Xu, Chin. J. Imaging Technol. 22, 971 20. C. Wu et al., paper presented at INS – 9th World Congress, Seoul, 35. B. Sun, W. Liu, Surgical Treatments for Drug Addictions in Humans. Korea, September, 2009.
(2006).
21. J. Xu et al., Chin. J. Stereotact. Funct. Neurosurg. (Chin.) 18, 140 Deep Brain Stimulation (Springer-Verlag, Berlin/Heidelberg,
2012).
(2005).
22. B. Sun , D. Li , C. Wu, Neuromodulation (Academic Press, London, 36. S. Liang, Chin. J. NeuRosurg. (Chin.) 28, 802 (2012).
37. L. Wang et al., Chin. J. Stereotact. Funct. Neurosurg. (Chin.) 15, 166 UK, 2009).
Figure 1. Strategies to Generate iNPCs from Somatic
Cells. iNPCs can be induced by conceptually different
mechanisms. (A) Ectopic expression of Yamanaka factors
converts somatic cells to induced pluripotent (iPS) cells that
can then be differentiated into primitive iNPCs, and subsequently into different subtypes of neurons based on differentiation conditions. (B) Ectopic expression of Yamanaka
factors converts somatic cells to primitive iNPCs through
an intermediate unstable pluripotent stage. (C) Direct reprogramming of primitive iNPCs by expression of lineagespecific transcription factors. (D) Direct reprogramming
of region-specific iNPCs by expression of lineage-specific
transcription factors and location-specific determinants. (E)
Generation of neuronal subtypes through direct conversion
of cells in vitro and in vivo. (F) Generation of neuronal progenitor subtypes through direct conversion of somatic cells
using sets of defined transcription factors.
23. B. Sun et al., Chin. J. Nerv. Ment. Dis. 29, 410 (2003).
24. Y. Ma et al., Chin. J. Neurosurg. (Chin.) 23, 452 (2007).
38. Z. Zhang et al., Lancet 365, 595 (2005).
25. C. Zuo et al., Chin. J. Nucl. Med. 27, 335 (2007).
39. Y. Ma et al., Chin. J. Neruosurg. (Chin.) 29, 415 (2013).
26. Y. Li et al., Chin. J. Neurosurg. (Chin.) 17, 350 (2001).
40. B. Chen et al., Chin. J. Neruosurg. (Chin.) 28, 855 (2012).
(2002).
27. B. Sun et al., Chin. J. Neurosurg. (Chin.) 22, 717 (2006).
28. Y. Li et al., Chin. J. Neurosurg. (Chin.) 25, 605 (2009).
Acknowledgments: We thank Shikun Zhan, Dianyou Li, Sijian Pan, and
29. C. Cao et al., Mov. Disord. 12, 1877 (2013).
Chunyan Cao for their valuable advice and assistance in the preparation
30. X. Hu et al., Shanghai Med. J. 29, 353 (2006).
of this manuscript. These studies were supported by the National Natural
31. F. Meng et al., Nat. Med. J. China 92, 1037 (2012).
Science Foundation of China (81271518) and the Natural Science
32. B. Sun et al., Chin. J. Nerv. Ment. Dis. 29, 81 (2003).
Foundation of Shanghai, China (11ZR1422600).
Neural Progenitors by Direct Reprogramming:
Strategies for the Treatment of Parkinson’s and Alzheimer’s Diseases
Changhai Tian1,2 and Jialin C. Zheng1,2,3*
Age-related neurodegenerative diseases, such as Parkinson’s
(PD) and Alzheimer′s (AD) disease, are seriously threatening the
health of Chinese citizens, with PD and AD patient populations
increasing at a rate of 100,000 and 300,000 per year, respectively (1, 2). Although there is no cure for AD or PD, a variety
of medications can provide relief from some of the symptoms.
Recently, the development of direct reprogramming technologies has shed light on promising strategies for stem cell-based
therapies for AD and PD. This review will focus on neural progenitor cells and induced neural progenitor cells studied in our
laboratory, and discuss their potential applications in AD and
PD treatment.
1
Center for Translational Neurodegeneration and Regenerative
Therapy, Shanghai Tenth People’s Hospital affiliated to Tongji
University School of Medicine, Shanghai, China
2
Department of Pharmacology & Experimental Neuroscience, and
3
Department of Pathology and Microbiology, University of Nebraska
Medical Center, Omaha, Nebraska
*Corresponding Author: [email protected]
50
Neural Progenitors from Fetal Tissues
Multipotent neural progenitor cells (NPCs) exist in the mammalian developing and adult nervous system, and are capable of
migrating toward the specific sites and giving rise to the main
components of the nervous system. Transplantation of NPCs is
a promising therapy for various neurodegenerative diseases and
brain injuries (3–5). It has been reported that the transplantation
of NPCs into the brain contributes to the improvement of cognitive impairment in animal models of PD and AD. This occurs
through cell replacement, the release of specific neurotransmitters, and the production of neurotrophic factors that protect injured neurons and promote neuronal growth (6–10). The transplantation of NPCs derived from fetal brain raises serious ethical
concerns, particularly in the acquisition of the fetal tissue, but
the understanding gained from these in vitro experiments will be
useful for the future application of NPCs obtained from non-fetal
sources. Our laboratory has been working on human NPCs for
many years and has revealed that C-X-C motif chemokine 12
(CXCL12, also known as stromal cell-derived factor 1) plays an
important role not only in increasing cell proliferation through
the Akt/Foxo3a signaling pathway, but also in protecting against
NPC apoptosis through chemokine receptor CXCR7- and CXCR4-mediated endocytotic signaling pathways (11, 12). These
data provide insight into the essential role for CXCL12 in neurogenesis and also suggests a novel role for CXCR7 in NPC survival contributing to neurogenesis, as well as potentially offering
some theoretical guidance for NPC-based therapy.
Neural Progenitor Cells Generated
by Direct Reprogramming
Selective degeneration of functional neurons is associated with
the pathogenesis of neurodegenerative disorders, such as degeneration of midbrain dopaminergic neurons in PD (13) and
forebrain cholinergic neurons in AD (14). How to achieve sufficient cell replacement to halt PD or AD progression, or possibly even provide a cure, is the main challenge. The discovery
of induced pluripotent stem (iPS) cells has facilitated the derivation of stem cells from adult somatic cells for the personalized treatment of PD and AD without depending on fetal tissues. However, ethical and safety concerns still exist. Recently,
neuronal subtypes including dopaminergic and cholinergic
neurons have been generated successfully through direct reprogramming of somatic cells by expression of developmental
genes (15–17). The low yield of neurons from this method has,
however, limited its broad application in cell transplantation. In
addition to NPCs obtained from iPS cell differentiation (Fig. 1,
path A) and neuronal subtypes induced by direct reprogramming (Fig. 1, path E), the direct reprogramming of NPCs from
differentiated, non-neuronal somatic cells (Fig. 1, paths B and
C) (18–21) will not only provide an alternative to fetal tissue or
pluripotent cells as precursors, but also furnish a potentially unlimited source of neurons. Our laboratory was among the first to
successfully convert fibroblasts into induced NPCs (iNPCs) by
ectopic expression of transcription factors. iNPCs share many
characteristics with primary NPCs and are able to differentiate
into neurons (Fig. 2). Recently, the direct reprogramming of iNPCs has also been achieved by forced expression of a single
factor (22) or by 3-D sphere culture of fibroblasts on low attachment surfaces (23). These findings suggest that neural progenitor cell fate can be reprogrammed by modifying intrinsic and
extrinsic cues.
Stem cell-based therapy for AD and PD is essentially based
on the regeneration of different neuronal subtypes, such as dopaminergic neurons in PD and forebrain cholinergic neurons in
AD. Recently, we have been working on the direct reprogramming of somatic cells into region-specific iNPCs (Fig. 1, path
D) as well as subtype-specific iNPCs (Fig. 1, path F) by over-
51
THERAPEUTIC STRATEGIES
Produced by the Science/AAAS Custom Publishing Office
Figure 2. Direct Conversion of Fibroblasts into NPCs by a Novel Combination of Transcription Factors.
