All neuropathies great and small

Abstract

Autosomal-dominant pure hereditary spastic paraplegia (AD-HSP) is characterized by the degeneration of long axons in corticospinal tracts and dorsal columns, resulting in spasticity and difficulty walking. Mutations in the SPG4 gene product spastin are the predominant genetic lesions associated with this inherited disease. In this issue, Orso et al. examine and reconcile existing Drosophila mutants of spastin and generate a new model for HSP by overexpression of a fly spastin transgene that carries a mutation prevalent in human AD-HSP. Expression of this mutant spastin protein produces pathology in flies reminiscent of the human disease, including adult locomotion defects, in addition to causing aberrant synaptic morphology and altered microtubule stability. Both movement and synaptic defects in fly mutants were ameliorated by treatment with the microtubule-modifying agent vinblastine. The results are consistent with disease-causing mutations in human spastin producing dominant-negative proteins and confirm the usefulness of Drosophila genetic techniques to understand HSP and other neurodegenerative diseases.

Good things come in small packages, and for nearly a century the minuscule fruit fly Drosophila melanogaster has been generous to researchers, leading to the development of a genetic toolkit of unrivaled sophistication. With a rapid generation time and inexpensive culture requirements, huge numbers of Drosophila can be screened in order to find rare mutations and fully exploit the advantages of a completely sequenced genome. Genetics can be coupled with molecular and cell biology, biochemistry, electrophysiology, and so on for a comprehensive phenotypic analysis. The combination has proven powerful, yielding fundamental insights into the mechanisms of development, physiology, and neurobiology, for even though flies are small, their brains are estimated to have over 250,000 neurons, allowing relatively complex behaviors. The beauty of Drosophila is not only skin (or even cuticle) deep, however — a comparison of the fly and human genomes reveals striking similarities. For example, 77% of the approximately 1,000 genes that have been linked to specific human diseases have a readily identifiable Drosophila homolog (1). An increasing number of researchers are taking advantage of this conservation to study the basic molecular and cell biology of disease-causing genes and search for interacting molecules using the genetics of this tractable model organism.

A number of approaches are available for examining a disease-related gene in Drosophila (Figure 1). The traditional and most reliable way is to generate a mutant for the fly homolog of the disease-associated gene. With large-scale programs underway to generate mutations in the majority of Drosophila’s approximately 14,000 genes, this approach is becoming easier and more rapid (2). If a disease protein has a dominant effect, another approach is to overexpress the protein in an otherwise normal fly. Thanks to the Gal4/upstream activating sequence (Gal4/UAS) binary expression system (3), transgenic proteins can be expressed in many tissues or in just small subsets of cells. For example, eye-specific overexpression of many disease-associated proteins disrupts the organization of cells in the eye, which can easily be “scored” by microscopic examination, providing a productive in vivo platform to search for mutations that suppress the deleterious effects of disease genes (4). The latest addition to the flies’ genetic arsenal is transgenic RNA interference (RNAi), which relies on the ability of double-stranded RNAs to induce the degradation of homologous mRNAs by conscripting the enzymatic machinery normally utilized to protect against viruses or produce microRNAs (5). To target a gene for interference, a transgenic construct is generated with inverted regions of the gene of interest. When RNA is expressed from this transgene, it forms a hairpin that is cleaved by endogenous fly enzymes into small interfering RNAs (siRNAs). These siRNAs are used as a guide by the RNAi silencing complex (RISC) to recognize and deplete homologous mRNA sequences, knocking down protein expression of the targeted gene. Unlike in plants and Caenorhabditis elegans, the RNAi effect in Drosophila seems to be cell autonomous, allowing transcripts to be targeted only in the cells where the RNAi transgene is expressed using the Gal4/UAS system.

Figure 1.

Currently available approaches for the examination of disease-related genes in Drosophila. RISC, RNAi silencing complex; siRNA, small interfering RNA; Dicer, a protein that produces siRNA from long, double-stranded RNAs.

