ARTICLE IN BRIEF
Investigators used whole-exome sequencing in combination with network analysis to identify 18 previously unknown putative hereditary spastic paraplegia genes, and validated nearly all of these genes functionally or genetically.
A novel combination of cutting-edge gene discovery techniques has turned up 18 new genes for hereditary spastic paraplegia (HSP), including genes for cellular processes not previously implicated in the disease. The study, published in the Jan. 31 issue of Science, also highlighted potential links between HSP and other neurodegenerative diseases that may be targeted therapeutically.
The study began with 55 consanguineous families displaying an autosomal recessive form of the disease. Consanguinity increased the likelihood of finding disease-causing recessive mutations. Recessive mutations were required for this study because of what happened next. DNA from individuals with the disease was first analyzed with whole-exome sequencing, determining the sequencing of every coding region in the genome. For most genes, the sequences of the two alleles, one inherited from each parent, were different. These could be ruled out as the source of the disease, because a recessive disease requires two identical copies of the gene, explained lead author Joseph Gleeson, MD, professor of neurosciences and pediatrics at the University of California, San Diego.
That left a much smaller number of homozygous genes, many of which could be eliminated based on the absence of any unusual sequence variation within them, compared with the population at large. Genes were further ruled out if the variant did not segregate with the disease.
This method turned up 13 known HSP genes, and identified 15 candidate genes not previously associated with HSP. These were examined further, first by screening for them in 200 HSP patients without a previous genetic diagnosis; five of the 15 genes were confirmed in this group.
For five genes occurring in only a single family, Dr. Gleeson tested the consequence of their loss in zebrafish, by knocking each down using RNA interference. Each produced a locomotor phenotype consistent with defects in motor neurons.
“Although more work is warranted to conclusively uncover the role of the tested genes in corticospinal tract degeneration,” Dr. Gleeson said, “our in vivo functional validation supports the genetic data.”
Next, Dr. Gleeson turned to “network analysis,” a computer-driven method to understand the relationships among a set of genes, as a way to both understand the functional consequences of the loss of these genes in HSP, and to predict other genes that may be involved in the disease.
Network analysis draws upon the known interactions among the entire human proteome, as recorded in the scientific literature and annotated by data-mining companies. Some of these interactions are fully understood — for example, when one protein acts as a substrate for another in a well-known biochemical pathway — while others are tentative, for instance when two proteins are isolated together in a gel.
Dr. Gleeson constructed a “seed network” using the previously known autosomal recessive HSP genes. Some of these genes were closely linked to others in the network, reflecting the known interactions among them: AP4B1 (adaptor-related protein-4 complex subunit beta-1) and AP4S1 (adaptor-related protein-4 complex subunit sigma-1), for instance, are both part of a larger protein complex that sorts membrane proteins. In the seed network, they are connected by a single straight line, or “edge” in the language of network analysis. In contrast, SPG21 (spastic paraplegia 21) is separated from these two by multiple edges, representing its separate function and lack of known direct interaction.
To the seed network, he then added the new candidate genes, each connected more or less closely to the existing genes. For instance, the newly identified ERLIN1 (endoplasmic reticulum lipid raft-associated protein 1) gene, involved in endoplasmic reticulum-associated degradation of proteins, was connected by a single edge to the known HSP gene ERLIN2, but was many edges away from AP4B1. This expanded network was then statistically compared to 10,000 randomly chosen, similarly sized additions to the seed network. “We found that the new candidate genes were much more tightly linked to the seed network than would be expected by chance,” Dr. Gleeson said.
Dr. Gleeson extended the group by including proximal interactors with seeds and candidate genes, a total of 589 genes, which together comprise what he dubbed the “HSPome.” “This allows a global view of HSP,” he said, and focuses attention on other genes that may be mutated in HSP patients. Indeed, further testing of families not included in the original network analysis revealed mutations in three more genes. “Although further validation of these three candidates is necessary in larger cohorts, the data suggest the HSPome can be useful to identify HSP-relevant pathways and genes,” he said.
The 18 new genes bring the total number of HSP genes to 70, which largely cluster into a much smaller number of pathogenic categories. Besides ER-associated degradation and membrane traffic control, these include lipid and myelin biosynthetic pathways, as well as axonal guidance and synapse-related genes. A new category implicated in this study is nucleotide metabolism, with three new genes identified encoding enzymes that metabolize purines. How mutations in these genes cause HSP is unknown.
Finally, Dr. Gleeson asked about commonalities among the HSP network and those of other neurologic diseases, including neurodegenerative, developmental, and seizure disorders. “We found that the set of HSP seeds plus candidates significantly overlaps with sets of genes previously implicated in three neurodegenerative disorders: amyotrophic lateral sclerosis, Alzheimer's disease, and Parkinson's disease,” but not with the other types of disorders.
“This suggests there are common vulnerabilities that perhaps all neurons have,” with the type of neuron affected varying by disease, but all experiencing stresses through common pathways, Dr. Gleeson said. “These neurodegenerative diseases are connected in ways that we need to explore in more detail.” They may reveal key targets for developing new treatments, he suggested.
This gene discovery approach can't be used for every disease, Dr. Gleeson noted. “Our work benefited in part because this condition is so genetically heterogeneous, but is still tractable.” A single-gene disorder such as Rett syndrome would not reveal multiple new genes, while a more complex disorder, such as Alzheimer's disease, would be too difficult to analyze, since many genes are thought to contribute a small amount of risk. Dr. Gleeson is currently applying this approach to other neurodevelopmental disorders.
“Even before this study, HSP was one of the most genetically diverse diseases,” noted Craig Blackstone, MD, PhD, senior investigator at the National Institute of Neurological Disorders and Stroke, who was not involved with the study. “But even though there are many genes, there are a much smaller number of themes, and this study reinforces that idea.”
While many of the genes discovered are common to only one of a few families, “they may still give us the missing piece, the insight that might tell us which direction we should be looking” to understand the disease. The discovery of new genes also gives neurologists a chance to provide more of their patients with a diagnosis. “Giving them an answer is important,” he said.
John Fink, MD, associate professor of neurology at the University of Michigan in Ann Arbor, agreed. “This study provides a treasure trove of new insights into the disease process. What is unique about this paper is not the identification of multiple pathways, but the abundance of new genes all at once.”
Both experts agreed there are likely to be dozens more HSP genes yet to be discovered, a consequence perhaps of the extreme length and resulting vulnerability of the upper motor neurons affected in HSP.
While there appears to be some overlap with other neurodegenerative disorders, Dr. Fink said, the relative lack of common genes between HSP and amyotrophic lateral sclerosis (ALS) — both degenerative motor neuron disorders — may also be significant. “That suggests there is a fundamentally different molecular mechanism between the various types of ALS on the one hand, and the HSPs on the other.”