(A) Schematic showing direct reprogramming of fibroblasts derived from
E/Nestin:EGFP transgenic mice into
NPCs using defined transcription factors including Brn2, Sox2, Bmi1, TLX,
and c-Myc. (B) Differentiation of iNPCs
into neurons (panel a; MAP-2, red),
astrocytes (panel b; GFAP, red), oligodendrocytes (panel c; O4, red). Panel
d shows synapse formation between
neurons (β-III Tubulin, green; Synaptophysin, red), with magnification of
white box in d shown in panel e. Nuclei are stained with DAPI (blue). EGFP,
enhanced green fluorescent protein;
MAP-2, microtubule-associated protein 2; GFAP, glial fibrillary acidic protein; O4, marker for oligodendrocytes.
expression of defined growth factors. Both of these pathways
show promise for AD and PD therapies. Direct in vivo conversion of somatic cells, such as fibroblasts and astrocytes, into
functional neurons has also been achieved (24), providing proof
of principle that it may be feasible to convert somatic cells—
such as activated astrocytes—into region- or subtype-specific
iNPCs in the brains of AD and PD patients. In the future, the
further development of this technology will undoubtedly provide promising strategies for the effective treatment of AD
and PD.
REFERENCES
12. B. Zhu et al., Stem Cells 30, 2571 (2012).
13. W. Dauer, S. Przedborski, Neuron 39, 889 (2003).
14. D. S. Auld, T. J. Kornecook, S. Bastianetto, R. Quirion, Prog. Neurobiol. 68, 209 (2002).
15. M. Caiazzo et al., Nature 476, 224 (2011).
16. U. Pfisterer et al., Proc. Natl. Acad. Sci. U.S.A. 108, 10343 (2011).
17. M. L. Liu et al., Nat. Commun. 4, 2183 (2013).
18. C. Tian et al., Curr. Mol. Med. 12, 126 (2012).
19. E. Lujan, S. Chanda, H. Ahlenius, T. C. Sudhof, M. Wernig, Proc. Natl. Acad. Sci. U.S.A. 109, 2527 (2012).
2. J. Wang, Asia-Pacific Traditional Medicine (Chin.) 7, 157 (2011).
21. D. W. Han et al., Cell Stem Cell 10, 465 (2012).
3. A. Bjorklund, O. Lindvall, Nat. Neurosci. 3, 537 (2000).
22. K. L. Ring et al., Cell Stem Cell 11, 100 (2012).
4. O. Lindvall, Z. Kokaia, Nature 441, 1094 (2006).
23. G. Su et al., Biomaterials 34, 5897 (2013).
5. J. K. Ryu, T. Cho, Y. T. Wang, J. G. McLarnon, J. Neuroinflammation 24. O. Torper et al., Proc. Natl. Acad. Sci. U.S.A. 110, 7038 (2013).
6, 39 (2009).
(2009).
(2007).
Acknowledgments: This work was supported in part by research grants
from the National Institutes of Health (R01 NS 41858-01 and R01 NS
Jun Liu and Sheng-Di Chen*
Parkinson’s disease (PD), the second most common neurodegenerative disease, is characterized by progressive loss of dopaminergic neurons in the substantia nigra pars compacta and
formation of intracytoplasmic Lewy inclusion bodies (1). Its cardinal clinical features are resting tremor, rigidity, bradykinesia,
and postural instability. The most effective approach currently
for managing PD patients is utilizing a dopamine replacement
strategy. Even though they provide considerable benefit during
the early stages of the disease, these treatments exhibit drawbacks with long-term use. Most critically, they are thought to be
only symptomatically effective, not disease-modifying. We have
therefore explored four avenues to provide favorable changes
in dopaminergic cell loss and neurologic symptoms in PD models: traditional Chinese herbs, small peptides, synthesized compounds, and stem cell therapies. Results have been encouraging, showing neuroprotective effects in vitro and in vivo through
a variety of different molecular pathways. It is hoped that these
agents may provide promising new drug candidates for PD intervention by blocking the ongoing degeneration of dopaminergic
neurons and slowing the disease progression.
Traditional Herb-based Treatments
1. Curcumin
Curcumin, isolated from the roots of the turmeric plant (Curcuma
longa), is a natural polyphenol and the primary active component
of turmeric, a widely used spice and the major constituent of curry
powder in South and Southeast Asian cuisine, particularly in Indian. Over the past several years, curcumin has been extensively
studied for its neuroprotective effects and therapeutic potential in
both in vitro and in vivo models of PD. Our laboratory and other
investigators have demonstrated that curcumin can modulate
multiple processes, including antioxidation, antiapoptosis, antineuroinflammation, and antiaggregation of proteins (2).
In Vitro Studies
The antioxidant and anti-inflammatory effects of curcumin in
6-hydroxydopamine (6-OHDA)-treated MES23.5 cells have
061642-01), the State of Nebraska (DHHS-LB606 Stem Cell 2009-10 to
been reported (3). Curcumin counteracts oxidative stress by not
only downregulating reactive oxygen species (ROS), but also
by upregulating the antioxidant enzyme superoxide dismutase
(SOD). It alleviates inflammation and mitochondrial dynamic
dysfunction by suppressing the activation and nuclear translocation of nuclear factor кB (NF-кB) or inhibiting extracellular
signal-regulated kinase (ERK) (4). Moreover, Chen et al. demonstrated that curcumin possesses antiapoptotic properties, reducing 1-methyl-4-phenylpyridinium ions (MPP+)-induced PC12 cell
death mediated via the Bcl-2/mitochondrial/ROS/iNOS (inducible
nitric oxide synthase) pathways (5). Our study demonstrated the
neuroprotective effects of curcumin in SH-SY5Y cells exposed
to α-synuclein and its efficient inhibition of α-synuclein fibril accumulation through downregulation of mTOR/p70S6K signaling,
which is a classical suppressive pathway of autophagy and recovery of macroautophagy (6).
In Vivo Studies
Some studies have evaluated the neuroprotective value of curcumin against both 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine
(MPTP)- and 6-OHDA-induced dopaminergic degeneration in
rodents. Recent experiments indicated that intraperitoneally administered curcumin could protect dopaminergic neurons from
apoptosis in a mouse model of PD (7). Our studies showed that
curcumin inhibited MPTP-induced hyperphosphorylation of c-Jun
N terminal Kinase (JNK), especially JNK3, and its downstream
molecules, which contributed to the prevention of mitochondriamediated apoptosis (8). In addition, our laboratory has provided
evidence that a reduction in ROS levels and the inhibition of microglial activation play important roles in curcumin’s neuroprotective effect in MPTP-treated PD mice (9). Furthermore, others
have demonstrated that intravenous and oral curcumin exhibited
protective effects against dopaminergic neurotoxicity induced by
MPTP- or 6-OHDA in mice (10, 11). Dietary supplementation with
curcumin could protect dopaminergic neurons against MPTP-induced neurodegeneration through inhibition of protein nitration,
elevation of glutathione levels, and by boosting mitochondrial
complex I activity (12). Recent studies showed that curcumin improved motor function in a mouse model of PD (13).
J.Z. and LB606 Stem Cell 2010-10 to S.D. and C.T.), the China National
8. S. Wu et al., Pathobiology 75, 186 (2008).
Science and Technology major project (2014CB965001), and the National
9. T. Yasuhara et al., J. Neurosci. 26, 12497 (2006).
Natural Science Foundation of China (81028007 and 81329002 to J.Z.,
10. L. Madhavan, B. F. Daley, K. L. Paumier, T. J. Collier, J. Comp. 52
11. Y. Wu et al., J. Neurochem. 109, 1157 (2009).
20. M. Thier et al., Cell Stem Cell 10, 473 (2012).
7. D. E. Redmond, Jr. et al., Proc. Natl. Acad. Sci. U.S.A. 104, 12175 Neurol. 515, 102 (2009).
1. J. Xian, Public Health (Chin.) 36, 25 (2013).
6. M. Blurton-Jones et al., Proc. Natl. Acad. Sci. U.S.A. 106, 13594 Potential Neuroprotective Therapies for
Parkinson’s Disease
and 81271419 to C.T.).
Department of Neurology & Institute of Neurology, Rui Jin Hospital
affiliated to Shanghai Jiao Tong University School of Medicine,
Shanghai, China
*Corresponding Author: [email protected]
2. Salidroside
Salidroside, a remedy used in Tibet, is a phenylpropanoid glycoside isolated from Rhodiola rosea L. A recent study indicated that
53
THERAPEUTIC STRATEGIES
Produced by the Science/AAAS Custom Publishing Office
salidroside played antioxidant and antiapoptotic roles, protecting PC12 cells in culture against MPP+ toxicity by inhibiting the
nitric oxide pathway (14). Our work has shown that salidroside
could partially block the loss of dopaminergic neurons induced
by MPTP in a PD mouse model. The mechanism of action involved activation of the PI3K/Akt and PI3K/protein kinase B
signaling pathways (15) and may be related to the elevated secretion of endogenous glial cell line-derived neurotrophic factor
(GDNF) (16).