Miniature models of hereditary spastic paraplegia

In this issue of the JCI, Orso and colleagues utilize a combination of genetic techniques to study the neurodegenerative disorder hereditary spastic paraplegia (HSP) in a Drosophila model (6). HSPs are a heterogeneous group of disorders that vary in their pattern of inheritance, time of onset, and pathology; however, all are characterized by spasticity associated with the degeneration of spinal cord axons (7). Pure HSPs exhibit retrograde axonopathy of the terminal ends of axons in corticospinal tracts and the fasciculus gracilis and were classically thought to be uncomplicated by neurological symptoms other than spasticity, although more recent evidence suggests that neuronal dysfunction can also occur outside of the motor system (8). Eleven genetic loci, including 4 characterized genes, are associated with autosomal-dominant pure HSP (AD-HSP). All 4 of the characterized AD-HSP genes — spastin (SPAST), kinesin family member 5A (KIF5A), atlastin (also known as spastic paraplegia 3A, SPG3A), and heat shock protein 60 (HSP60) — have Drosophila homologs. Greater than 40% of human AD-HSPs are caused by mutations in spastin, which belongs to the AAA family of adenosine triphosphatases. Spastin is related to katanin p60, which has biochemically been shown to sever microtubules. Spastin, too, can cleave microtubules when overexpressed in cell culture (7, 9, 10). Microtubules provide structural support to axons and act as tracks along which proteins, vesicles, and other essential cargo are transported to the synapse, making them essential for the development and normal function of neurons and synaptic terminals.

Suspicion of inhibition

Last year, 2 independent groups employed different techniques to study spastin function in Drosophila (11, 12). Both groups determined that spastin overexpression within the nervous system was highly toxic, causing early mortality. Overexpression was associated with decreased staining of microtubules, suggesting that spastin can cleave microtubules in vivo, which has previously been demonstrated in vitro (10). However, their data conflicted concerning the effects of spastin loss of function on adult viability and neuromuscular junction (NMJ) morphology and function. Trotta et al. (11) knocked down fly spastin expression using transgenic RNAi, while Sherwood et al. (12) generated mutations in the fly spastin gene. Approximately 20% of complete loss-of-function spastin mutants reached adulthood, with those that survived exhibiting severe motor defects, which indicates that spastin loss of function produces pathology in Drosophila reminiscent of human HSP (12). In contrast, inhibition of spastin RNA by the ubiquitous expression of a transgenic RNAi construct was reported to cause complete lethality during development (9).

Both groups also examined the requirement for spastin in the development of motor neuron synapses. The Drosophila NMJ is one of the best-characterized synapses in any organism at the morphological and physiological level. During the larval stage, Drosophila NMJ synapses grow rapidly by the addition of synaptic branches and boutons (13). Examination of NMJ morphology in spastin mutants revealed smaller, more numerous synaptic boutons at the NMJ than in control animals as well as reduced numbers of microtubule bundles (10). Electrophysiological recordings from these mutants showed that neurotransmitter release was reduced. In contrast, when spastin was inhibited in the nervous system using RNAi, the NMJ synaptic area was decreased, and staining of posttranslationally modified tubulin was increased (9). Electrophysiological examination of these RNAi-inhibited animals found increased amplitudes of excitatory junctional currents (9). No functional protein was produced by spastin mutants (10), so it was worrying that RNA interference of spastin caused an opposing phenotype. Was the transgenic RNAi technique doing something unexpected?

Since RNAi was initially discovered, there has been speculation that problems such as lack of specificity may limit the usefulness of the technique (5). In vertebrate cell culture, RNAi has been found to cause the degradation of unrelated targets and activation of the IFN pathway (14, 15), though this should not be a problem in Drosophila (5). The incongruous results produced by spastin mutations versus RNAi of spastin sparked an intense discussion about the relevance and pitfalls of RNAi models in Drosophila. It was speculated that the spastin RNAi construct might have off-target activity and produce effects unrelated to spastin function. Furthermore, similar problems might arise in other RNAi models, limiting their usefulness. This was particularly disappointing because transgenic RNAi models are generally faster to produce than gene mutations and have advantages such as the ability to knock down gene expression in specific cell types or times during development, even though RNA interference suffers the disadvantage of commonly only producing partial loss-of-function phenotypes (5).

