With large genome-wide association studies having yielded perhaps all the genes for Alzheimer's they can, scientists are turning to new methods. In the January 31 Science, researchers led by Joseph Gleeson at the University of California, San Diego, describe how a combination of exome sequencing and gene interaction network analysis identified 18 new genes that cause hereditary spastic paraplegias (HSPs), a group of rare movement disorders. The network, or “HSPome,” they derived comprises nearly 600 proteins and brings into view the biological pathways that are most affected by the disease. These include processes such as protein trafficking and degradation, and lipid biology, which have also been implicated in other neurodegenerative diseases such as Alzheimer’s and Parkinson’s. In fact, analysis showed that the HSPome overlapped with known genes for AD, PD, and amyotrophic lateral sclerosis (ALS). “This told us that common neurodegenerative diseases share similar networks and cellular vulnerabilities,” Gleeson told Alzforum. “Maybe we need to think about these less as individual diseases, and more as a problem of neuronal susceptibility.”

Other researchers found the method exciting and said they believe it holds promise for gene discovery in other disorders. In an accompanying commentary, Andrew Singleton at the National Institute on Aging (NIA), Bethesda, Maryland, wrote that the authors “perform what is perhaps the most complete genetic analysis of the neurological disorder HSP,” and that the work “shows not only the power of comprehensive genetic analysis in identifying the pathogenic networks involved in that disease, but also the potential of such work to inform outside of the disease in question.” Mark Cookson at NIA noted, “As of tomorrow, many labs will have their informatics experts studying this.”

For many neurodegenerative diseases, GWAS have now found most of the common variants that contribute small risks of disease. However, GWAS typically cannot identify less-common variants that confer moderate or high risk. These are typically found by performing more comprehensive genetic sequencing in affected families. Many initiatives have sprung up to sequence exomes or whole genomes in search of these genes (see Nov 2010 news storyNov 2012 news storyOct 2012 news storyAug 2013 news story; and Dec 2013 news story). 

Gleeson and colleagues wanted to use this type of approach to find genetic variants responsible for HSP, a heterogeneous group of motor neuron diseases characterized by loss of corticospinal tract function. Joint first authors Gaia Novarino, Ali Fenstermaker, and Maha Zaki chose 55 families with high rates of intermarriage where the disease was inherited in an autosomal recessive fashion. The authors sequenced the exomes, that is, all expressed regions of the genome, in two members of each family. They looked in homozygous regions for rare gene variants that segregated with disease and would likely inactivate the encoded protein. By this method, they found known HSP genes in a third of the families and 15 new gene candidates in another 40 percent of the families.

One of the biggest challenges in gene discovery is to prove that a variant associated with a person’s clinical status actually causes their disease. To validate these 15 genes, the authors looked for the same variants in 200 additional HSP patients. They found five of the candidates. Knockdown of several of the others in zebrafish produced movement problems, suggesting that these genes might contribute to HSP, as well.

The authors then built a gene interaction network by combining the 43 known HSP genes and the 15 new candidates. Analysis showed that the genes were significantly more related than would be expected at random, strengthening the idea that they act in biological pathways. To expand the network, the authors searched for proteins known to bind to the HSP set in databases of protein interactions such as HumanNet (see Lee et al., 2011) and STRING (see Szklarczyk et al., 2011). These searches returned an HSPome of 589 proteins (see image below). Reasoning that these interacting proteins might also play a role in disease, the authors then re-examined their exome sequences for mutations in these additional genes and found three more candidates they initially had missed. One of these genes was mutated in two separate HSP families, a second produced spinal defects in knockout mice, and the third was recently independently identified as a cause of dominant HSP (see Oates et al., 2013), suggesting that these finds represent genuine HSP genes.

Protein interaction network built around known and candidate HSP genes forms the “HSPome.” [Image courtesy of Science/AAAS]

The HSPome network flagged a handful of biological processes as likely to be involved in the disease. These include protein trafficking, endosome sorting, and protein degradation, which have been implicated in Alzheimer’s, Parkinson’s, ALS, and FTD as well (see, e.g., Jul 2008 conference story; Jun 2010 news storyJul 2011 news story; and Jun 2012 news story). Other processes of interest were lipid metabolism, axon and synapse development, and purine nucleotide metabolism. Further analysis showed similarities between the HSPome and sets of known risk genes for Alzheimer’s, Parkinson’s, and ALS, but not with neurodevelopmental disorders, suggesting common mechanisms of neurodegeneration. 