3. Tripchlorolide
Tripchlorolide (TW397), an analogue of triptolide (T10) and one
of the active ingredients from Tripterygium wilfordii Hook F.—a
traditional Chinese herb widely used in the amelioration of a host
of inflammatory and autoimmune diseases—possesses potent
anti-inflammatory and immunosuppressive properties (17). Li et
al. reported that TW397 exerted a neuroprotective effect against
dopaminergic lesions and restored their function after neurotoxicity induced by MPP+ in vitro (18). Our own study demonstrated
a neuroprotective action of TW397 in vivo in the MPTP-lesioned
PD mouse model by suppression of the astroglial response within the striatum. We found that intraperitoneal injection of TW397
not only dramatically attenuated the death of tyrosine hydroxylase-immunoreactive (TH-IR) neurons in the substantia nigra
pars compacta and TH-IR fibers in the striatum after MPTP treatment, but also significantly improved the level of dopamine in the
substantia nigra and striatum. In addition, intriguingly, in MPTPtreated animals the locomotive performance of those treated with
TW397 improved significantly (19).
Artificial Neuroprotective Compounds
Our previous studies have indicated that apoptosis mediated by
the JNK signaling pathway plays an important role in animal models with MPTP- or 6-OHDA–induced PD. We have since reported
the inhibitory effect of a synthesized compound, K252a, on JNKs
activation, and also generated a novel small peptide, Tat-JBD,
comprised of the membrane transduction domain (Tat) of human
immunodeficiency virus-type 1 (HIV-1) plus the sequence of an
11-amino acid peptide corresponds to the JNK-binding domain
(JBD) found in JNK-interacting protein-1 (JIP-1) (20). Tat-JBD
interfered with JIP-1-JNK complex formation, suppressed the activation of JNK induced by MPTP, and inhibited the downstream
apoptosis cascade.
Overall, the loss of dopaminergic neurons in the substantia
nigra and their axons in the striatum was significantly minimized
by administration of Tat-JBD in MPTP-treated mice.
K252a has numerous modes of action, including inhibition of
Trk, MLK3, and apoptosis-inducing kinase 1 (ASK1), as well as
activation of mitogen-activated protein kinase signaling pathways
through its interaction with multiple molecular targets. It has also
been shown that K252a protects dopaminergic neurons against
54
6-OHDA-induced cell death, and partially improved the behavioral asymmetry in 6-OHDA-treated rats (20, 21). Moreover, we
found that the antioxidant N-acetylcysteine and the N-methyl-Daspartate receptor antagonist ketamine also exerted a preventive effect on MPTP-induced loss of dopaminergic neurons by
suppressing the nuclear translocation and activation of JNK3,
the only neural-specific isoform of JNK (22).
In summary, these findings suggest that Tat-JBD and K252a
possess neuroprotective properties following MPTP or 6-OHDA
insult through inhibition of the JNK signaling pathway.
Stem Cell Therapy
In previous work, we have explored the curative potential of neural stem cells (NSCs) and bone marrow stromal cells (BMSCs) in
PD mouse models (23, 24). Genetically engineered neural stem
cells (NSCs) have been found to be a promising vector for delivery of neurotrophic factors, which has been used to treat various
nervous system disorders (25). We genetically engineered NSCs
to delivery neurturin (NTN), a GDNF family member, into the
brains of a rat model of PD and demonstrated that it promoted
the survival of embryonic dopaminergic neurons. These specialized NSCs not only engrafted and integrated in the host striatum
successfully, but also differentiated into neurons, astrocytes, and
oligodendrocytes with stable, high-level NTN expression. Our
study revealed that NTN expression decreased the loss of dopaminergic neurons in the substantia nigra and improved motor
symptoms for at least four months after the NSC graft. These
findings suggested that transplantation of NTN-secreting NSCs
exerted a protective effect in a PD rat model (23).
BMSCs, non-hemopoietic stem cells found in adult mammals
or in human bone marrow, possess several advantages over
stem cells—abundant availability, and ease of preparation, isolation, purification, and culture—as well as trans-differentiation
properties that allow BMSCs to differentiate into various types
of cells under certain conditions, including neuron- and glia-like
cells (26, 27), providing a new and reliable source for cell substitution and gene transfection. One study revealed that differentiated BMSCs contributed to higher dopamine levels in the lesioned striatum and significant behavioral improvement over five
months (28). BMSCs containing NTN also significantly attenuated behavioral symptoms in 6-OHDA–lesioned rats. Interestingly,
in contrast to NSCs transfected with NTN genes, the mechanism
of action of NTN-expressing BMSCs appeared to involve functional enhancement of residual dopaminergic neurons by NTN,
rather than an increase in the numbers of existing dopaminergic
neurons (29). These results may help us to further evaluate the
neuroprotective value of BMSC transplantation in PD treatment.
Moreover, it is well known that regulation of the immune response after grafting is a critical issue in PD transplantation
(30). Our work found that NSCs transfected with the interleukin-10 gene, an important regulator of the immune system,
could downregulate intracerebral cellular and humoral immune
responses in 6-OHDA-lesioned PD rats (31). Furthermore, selenite—the source of the essential micronutrient selenium—intraperitoneally injected into 6-OHDA–lesioned rats before the
transplantation of DA neurons derived from embryonic stem
cells, could inhibit the expression of proinflammatory factors
such as TNFα and iNOS. It also improved the survival of implanted neurons and the rotational behavior in the recipient
animals (32).
In summary, the agents above have been shown to be neuroprotective in both in vitro and in vivo experiments, and may slow down
disease progression and provide new therapeutic approaches
for PD. However, some challenges still lie ahead. More precise
knowledge is needed of the signaling pathways responsible for
neuronal death and the mechanisms underlying the protective
effects. More critically, the effectiveness of these potential therapies will need to be confirmed in large scale clinical trials with PD
patients.
11. V. Zbarsky et al., Free Radic. Res. 39, 1119 (2005).
12. R. B. Mythri, J. Veena, G. Harish, R. B. S. Shankaranarayana, B. M. M. Srinivas, Br. J. Nutr. 106, 63 (2011).
13. W. Tripanichkul, E. O. Jaroensuppaperch, Int. J. Neurosci. 122, 263 (2012).
14. X. Li et al., Brain Res. 1382, 9 (2011).
15. Y. H. Zhang et al., J. Clin. Neurol. 21, 133 (2008).
16. Y. H. Zhang et al., Chin. J. Neurol. 39, 540 (2006).
17. W. Z. Gu, S. R. Brandwein, Int. J. Immunopharmacol. 20, 389 (1998).
18. F. Q. Li et al., Exp. Neurol. 179, 28 (2003).
19. Z. Hong et al., Eur. J. Neurosci. 26,1500 (2007).
20. J. Pan et al., Lab. Invest. 90, 156 (2010).
21. J. Pan et al., Mol. Pharmacol. 72, 1607 (2007).
22. J. Pan et al., Neurochem. Int. 54, 418 (2009).
23. W. G. Liu, G. Q. Lu, B. Li, S. D. Chen, Parkinsonism Relat. Disord. 13, 77 (2007).
24. M. Ye et al., Parkinsonism Relat. Disord. 13, 44 (2007).
25. C. Chen, Y. Wang, G. Y. Yang, Curr. Drug Targets 14, 81 (2013).
26. B. A. Horger et al., J. Neurosci. 18, 4929 (1998).
REFERENCES
2. A. Jitoe-Masuda, A. Fujimoto, T. Masuda, Curr. Pharm. Des. 19, 28. J. Sanchez-Ramos et al., Exp. Neurol. 164, 247 (2000).
2084 (2013).
3. J. Wang, X. X. Du, H. Jiang, J. X. Xie, Biochem. Pharmacol. 78, 178 27. J. Sanchez-Ramos, F. Cardozo-Pelaez, S. Song, Mov. Disord. 13S, 1. K. Jellinger, Movement Disord. 2, 124 (1987).
(2009).
122 (1998).
29. M. Ye et al., Brain Res. 1142, 206 (2007).
30. K. E. Soderstrom et al., Neurobiol. Dis. 32, 229 (2008).
31. X. J. Wang, W. G. Liu, Y. H. Zhang, G. Q. Lu, S. D. Chen, Neurosci. Lett. 423, 95 (2007).
4. Y. X. Gui et al., Neurobiol. Aging 33, 2841 (2012).
5. J. Chen et al., Apoptosis 11, 943 (2006).
32. L. P. Tian et al., Curr. Mol. Med. 12, 1005 (2012).
6. T. F. Jiang et al., J. Neuroimmune Pharmacol. 8, 356 (2013).
7. Z. Mansouri, M. Sabetkasaei, F. Moradi, F. Masoudnia, A. Ataie, J. Mol. Neurosci. 47, 234 (2012).
Acknowledgments: These studies were supported by the National
Program of Basic Research of China (2010CB945200 and
8. J. Pan et al., Translational Neurodegeneration 1, 16 (2012).
2011CB504104), the Natural Science Fund (30971031, 81129018,
9. J. Pan, J. Q. Ding, S. D. Chen, Chin. J. Contemp. Neurol. Neurosurg. and 81371407), the Shanghai Key Project of Basic Science Research
7, 421 (2007).