Concord, not conflict

The study by Orso and colleagues in this issue of the JCI (6) has relieved some of these worries, if not resolved them. A problem with previous studies is that like was not compared with like. Whereas spastin mutations remove protein function throughout the animal, RNAi experiments had primarily characterized the effects of removing spastin expression solely in the nervous system. Orso et al. have examined the effects of interfering with spastin RNA in all tissues, not just in the nervous system. The authors demonstrate that ubiquitous spastin RNAi expression and spastin-null mutants have a similar phenotype: low eclosion rates, a decreased NMJ synaptic area with an increased number of synaptic boutons, and increased amounts of acetylated tubulin. Electrophysiological recordings were not carried out in the current study, but presumably ubiquitous RNAi expression would produce electrophysiological defects similar to those observed in spastin mutants.

Orso et al. have developed another way to model HSP, by transgenic expression of a mutant spastin protein (6). They generated a point mutation in a transgenic Drosophila spastin at a location equivalent to a prevalent mutation responsible for HSP in humans. Overexpression of this mutant protein produced pathology similar to that caused by RNAi of spastin. Whereas overexpression in the eye of wild-type spastin alone caused cellular disarray, coexpression with the mutant spastin suppressed this toxic effect. These experiments were revealing because they strongly suggest that disease-associated spastin mutations produce dominant-negative proteins that disrupt normal spastin function.

While Orso et al. (6) have provided compelling evidence that RNA interference of spastin can be effectively used to study HSP, some differences among fly spastin models still exist. In particular, how spastin loss of function affects microtubules remains unclear. Previous studies used very different methods to assess the presence of microtubules. Sherwood et al. (12) found decreased levels of α-tubulin and the microtubule-associated protein MAP1B. Trotta et al. (11) found that although spastin RNAi increased the level of posttranslationally-modified tubulin, total tubulin staining was decreased, perhaps due to depletion of free tubulin dimers. This may account for Sherwood et al.’s finding of decreased α-tubulin at spastin mutant synapses. Decreased MAP1B staining in spastin mutants indicates that microtubule structure may be disrupted and suggests that MAP1B may be involved in the HSP pathologic mechanism, as mutations in other microtubule-associated proteins can produce neurodegeneration (16).

Alighting on new therapies

The path is now clear to apply the full power of Drosophila genetics to elucidate the molecular and cellular mechanisms behind HSP and other neurodegenerative disorders. Several genes associated with HSP other than spastin are also required for cellular transport and trafficking. KIF5A is a neuronal kinesin superfamily member involved in fast axonal transport, while atlastin encodes a GTPase with homology to dynamins, which are associated with vesicle trafficking. Disruption of axonal transport may be a common mechanism by which mutations in these genes cause HSP and could explain why spinal neurons with long axons are particularly vulnerable to HSP. Drosophila is an ideal organism in which to study the genetic interactions among HSP genes and to decipher whether these proteins act in the same pathological pathway. Furthermore, by identifying mutations that enhance or suppress the spastin mutant phenotype, new targets for therapeutic intervention may be isolated. Orso et al. (6) show that vinblastine, a microtubule-modifying agent already used in human patients for cancer chemotherapy, ameliorates the behavioral and morphological phenotypes produced by spastin loss of function. Chemotherapy with vinblastine causes neuropathy, perhaps due to disruption of neuronal microtubules, which indicates that at clinically tolerated doses, vinblastine might have therapeutic potential for HSP (17). A number of drugs being evaluated as therapies for Huntington disease, Parkinson disease, and Alzheimer disease are effective in Drosophila models of these disorders (18), suggesting that flies may be useful for initial validation or perhaps identification of new drug treatments. If Drosophila can contribute to the development of effective therapeutic treatments for neurological diseases like HSP, more than geneticists will benefit from their gifts.