John Hardy at University College London noted that similarities between different neurodegenerative disease mechanisms crop up repeatedly, with another example being perturbed calcium homeostasis in both Alzheimer’s disease and ataxia disorders. “These defective molecular pathways give us clues about the basis of selective neuronal vulnerability,” he wrote to Alzforum.

Commentators agreed that the method described in this paper would work well for other hereditary recessive diseases, but might be difficult to apply to disorders with dominant or complex inheritance, such as AD. In most cases, AD researchers cannot narrow down gene candidates by looking for homozygous variants, although there are exceptions (see Aug 2013 news story). Even so, researchers enthusiastically endorsed the idea of building interaction networks starting from known disease genes and using those to guide discovery. “In my view, this will be a useful method to prioritize the most likely candidates from GWAS,” Cookson told Alzforum. 

“[Protein] interaction databases will continue to get better, and I think approaches such as the one used in this study will only become more useful,” wrote Jeremy Miller at the Allen Institute for Brain Science, Seattle. “While I think applying a similar strategy in AD would be more complicated, the idea is sound, and could probably be used to learn more about the disease.”

Several groups are already taking such approaches (see, e.g., May 2013 webinar). Rita Guerreiro at University College London said her group is doing network analysis centered on the TREM2 gene, which has revealed common inflammatory pathways that seem to be affected in AD, multiple sclerosis, and motor neuron diseases (see Forabosco et al., 2013). “The commonality of etiological pathways between different neurodegenerative diseases has been an old topic of discussion that has gained a new view with the availability of next-generation genetic technologies. If confirmed, this finding will be the basis of different lines of research into neurodegenerative processes in the near future,” she wrote to Alzforum.—Madolyn Bowman Rogers.

Comments

Make a Comment

To make a comment you must login or register.

Comments on this content

No Available Comments

Comments on Primary Papers for this Article

No Available Comments on Primary Papers for this Article

References

News Citations

  1. Next-Generation Sequencing: Boldly Going Where No Geneticist...
  2. Genetics Project Update: Over 1,000 Genomes and Counting
  3. Thousands of Whole Genomes to Be Mined for New Clues to AD
  4. Scientists Hunting Rare Alzheimer’s Mutations Eye Phospholipase D3
  5. Alzheimer’s Whole-Genome Data Now Available From the NIH
  6. Paris: Intracellular Traffic and Neurodegenerative Disorders
  7. Death of the Neatnik: Neurons Perish When Trash Clutters Their Space?
  8. Sorting Out Parkinson’s: Exome Sequencing Points to Recycling Defect
  9. Evidence Piles Up for Lysosomal Dysfunction in Parkinson’s
  10. Racing Toward New Alzheimer’s Mutations with Runs of Homozygosity

Webinar Citations

  1. Can Network Analysis Identify Pathological Pathways in Alzheimer’s

Paper Citations

  1. . Prioritizing candidate disease genes by network-based boosting of genome-wide association data. Genome Res. 2011 Jul;21(7) Epub 2011 May 2 PubMed.
  2. . The STRING database in 2011: functional interaction networks of proteins, globally integrated and scored. Nucleic Acids Res. 2011 Jan;39(Database issue) Epub 2010 Nov 2 PubMed.
  3. . Mutations in BICD2 cause dominant congenital spinal muscular atrophy and hereditary spastic paraplegia. Am J Hum Genet. 2013 Jun 6;92(6) Epub 2013 May 9 PubMed.
  4. . Insights into TREM2 biology by network analysis of human brain gene expression data. Neurobiol Aging. 2013 Dec;34(12):2699-714. PubMed.

Further Reading

News

  1. Barcelona: What Lies Beyond Genomewide Association Studies?
  2. In Shout-Out for Community Studies, Vantaa Finds Protective Mutation
  3. New Method to Boost Single Cell Genomics
  4. Do Copy Number Variations Point to Potential AD Genes?