10. O. Vajragupta et al., Free Radic. Biol. Med. 35, 1632 (2003).
(10411954500), and the Program for Outstanding Medical Academic
Leader (LJ 06003).
55
THERAPEUTIC STRATEGIES
Produced by the Science/AAAS Custom Publishing Office
Exploring Therapeutic Mechanisms of Traditional
Chinese Medicine Extracts and Electroacupuncture
in Parkinson’s Disease
Yan Zheng, Jun Jia, and Xiao-Min Wang*
Introduction
Parkinson’s disease (PD) is a common neurodegenerative disorder of the central nervous system (CNS), found particularly in the
elderly and affecting approximately two million people in China
(1, 2). It is characterized by the loss of dopaminergic (DAergic)
neurons in the substantia nigra of the midbrain and the subsequent decrease in dopamine (DA) levels in the striatum. The
imbalance in the substantia nigra-striatum basal ganglia neural
signaling resulting from DA insufficiency can be corrected with
DA replacement treatments such as levodopa (L-dopa). However, severe side effects result from long-term administration. The
complicated pathogenesis of PD makes the development of remedies difficult, driving investigators to continually look for novel
treatments. China is fortunate to be able to draw on abundant
traditional Chinese medicine (TCM) resources, including herbal
medicines and acupuncture, which are increasingly being recognized as providing potentially effective treatments. Our research
into natural extracts and electroacupuncture (EA) targeting PD
pathogenesis have shown therapeutic potency. However, their
mechanisms of action are not fully understood, creating a barrier to the popular acceptance of TCM treatments. For this reason, we investigated the therapeutic mechanisms and efficacy of
these therapies.
56
(MPTP)/1-methyl-4-phenylpyridinium ion (MPP+) or 6-hydroxydopamine (6-OHDA), and an inflammation model simulated by
microglia activation using lipopolysaccharides (LPS).
Natural Traditional Chinese Medicine Extracts
Diverse factors are involved in the loss of DAergic neurons in
PD, among which inflammation and oxidative stress play critical roles. They interact with each other to activate and drive the
course of neuronal death. Three natural TCM extracts have been
discovered that target these pathways: triptolide, fucoidan, and
tenuigenin (Chinese patents ZL00107779.1, ZL200710099008.8,
and ZL 201010200579.8, respectively). The therapeutic and neuroprotective effects of these compounds have been confirmed
in vitro and in rodent PD models, including an oxidative stress
model induced by 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine
Triptolide
Triptolide (T10) is the major active component extracted from
Tripterygium wilfordii Hook. F. (TWHF) and has been used to
treat inflammatory diseases for centuries in China. Our data
showed for the first time that T10 and tripchlorolide, an analogue
of T10, prevented the death of DAergic neurons in neurotoxic lesions induced by MPP+ or LPS (3, 4). In primary microglial cultures, T10 significantly suppressed the activation of microglia
and the release of interleukin-1β (IL-1β), tumor necrosis factor
α (TNFα), and nitric oxide (NO) from LPS-activated microglia
(5), indicating that T10 exhibited neuroprotective effects by reducing inflammation. As expected, in LPS- or MPP+-induced
rodent PD models, the protective effect of T10 on DAergic neurons and the inhibition of microglial activation were detected
simultaneously, contributing to behavioral improvement (6, 7).
Two parallel signaling pathways may underlie the suppression
effect of T10 on microglial activation: T10 inhibited prostaglandin
E2 (PGE2) expression by blocking the phosphorylation of the
c-jun NH2-terminal kinase (JNK); T10 reduced the transcriptional
activity of NF-kB, which in turn down-regulated cyclooxygenase-2 (COX2) expression and PGE2 release (8). In addition
to impacting microglia, T10 promoted the synthesis and release
of nerve growth factor (NGF) from primary astrocyte cultures.
Surprisingly, it had no effect on brain-derived neurotrophic factor (BDNF) and glial cell-derived neurotrophic factor (GDNF)
levels (9). Tripchlorolide, however, promoted BDNF gene expression in astrocytes and stimulated the neurite growth of
DAergic neurons (3). This evidence suggests that T10 and its
derivatives exert their neuroprotective function in a number of
ways, especially through anti-inflammatory and neurotrophic
effects (10).
Department of Neurobiology, Department of Physiology, Capital
Medical University, Key Laboratory for Neurodegenerative Disorders
of the Ministry of Education, Beijing Institute for Brain Disorders,
Beijing, China
*
Corresponding Author: [email protected]
Fucoidan and Tenuigenin
Fucoidan is a sulfated polysaccharide from edible brown seaweeds, while Tenuigenin (TEN) is an active compound found
in polygala tenuifolia root extracts. Our group first showed their
value as a potential PD therapies by demonstrating that they
appeared to prevent motor deficits, alleviate apomorphine-
induced abnormal rotation, and
attenuate DAergic neuron loss
and decreases in DA levels in
an MPTP-induced mouse model
and LPS-induced rat model of PD
(11–13). The results indicate that
these compounds likely have an
anti-inflammatory mode of action
in addition to the anti-oxidative
effects reported previously. In
vitro studies revealed that fucoidan blocked NO synthesis (14)
and TNFα release (12)—both
essential to the inflammatory response—and, interestingly, also
dramatically reduced the production of reactive oxygen species
(ROS) by activated microglia
(12), implying that fucoidan might
impede the self-propagating cyFigure 1. Possible therapeutic mechanisms of traditional Chinese medicine extracts in Parkinson’s disease. Neucle of neurodegenerative progrotoxins (MPP+/LPS/6-OHDA) activate microglia to release pro-inflammatory factors (PGE2/IL-1β/TNFα) and ROS, and
ress in PD since ROS acts as a
also inhibit neurotrophic activity of astrocytes, both of which result in DAergic neuron degeneration. Some of these neurotoxins can also be transported into the cytosol through a dopamine transporter (DAT) located on the membrane of DAebridge between inflammation and
rgic neurons, initiating oxidative stress. TCM extracts may inhibit neuroinflammation and oxidative stress (1), stimulate
oxidative stress. As for TEN, it
the production and release of neurotrophic factors from astrocytes (2), prevent neuronal damage from neurotoxicity (3),
stabilized mitochondrial memand downregulate α-synuclein hyperphosphorylation (4). DA, dopamine; MPP+, 1-methyl-4-phenylpyridinium ion; LPS,
lipopolysaccharide; 6-OHDA, 6-hydroxydopamine; PGE2, prostaglandin E2; IL-1β, interleukin-1β; TNFα, tumor necrosis
brane potential and decreased
factor α; ROS, reactive oxygen species.
neurotoxic damage through
increased
antioxidant
expression (15). Recently, it has been observed that TEN at- forebrain bundles (MFBs) axotomy) rodent PD model (19–22)
tenuated phosphorylation of α-synuclein, a major com- and in rodents suffering 6-OHDA- (23, 24) or MPTP-induced
ponent of Lewy bodies in degenerating DAergic neurons, damage (25).
thereby protecting neurons from toxicity induced by α-synuclein
overexpression (16).
Neuroprotective Effects of
Taken together, these promising PD treatments exhibited Electroacupuncture
protective effects through multiple mechanisms including anti- Previous investigations to elucidate acupuncture mechanisms
inflammation, anti-oxidation, neurotrophy, and impacting post- of action have hinted at a neuroprotective role in animal modtranslational protein modifications (Fig. 1).
els of PD that functions through various pathways in the substantia nigra and striatum (26). Similarly, in rodents suffering
MPTP (27, 28), 6-OHDA (24), or MFB insult (19, 21), EA inElectroacupuncture for Symptom Alleviation
hibited the loss of DAergic neurons, possibly attributable to
in Parkinson’s Disease
Acupuncture is an ancient TCM technology that has been ap- EA-enhanced expression of GDNF (20), BDNF (29), or tyroplied to alleviate the symptoms of PD patients and improve sine kinase B (the BDNF receptor) (24, 30). It has been also
their quality of life (17, 18); however, its therapeutic mode of ac- determined that EA significantly attenuated the activation of
tion remains unconfirmed. To clarify these issues, electroacu- microglia, reduced the levels of TNFα and IL-1β mRNA in venpuncture (EA) was developed by adding a pulsed electric tral midbrains (20), and limited the upregulation of COX-2 and
current through the needles in place of manual manipulation inducible nitric oxide synthase (iNOS) (28). Additionally, EA ef(2). One of the big advantages is that the stimulation param- fectively inhibited the oxidative stress response and increased
eters—such as frequency, pulse width, intensity, and period the activity of antioxidants in the striatum (27). These results inof treatment can be controlled precisely. It has been demon- dicate that EA could serve as a neuroprotective intervention by
strated that high-frequency EA (100 Hz) can significantly miti- inhibiting the inflammatory response and restoring the oxidantgate motor deficits in a mechanically-damaged (by medial antioxidant balance.
57
THERAPEUTIC STRATEGIES
Produced by the Science/AAAS Custom Publishing Office
Non-dopamine Approach for Therapeutic Effects
of Electroacupuncture
Contrary to current thinking that increasing DA is crucial for PD
therapy, data from our group and others (31) have shown that the
efficacy of EA was independent of DA levels in the striatum (21,
22, 29, 32). This implies that a non-DA–dependent neuroprotective pathway exists. EA was able to modulate neurotransmitter
stability and reverse the neurotransmitter imbalance caused by
DA depletion. EA impacted many neurotransmitters in basal ganglia circuitry, including neurotrophic factors (19, 29), substance
P (21), g-aminobutyric acid (GABA) (22), glutamate, and acetylcholine (32), and also affected synaptic plasticity (31). Taken
together, these studies suggest that restoration of homeostasis
in the basal ganglion circuitry was the predominant therapeutic
effect generated by EA.
Genome-wide analysis was used to study the gene expression of those molecules regulated by EA. In a mouse model of
PD caused by MPTP, EA significantly shifted gene expression
profiles, and comparative analyses suggested that EA improved
motor abnormalities by correcting imbalances in the expression
of functional gene clusters in both the cortex and striatum (24).
In summary, high-frequency EA stimulation could effectively alleviate behavioral disorders in PD animal models. Further work
to elucidate EA mechanisms will focus on understanding how
neuronal network balance can be restored in the basal ganglion
circuitry through the application of high throughput assays and
functional imaging technologies. TCM extracts and EA are clearly promising prospects for PD therapy, even though details of
their therapeutic mechanism of action remain to be clarified. It
is prudent to acknowledge that they may not completely replace
treatments such as L-dopa, but rather may be useful for prolonging the effect of current therapies and alleviating the side effects
of those treatments.
4. F. Q. Li et al., J. Neuroimmunol. 148, 24 (2004).
5. H. F. Zhou et al., NeuroReport 14, 1091 (2003).
6. H. F. Zhou et al., Neurobiol. Dis. 18, 441 (2005).
Novel Therapeutic Strategies for Amyotrophic
Lateral Sclerosis
7. J. P. Gao, S. Sun, W. W. Li, Y. P. Chen, D. F. Cai, Neurosci. Bull. 24, 133 (2008).
Weidong Le1* and Xiaojie Zhang2
8. Y. Gong, B. Xue, J. Jiao, L. Jing, X. Wang, J. Neurochem. 107, 779 (2008).
9. B. Xue et al., Neurochem. Res. 32, 1113 (2007).
10. Y. Zheng, W. J. Zhang, X. M. Wang, CNS Neurosci. Ther. 19, 76 (2013).
11. D. Luo et al., Eur. J. Pharmacol. 617, 33 (2009).
12. Y. Q. Cui, Y. J. Jia, T. Zhang, Q. B. Zhang, X. M. Wang, CNS Neurosci. Ther. 18, 827 (2012).
13. H. L. Yuan et al., CNS Neurosci. Ther. 18, 584 (2012).
14. Y. Q. Cui et al., Clin. Exp. Pharmacol. Physiol. 37, 422 (2010).
15. Z. Liang et al., Neurosci. Lett. 497, 104 (2011).
16. J. X. Zhou et al., CNS Neurosci. Ther. 19, 688 (2013).
17. A. A. Rabinstein, L. M. Shulman, Neurologist 9, 137 (2003).
18. L. M. Shulman et al., Mov. Disord. 17, 799 (2002).
19. X. B. Liang et al., NeuroReport 14, 1177 (2003).
20. X. Y. Liu et al., Exp. Neurol. 189, 189 (2004).
21. J. Jia et al., Behav. Brain Res. 205, 214 (2009).
22. J. Jia et al., Behav. Neurosci. 124, 305 (2010).
23. H. J. Park et al., Exp. Neurol. 180, 93 (2003).
24. L. R. Huo et al., Evid. Based Complement Alternat. Med. 2012, 908439 (2012).
25. H. Wang et al., PLOS ONE 8, e64403 (2013).
26. T. H. Joh, H. J. Park, S. N. Kim, H. Lee, Neurol. Res. 32, 5 (2010).
27. H. Wang et al., PLOS ONE 6, e19790 (2011).
28. J. M. Kang et al., Brain Res. 1131, 211 (2007).
29. X. B. Liang et al., Mol. Brain Res. 108, 51 (2002).
30. Y. K. Kim et al., Neurosci. Lett. 384, 133 (2005).
31. S. N. Kim et al., PLOS ONE 6, e27566 (2011).
32. Z. L. Sun et al., Neurosci. Lett. 520, 32 (2012).
REFERENCES
1. Z. X. Zhang et al., Lancet 365, 959 (2005).
Acknowledgments: This work was supported by the National Basic
2. X. Wang et al., Neurochem. Res. 33, 1956 (2008).
Research “973” Program (2006CB500700 and 2011CB504100) and the
3. F. Q. Li et al., Exp. Neurol. 179, 28 (2003).
National Natural Science Foundation of China (81030062 and 81072858).
Amyotrophic lateral sclerosis (ALS)
is a devastating, adult-onset neurodegenerative disease. The pathological hallmark of ALS is the selective death of motor neurons (MNs) in
the spinal cord and brain (1, 2). The
only drug approved by the U.S. Food
and Drug Administration (FDA) for
ALS treatment is riluzole, but it has
limited impact on disease symptoms
and patients survival (3). Although the
underlying causes of ALS have not
been fully defined, progress has been
made in recent years, which has led
to the development of several neuroprotective agents and potential therapeutic strategies for treatment. Our
recent work has shown that selective
degeneration of MNs is usually associated with impairment of the autophagy pathway, which appears to play a
critical role in the pathogenic process
leading to neuron degeneration (4, 5).
We have also reported on a range of
biological effects using mammalian
target of rapamycin (mTOR)-dependent and independent autophagic
activators in ALS mice (4, 5). Furthermore, we and others have found
that ALS patients have higher levels
of homocysteine (Hcy) in their blood
and spinal fluid, suggesting that Hcy
may be involved in the pathogenesis
of ALS and that Hcy-lowering treatments might have therapeutic potential (6, 7). We present hereby our
findings of autophagy and Hcy in ALS
(summarized in Fig. 1).
Figure 1. Novel findings and possible mechanisms of action of autophagy and Hcy in ALS. Small molecule
autophagy enhancers induce autophagy activation via mTOR-dependent or independent pathways in motor neurons
(MNs). If the autophagic flux is intact, the activated autophagy shows neuroprotective effects by enhancing the
degradation of toxic protein aggregates. However, if the autophagic flux is defective, the enhanced autophagy might
be harmful to the MNs due to accumulated toxic protein aggregates. Rescue of the impaired autophagic flux can
induce protein degradation and improve MNs survival. Low levels of folic acid and 5-methyltetrahydrofolate (5MTHF), as well as other unknown factors, may contribute to elevated homocysteine (Hcy) levels in ALS. Hcy-lowering
therapies such as folic acid, B12, or GOS-rich prebiotic yogurt show neuroprotective effects mediated by lowering
Hcy in ALS mice.
1 Affiliated Hospital, Dalian Medical University, Dalian, China
Ruijin Hospital, Shanghai 2nd Medical University, Shanghai, China
*
Corresponding Author: [email protected]
1 st
2
58
Autophagy as Therapeutic Target for ALS
Protein aggregation and inclusion body formation have been recognized as common cellular and molecular processes in many
neurodegenerative diseases, including Alzheimer’s disease
(AD), Parkinson’s disease (PD), Huntington’s disease (HD), and
ALS (8). Postmortem neuropathological studies have revealed
the increased number of autophagosomes (AVs) in MNs of
sporadic ALS (SALS) and familial ALS (FALS) patients (4, 9).
In mammals, the predominant route for the degradation of long-
59
THERAPEUTIC STRATEGIES
Produced by the Science/AAAS Custom Publishing Office
lived proteins, aggregated proteins, and injured organelles is
through the autophagosome-lysosome pathway. We investigated the activation status of autophagy in transgenic SOD1-G93A
mice (hereafter referred to as ALS mice, which carry a G93A
mutation in the superoxide dismutase 1 gene), demonstrating
for the first time that the level of autophagy marker microtubuleassociated protein 1 light chain 3-II (LC3-II) is markedly and specifically increased in the spinal cord of these mice (4, 5). Electron
microscopy and immunochemistry studies have shown that AVs
are significantly accumulated in the dystrophic axons of spinal
cord MNs in ALS mice. These changes appeared by the presymptomatic stage (90 days) and became more severe at the
end stage of the disease (120 days) (5).
Several previous studies suggest that the activation of autophagy may be neuroprotective in neurodegenerative diseases by
enhancing the removal of toxic protein aggregates (10). Specific
ablation of autophagy-related genes Atg 5 or Atg 7 in mouse
brain results in a neurodegenerative phenotype and the formation of protein aggregates (11). Recent findings indicate that
increased autophagy can protect Caenorhabditis elegans motor neurons against the toxicity of mutant SOD1 (12). Based on
these findings, we propose that autophagy is a potential target
for the development of antineurotoxicity treatments for ALS. The
classical pathway for autophagy control is through the serine/
threonine kinase, mTOR, which is a negative regulator of the process (13). Several pathways are thought to be involved, including insulin/growth factor pathways, the Ca2+-related pathway, as
well as other as-yet unknown pathways, all classified as mTORindependent pathways.
Rapamycin is a widely used mTOR-dependent autophagic enhancer that shows neuroprotective effects in HD and PD models
through autophagy-mediated protein degradation (14). Our own
studies have demonstrated that rapamycin unexpectedly accelerates disease progression in ALS mice, causing accumulation
of AVs, but fails to reduce the level of mutant SOD1 protein aggregates even though these are a known target of autophagymediated degradation (5). In addition, rapamycin treatment results in severe mitochondrial impairment, higher Bax levels, and
increased caspase-3 activation in ALS mice (14). Another study
reported that rapamycin had no effect on the disease progression in SOD1-H46R/H48Q mutant mice (15). Since the mTOR
pathway touches a broad variety of important cellular processes
including cell growth, mRNA translation, metabolism, and inflammation (13), rapamycin may not be a safe drug to treat ALS
patients. We therefore chose to investigate mTOR independent
autophagic inducers in ALS.
It is known that autophagy can be induced by lowering intracellular inositol 1,4,5-trisphosphate (IP3) levels. IP3-lowering agents,
such as lithium, carbamazepine, and sodium valproate, have
been demonstrated to induce autophagy in neurons (14). There
are contradictory data on lithium treatment in the disease pro-
60
gression of ALS (16, 17). Lithium has been reported to increase
survival and attenuate the disease progression in patients and
mouse models, partially due to induction of autophagy (16).
However, using the same strain of ALS mice and similar treatment methods, two other independent investigators did not find
neuroprotective effects of lithium (17). On the contrary, lithium
may cause an earlier disease onset and a reduction of survival
time in ALS mice (17). Moreover, concerns about the safety of
lithium carbonate as a therapy for ALS patients have been raised
after the halting of a trail that found serious side effects in its
participants (18). The efficacy of other IP3-lowering agents has
not been determined.
Trehalose is a non-reducing natural disaccharide with chemical
chaperone activity that induces autophagy in an mTOR-independent pathway by an unknown mechanism. A recent study found
that trehalose administration can significantly delay disease onset and prolong lifespan, and that such effects are associated
with activation of autophagy and improvement of autophagic flux
in the MNs of ALS mice (unpublished data). These results are
consistent with a report from Castillo and colleagues (19).
Autophagic Flux Defects and ALS
Autophagy is a highly dynamic and complex process, hence the
term “autophagic flux,” that can be regulated at multiple steps.
Early studies have demonstrated the accumulation of AVs in
a diverse range of neurodegenerative diseases. However, it is
difficult to determine whether this change is the result of induction of autophagy or impairment of autophagic flux (20). It has
been reported that the autophagic flux defect observed in ALS
could be caused by ALS-related mutant genes such as the dynein/dynactin complex, charged multivesicular body protein-2B
(CHMP2B), and ubiquilin 2 (21, 22). Our previous study showed
that rapamycin administration could cause further accumulation
of AVs, but failed to reduce the level of mutant SOD1 aggregation in the spinal cords of ALS mice (5); moreover, we found that
p62/SQSTM1 accumulates in the spinal cord, indicating a defective autophagy flux (5). These results suggest the possibility that
autophagy activity or flux may be impaired in SOD1 mutant ALS
mice.
In our present study using the mTOR-independent autophagy
activator trehalose, we documented that it can decrease SOD1
and p62 aggregation, and reduce the accumulation of ubiquitinated proteins in MNs of ALS mice. Furthermore, electron microscopy revealed that the reduction of protein aggregates by
trehalose is associated with improved autophagic flux in ALS
mice (unpublished data). These studies highlight the possibility that autophagy flux may be altered in SOD1-mediated ALS,
and enhancing autophagy with rapamycin may exaggerate the
degeneration of MNs. Instead, the use of multifunctional agents
that regulate autophagy at multiple points, or combinatorial
therapies to repair the autophagic flux defect and enhance ac-
tivation of the process, may have a more beneficial impact and
improve MNs function and survival.
Elevated Homocystin (Hcy) Levels
as Biomarker for ALS
The identification of ALS-specific biomarkers may provide opportunities to improve early disease diagnosis, monitor disease
progression, and possibly delay time to morbidity in clinical patients. Until now, no biomarkers have been identified that meet
the desired criteria for the diagnosis of ALS. Hcy is a sulfhydrylcontaining amino acid that is produced by demethylation of methionine. High plasma levels of Hcy have been reported to be
related to PD, AD, and other neurodegenerative disorders (24,
25). Using the ALS mice model, we detected high levels of Hcy
at early and late stages (90 and 120 days old) as compared with
the age-matched wild-type mice (6, 7). Further clinical study has
verified that Hcy levels are significantly higher in patients with
ALS compared to age- and sex-matched controls (26).
Folic acid and its metabolite, 5-methyltetrahydrofolate (5MTHF), play a key role in one-carbon metabolism, a process
that includes the remethylation of Hcy. We have documented
that the levels of folic acid and 5-MTHF are significantly lower in
ALS mice (7) and lower folic acid levels have also been reported
in ALS patients (26). Most interestingly, the decreased level of
5-MTHF is found at a young ages (30 days), indicating that the
transmethylation cycle might be altered very early, leading to
greater accumulation of Hcy by the later stages (7). Methylene
tetrahydrofolate reductase (MTHFR) catalyzes the conversion
of folic acid to 5-MTHF. A cross-sectional study has reported a
significant increase in ALS risk among clinical subjects with the
C677T mutation in the MTHFR gene (27).
The increase in plasma Hcy levels have been reported in patients with AD, PD, mild cognitive impairment, and ALS, so monitoring Hcy levels alone is not sufficient to diagnose ALS and track
disease progression. However, tracking the changes in folic acid
and 5-MTHF might help in early diagnosis, and preliminary evidence indicates that early intervention to reduce Hcy levels may
modify disease progression and even extend the lifespan of ALS
patients.
ics in the colon improve the absorption and synthesis of B vitamins. In ALS mice, administration of GOS-rich prebiotic yogurt
increased the levels of folate and B12, and reduced the level
of Hcy (28). Furthermore, chronic use of GOS-rich prebiotic yogurt attenuated MNs loss, improved mitochondrial activity, and
reduced atrophy of myocytes (28). These results suggest that
GOS-rich prebiotic yogurt may be an effective adjunct nutritional
therapy to lower Hcy in ALS sufferers.
A double-blind clinical trial conducted on ALS patients reported
that short-term high-dosage administration of methylcobalamin,
a form of vitamin B12, is effective in improving motor action of
patients, potentially through the reduction in Hcy (29). A clinical
review of 12 studies has confirmed that accumulation of Hcy increases the risk and progression of motoneuronal degeneration
in ALS patients (30).
While our studies indicate that altered autophagy status
and possibly an autophagic flux defect play a role in the pathology of ALS in our mouse model, the exact role of autophagy in ALS has not yet been fully elucidated. Additional
studies are required to more directly manipulate autophagy
in ALS model systems and detect the impact in disease progression and pathogenesis. Further, animal and clinical studies have provided growing evidence that elevated Hcy levels
impact ALS pathogenesis, but clearly more work is needed
to evaluate the efficacy and safety of directly targeting Hcy in
ALS patients.
REFERENCES
1. J. A. Rumfeldt, J. R. Lepock, E. M. Meiering, J. Mol. Biol. 385, 278 (2009).
2. P. Pasinelli, R. H. Brown, Nat. Rev. Neurosci. 7, 710 (2006).
3. M. C. Bellingham, CNS Neurosci. Ther. 17, 4 (2011).
4. L. Li, X. Zhang, W. Le, Autophagy 4, 290 (2008).
5. X. Zhang et al., Autophagy 7, 412 (2011).
6. X. Zhang, S. Chen, L. Li, Q. Wang, W. Le, Neuropharmacology 54, 1112 (2008).
7. X. Zhang, S. Chen, L. Li, Q. Wang, W. Le, J. Neurol. Sci. 293, 102 (2010).
8. C. A. Ross, M. A. Poirier, Nat. Med. 10 Suppl, S10 (2004).
9. S. Sasaki, J. Neuropathol. Exp. Neurol. 70, 349 (2011).
Hcy-Lowering Therapy in ALS
To investigate the efficacy of Hcy-lowering therapies, we first
tested folic acid and vitamin B12 (B12) in ALS mice to determine
their biological effects on Hcy levels and neuroprotection. We
found that a combination treatment with both folic acid and B12
can reduce the Hcy levels by up to 69% and, more importantly,
this regimen significantly delayed disease onset, prolonged lifespan, and enhanced MNs survival in ALS mice (6). A recent study
also demonstrated the neuroprotective effects of galactooligosaccharide (GOS)-rich prebiotic yogurt in ALS mice (28). GOSs
are oligomeric, non-digestible carbohydrates that help probiot-
10. Z. H. Cheung, N. Y. Ip, J. Neurochem. 118, 317 (2011).
11. M. Komatsu et al., Nature 441, 880 (2006).
12. J. Li, K. X. Huang, W. D. Le, Acta Pharmacol. Sin. 34, 644 (2013).
13. D. C. Rubinsztein, J. E. Gestwicki, L. O. Murphy, D. J. Klionsky, Nat. Rev. Drug Discov. 6, 304 (2007).
14. S. Sarkar, B. Ravikumar, R. A. Floto, D. C. Rubinsztein, Cell. Death Differ. 16, 46 (2009).
15. A. Bhattacharya et al., Neurobiol. Aging 33, 1829 (2012).
16. F. Fornai et al., Proc. Natl. Acad. Sci. U.S.A. 105, 2052 (2008).
17. A. Gill, J. Kidd, F. Vieira, K. Thompson, S. Perrin, PLOS ONE 4,
e6489 (2009).
61
THERAPEUTIC STRATEGIES
Produced by the Science/AAAS Custom Publishing Office
18. A. Chio et al., Neurology 75, 619 (2010).
27. P. Kuhnlein et al., Amyotroph. Lateral. Scler. 12, 136 (2011).
19. K. Castillo et al., Autophagy 9, 1308 (2013).
28. L. Song, Y. Gao, X. Zhang, W. Le, Neuroscience 246, 281
20. E. Wong, A. M. Cuervo, Nat. Neurosci. 13, 805 (2010).
21. S. Chen, X. Zhang, L. Song, W. Le, Brain Pathol. 22, 110 (2012).
29. R. Kaji et al., Muscle Nerve 21, 1775 (1998).
22. X. J. Zhang, S. Chen, K. X. Huang, W. D. Le, Acta Pharmacol. Sin. 30. S. Zoccolella, C. Bendotti, E. Beghi, G. Logroscino, Amyotroph. 34, 595 (2013).
(2013).
Lateral. Scler. 11, 140 (2010).
23. M. Nassif, C. Hetz, Autophagy 7, 450 (2011).
24. M. S. Morris, Lancet Neurol. 2, 425 (2003).
Acknowledgments: This study was funded by research grants from the
25. P. Quadri et al., Am. J. Clin. Nutr. 80, 114 (2004).
National Nature Science Foundation of China (81000541and 81171201)
26. S. Zoccolella et al., Neurology 70, 222 (2008).
and by the National Basic “973” Research Program (2011CB510003).
Novel Neuroprotective Strategy for Stroke—
Activating Inherent Neuronal Survival Mechanisms
Ming Chen, Bo-Xing Li, and Tian-Ming Gao*
Stroke is a leading cause of death, long-term disability, and socioeconomic cost worldwide. Thrombolysis is the only therapy
for acute ischemic stroke, but its use is limited by a narrow
therapeutic window. Although progress has been made in understanding the pathophysiology of stroke, clinically effective
neuroprotectants have remained elusive until now. Over the past
30 years, scientists worldwide had conducted many studies on
the neuronal death pathways induced by cerebral ischemia, and
have developed neuroprotective drugs targeting the key molecules mediating cell death; but thus far, all clinical trials have
failed (1). For example, clinical trials using N-methyl-D-aspartate
(NMDA) receptor antagonists—developed based on the calcium
(Ca2+) overload theory, the dominate theory in ischemic neuronal
cell death (2)—failed because of the short therapeutic time window and severe side effects of the drugs. Developing novel neuroprotective targets with a longer therapeutic window and fewer
side effects has become an urgent issue. The survival of neurons
after cerebral ischemia is determined not only by the activation
of cell death pathways, but also by the robustness of inherent
neuronal survival mechanisms. Activating these inherent survival
mechanisms is an attractive, novel neuroprotective strategy for
stroke. The survival of neurons depends on them experiencing
Department of Neurobiology, Key Laboratory of Neuroplasticity of
Guangdong Higher Education Institutes, Southern Medical University,
Guangzhou, China
*
Corresponding Author: [email protected]
62
physiological levels of electrical activity. This activity-dependent
survival depends on L-type Ca2+ channel-mediated Ca2+ influx
and the induction of neuronal survival-promoting gene expression (3–5). Here, we review our recent findings on the roles of ion
channels, which mediate Ca2+ influx and are required for neuronal
survival in ischemic neuronal death, and the protection against
ischemic cerebral injury afforded by neuregulin-1 (NRG1), the
expression of which is regulated by L-type Ca2+ channel activity
(Fig. 1).
Open L-type Ca2+ Channels—New Neuroprotective
Strategy with Long Therapeutic Window
Transient forebrain ischemia induces delayed, selective
neuronal death in the CA1 region of the hippocampus. The
induction of Ca2+ overload mediated by glutamate receptors is
the dominant theory in delayed neuronal death. Our previous
work has shown that the increase in L-type Ca2+ current 30
minutes after reperfusion might be one of the influx pathways
leading to acute postischemic Ca2+ overload during this early
period (6). Maintaining the proper intracellular Ca2+ levels (Ca2+
set point) is the key to the survival of neurons; too high or too low
Ca2+ levels is not conducive to neuronal survival. Studies have
reported that although intracellular Ca2+ concentration increases
in CA1 pyramidal neurons during the period of ischemia and
early reperfusion (30 minutes), at the late stage of reperfusion
the resting Ca2+ levels are below normal in hippocampal CA1
neurons. We therefore hypothesize that Ca2+ overload at the time
of ischemia and immediately after reperfusion is only a trigger,
but the later persistent lack of Ca2+ may be one of the
executors of delayed neuronal death. Using patch-clamp
techniques, we found that in vulnerable CA1 neurons,
L-type Ca2+ channel activity was transiently upregulated
and then persistently downregulated after ischemic insult,
whereas in invulnerable CA3 neurons, no such change
occurred. The lack of change in protein expression levels
of the L-type Ca2+ channel suggests that only functional
changes occurred after ischemia (7). Inhibition of the
L-type, but not the N-type or the P/Q-type Ca2+ channels,
significantly suppressed survival in hippocampal neurons
in culture. Treatment with a L-type Ca2+ channel agonist
after 1 to 12 hours of reoxygenation or as late as 24
hours after reperfusion significantly reduced cell death
induced by oxygen/glucose deprivation (OGD) and
ischemia-reperfusion injury in an animal model (7, 8). This
neuroprotection was mediated by the extracellular-signalregulated kinase (ERK) pathway (8). These results suggest
that the later, persistent L-type Ca2+ channel hypoactivity
may be responsible for the delayed neuronal death. These
findings also develop and correct the traditional Ca2+
overload theory and suggest a novel neuroprotective target
with a long therapeutic time window, namely the L-type
Ca2+ channel.
We further elucidated the mechanisms underlying the activity changes of the L-type Ca2+ channel after ischemia.
Nitric oxide is an important signaling molecule mediating ischemic neuronal death (9–11). At the early stage of
ischemia, elevated NO production enhances hippocampal
L-type Ca2+ channels by S-nitrosylation during the initial
phase of hypoxia and OGD (10, 11). During late stage of
reperfusion, oxidation modulation may be responsible for
the L-type Ca2+channel hypoactivity (7). These results suggest a new direction for drug design based on the disease
state, such as an effective agonist of the L-type Ca2+ channel in its oxidative state.
Figure 1. (A) Pathways of neuronal death following brain ischemia/reperfusion. Calcium
(Ca2+) overload occurs at early stage of reperfusion and may trigger oxidative modifications of ion channels. Later, persistent L-type Ca2+channel hypoactivity may be responsible for delayed neuronal death. In addition, BKCa channel hyperactivity decreased
membrane excitability, thereby further inhibiting L-type Ca2+ channel function. (B) Novel
neuroprotective strategies for stroke by activation of inherent neuronal survival mechanisms. Activation of L-type Ca2+ channels, blockade of BKCa channels, and enhancement of ErbB4 signaling are all candidate protective strategies. NMDAR, N-methyl-Daspartate receptor; BKCa, channel, large conductance Ca2+-activated potassium channel;
ERK, extracellular signal-regulated kinase; NRG1, neuregulin 1; ErbB4, v-erb-b2 avian
erythroblastic leukemia viral oncogene homolog 4; GABA, γ-aminobutyric acid; K+, potassium ion.
Blocking Large Conductance Ca2+-Activated
Potassium Channels—Alternative New
Strategy for Neuroprotection
In addition to regulation by L-type Ca2+ channels in the plasma
membrane, Ca2+ influx is also regulated by membrane excitability; a reduction in membrane excitability can inhibit L-type Ca2+
channel activation. By comparing the membrane properties of
ischemia-vulnerable CA1 pyramidal neurons and ischemic-invulnerable CA3 pyramidal neurons after mild and severe ischemia
with intracellular recording and staining techniques in vivo, we
found that only after severe ischemia did CA1 pyramidal neurons
show persistently decreased membrane excitability and a sustained increase in the magnitude of hyperpolarization mediated
by large conductance Ca2+-activated potassium (BKCa) channels.
By contrast, in ischemic-invulnerable neurons, no changes were
detected (12, 13). Therefore, the hyperactivity of BKCa channels
might persistently decrease membrane excitability and thereby
inhibit L-type Ca2+ channel activation, leading to ischemic neuronal death. This notion is supported by studies showing that BKCa
channel activity increased significantly in cultured hippocampal
neurons after hypoxia/reoxygenation and in CA1 neurons isolated from animals after forebrain ischemia/reperfusion (14, 15).
Administration of a BKCa channel-specific opener, NS1619, or in
vitro overexpression of BKCa channel genes induced apoptosis
in hippocampal neuronal cultures (15). Specific inhibition of BKCa
63
Produced by the Science/AAAS Custom Publishing Office
channels was neuroprotective in neurons subjected to hypoxic/
reoxygenation or OGD/reoxygenation, or in animals subjected to
forebrain ischemia/reperfusion (15, 16). Together, these results
indicate that BKCa channel hyperactivity mediates hippocampal
neuronal death after ischemia. The mechanism leading to cell
death may be due to decreased membrane excitability and inhibition of L-type Ca2+ channel function. Therefore, blocking BKCa
channels might provide a novel neuroprotective strategy.
We carried out a patch-clamp study to clarify the mechanisms
underlying the hyperactivity of BKCa channels after severe ischemia and reperfusion. Single-channel analysis revealed a sustained increase in open probability of BKCa channels in ischemic
CA1 pyramidal neurons, predominantly due to the increased Ca2+
sensitivity (17). Redox regulation of BKCa channels was investigated by comparing the sensitivity of BKCa channels to oxidants
and reductants in normal and ischemic neurons. We found that
more BKCa channels were in a reduced state in normal neurons,
while in ischemic neurons, more were in their oxidized state. This
suggests that oxidative modification may mediate the enhanced
BKCa channel activity following cerebral ischemia (17, 18).
ErbB4, Novel Target for Neuroprotection
NRG1 is a trophic factor that acts by stimulating ErbB receptor
kinases. The expression of NRG1 is regulated by L-type Ca2+
channel activity. Using transgenic mice expressing green fluorescent protein under the control of glutamic acid decarboxylase promoter (directs specific expression in γ-aminobutyric acid
(GABA) interneurons), we found that ErbB4 (v-erb-b2 avian
erythroblastic leukemia viral oncogene homolog 4) was localized in GABAergic presynaptic terminals (19). NRG1 enhanced
both the amplitude of evoked inhibitory postsynaptic currents
(eIPSCs) and the depolarization-induced release of GABA in a
dose-dependent manner (19). Further study showed that NRG1
significantly reduced the paired-pulse ratios (PPRs) of eIPSCs
in response to two successive stimuli, indicating that NRG1-enhanced GABAergic transmission works directly through the presynaptic terminals and not via postsynaptic mechanisms (19). To
verify that endogenous NRG1 regulates GABA release, we generated ecto-ErbB4, a neutralizing peptide containing the entire
extracellular region of ErbB4 that can specifically block NRG1
activation of ErbB kinases. Treatment with ecto-ErbB4 reduced
depolarization-induced GABA release and eIPSCs, indicating a
role for endogenous NRG1 in regulating these processes (19,
20). Two subtypes of ErbB receptors are expressed in neurons,
but only ErbB4 plays a role in the regulation of GABAergic transmission by NRG1 (19, 20). These findings identify a novel function of NRG1. More recently, we found that NRG1 modulation of
GABAergic transmission via ErbB4 occurred mainly in parvalbumin-positive interneurons, but not in pyramidal neurons (20).
We further studied whether NRG1 plays a neuroprotective role
through ErbB4 and GABAergic pathways in an OGD model. The
64
results showed that neurons in which ErbB4 had been knocked
out were more sensitive to OGD-induced injury, and both ectoErbB4 and GABA receptor antagonists could block the neuroprotective effect of NRG1. There was no synergism in the neuroprotective effects of NRG1 and GABA receptor agonists, suggesting
that the ErbB4 and GABAergic pathways are involved in the
protective effects of NRG1 (21). Our results reveal not only a
new mechanism underlying NRG1 neuroprotection against
ischemic brain injury, but also provide an experimental basis
for developing novel targets for neuroprotection such as the
ErbB4 receptor.
In summary, activating inherent neuronal survival pathways
is a novel neuroprotective strategy. L-type Ca2+ channels, BKCa
channels, and ErbB4 provide potentially exciting targets for treating ischemic stroke.
REFERENCES
1. S. I. Savitz, M. Fisher, Ann. Neurol. 61, 396 (2007).
2. D. W. Choi, Nature 433, 696 (2005).
3. P. L. Greer, M. E. Greenberg, Neuron 59, 846 (2008).
4. S. J. Zhang et al., PLOS Genet. 5, e1000604 (2009).
5. F. Leveille et al., J. Neurosci. 30, 2623 (2010).
6. X. M. Li et al., Prog. Biochem. Biophys. 30, 755 (2003).
7. X. M. Li et al., J. Neurosci. 27, 5249 (2007).
8. H. H. Hu et al., Mol. Neurobiol. 47, 280 (2013).
9. M. Chen et al., NeuroSignals 17, 162 (2009).
10. K. Jian et al., Biochem. Biophys. Res. Commun. 359, 481 (2007).
11. Y. W. Tjong et al., Free Radic. Biol. Med. 42, 52 (2007).
12. T. M. Gao, E. M. Howard, Z. C. Xu, J. Neurophysiol. 80, 2860 (1998).
13. T. M. Gao, W. A. Pulsinelli, Z. C. Xu, Neuroscience 90, 771 (1999).
14. L. Gong, T. M. Gao, X. Li, H. Huang, Z. Tong, Brain Res. 884, 147 (2000).
15. M. Chen et al., Mol. Neurobiol. Epub ahead of print (2013) doi: 10.1007/
s12035-013-8467-x.
16. H. Huang, T. M. Gao, L. Gong, Z. Zhuang, X. Li, Neurosci. Lett. 305, 83 (2001).
17. L. Gong, T. M. Gao, H. Huang, Z. Tong, Eur. J. Neurosci. 15, 779 (2002).
18. L. Gong, T. M. Gao, H. Huang, Z. Tong, Neurosci. Lett. 286, 191 (2000).
19. R. S. Woo et al., Neuron 54, 599 (2007).
20. Y. J. Chen et al., Proc. Natl. Acad. Sci. U.S.A. 107, 21818 (2010).
21. C. Y. Wu et al., Procedings of the Chinese Society for Neuroscience Ninth National Academic Conference and the Fifth Congress p.135 (2011).
Acknowledgments: This work was supported by grants from the National
Natural Science Foundation of China (81030022, 81070983, 81329003,
31371146, and U1201225), the Key Project of Guangdong Province
(9351051501000003 and CXZD1018), the Guangzhou Science and Technology
Project (7411802013939), and the Program for Changjiang Scholars and
Innovative Research Team in University (IRT1142).