For every other person with frontotemporal dementia (FTD), some form of the disease runs in the family. A few causative genes have already taken the blame for plaguing nearly half of these families, but much of the genetic burden remains unaccounted for. At the 9th International Conference on Frontotemporal Dementia, held last month in Vancouver, Canada, the ongoing hunt for more genetic culprits was on full display. Besides offering up a handful of new, potentially causal mutations for discussion, researchers presented a trove of data revealing how some genes steer the course of the disease in people who harbor a known variant. Another researcher probed the mechanistic underpinnings of a previously discovered mutation, and uncovered a fundamental finding about the way RNA moves throughout the cell.

Autosomal-dominant mutations in the genes encoding tau (MAPT), progranulin (GRN), and C9ORF72 account for the lion’s share of genetically inherited FTD cases in which a gene has been found. Mutations in the genes encoding valosin-containing protein (VCP), fused in sarcoma (FUS), and charged multivesicular body protein 2B (CHMP2B) are much rarer. Together, these known causal genes explain the disease in roughly 40 percent of affected families, plus about 10 percent among cohorts of unrelated patients. Complicating matters, even people who share the same causal mutation may present with drastically different disease characteristics. For example, people harboring C9ORF72 expansions may develop amyotrophic lateral sclerosis (ALS), FTD, or a devastating combination of both disorders. The age at which disease strikes can vary by decades. This bewildering picture has spurred researchers to push for progress in two directions: new genes that cause disease, and genetic factors that explain the range of clinical phenotypes.

The Modifiers
Rosa Rademakers of the Mayo Clinic in Jacksonville, Florida, kicked things off with factors that influence disease in carriers of the C9ORF72 expansion. Rademakers touched on recently published data suggesting that the two genes TMEM106B and ataxin-2 modify which clinical disease develops in a given C9ORF72 carrier. Specifically, TMEM106B mutations protect C9ORF72 expansion carriers from developing FTD, but not ALS; similarly, an intermediate number of expansions in the ataxin-2 gene skewed C9ORF72 carriers toward ALS (see van Blitterswijk et al., 2014, and Sep 2014 news story).

To dig for more genes that may influence when a C9 carrier becomes symptomatic, Rademakers and dozens of colleagues screened some 300 C9ORF72 expansion carriers for 36 different genetic variants that previously had been associated with FTD or ALS. This netted three variants associated with when disease struck, plus six associated with the length of survival thereafter (see van Blitterswijk et al., 2014). The candidate modifier genes include those involved in protein sorting (ubiquitin-associated protein 1), transcription (elongator acetyltransferase complex subunit 3), and antioxidant defense (metallothionein), as well as APOE4 and granulin. All need confirmation in larger cohorts, Rademakers said.

Rademakers also presented fresh data looking at how C9ORF72 expansions trigger changes in gene expression in the brain. Performing whole-genome expression analyses on samples from 32 expansion carriers and 30 controls, her group identified 40 genes that were differentially expressed in the cerebellum, and three in the frontal cortex. A majority hailed from the homeobox, or hox, family of genes, which is primarily implicated in development but also in neuronal repair and regeneration. In preliminary experiments in non-neuronal cell lines, Rademakers found that loss of C9ORF72 expression (as opposed to expression of C9ORF72 expansions), triggered hox gene upregulation. The C9ORF72 expansion is known to generate gobs of dipeptide repeat (DPR) inclusions in the cerebellum but, oddly, this region is not particularly hard hit with neurodegeneration, Rademakers said (see also Part 7 of this series). The expansions also downregulate expression of the C9ORF72 gene itself. Rademakers speculated that the hox response could point to a protective mechanism that counters negative effects of the C9 expansion in the cerebellum.

In a packed talk, Christine Van Broeckhoven of the VIB, University of Antwerp, Belgium, laid out more findings about genetic modifiers. For one, Van Broeckhoven looked in C9ORF72 expansion carriers for evidence of genetic anticipation. In this phenomenon, known from Huntington’s disease, children develop the disease at a younger age than their parents did. Drawing on blood and brain samples from C9ORF72 family pedigrees, Ilse Gijselinck, a postdoc in Van Broeckhoven's group, found that expansions grew longer from one generation to the next, and that this increase correlated with an earlier age at onset. The researchers also found that longer expansions were more hypermethylated, in both a nearby CpG island and the CG-rich expanded-repeat sequences themselves. This in turn dampened expression of the C9ORF72 gene. Longer repeats therefore correlated with less C9ORF72 expression as well as with younger onset, Van Broeckhoven said. Further, Van Broekhoven showed that C9ORF72 expansions as short as 47 repeat units co-segregated with disease in two families and had the same DPR pathology as those with a longer expansion.

For another modifier of age at onset in progranulin-associated FTLD, Van Broeckhoven presented preliminary data of Eline Wauters, a Ph.D. student in her group. In the Belgian founder family who led to the identification of progranulin’s link to FTD, all carriers harbor a granulin mutation that prevents proper splicing of the gene. However, some relatives fall ill at age 45, others are unaffected at age 84. Van Broeckhoven’s group searched for loci containing variants that correlated with age at onset, and zeroed in on a 7 Mb region on chromosome 12 that accounts for a majority of the variance. Van Broeckhoven said the modifier is likely to reside within an intergenic region, where it may affect the expression of nearby or distant genes. “It could also be that this modifier only modifies age at onset in granulin mutation carriers, but maybe in C9ORF72 expansion carriers or sporadic patients as well,” she told Alzforum. “It’s best to keep an open mind.”

Newbies, Anyone?
Scientists at ICFTD were thrown a handful of new FTD-linked genes to unwrap, some of them better-vetted than others. Van Broeckhoven offered two—vacuolar protein sorting 13 homolog C (VPS13C) and filamin C (FLNC). As presented in a talk by her Ph.D. student Stéphanie Philtjens, mutations affecting the VPS13C protein emerged from whole-genome sequencing on 16 FTD patients who harbored no known pathogenic mutations. The researchers later identified 20 different missense mutations in this gene in 24 of 590 Belgian FTLD patients, absent in more than 1,300 matched controls. The researchers also showed a decreased expression of endogenous VSP13C in lymphoblast cells of mutation carriers. The normal function of the 86-exon-strong VPS13C gene is unknown. However, the closely related gene VPS13A encodes chorein, a protein known to play a role in regulating the neuronal and erythrocyte cytoskeleton, particularly actin filaments. Chorein mutations cause a rare, autosomal-recessive neurodegenerative disease, the movement disorder chorea-acanthocytosis.

The FLNC, whose protein also crosslinks actin and stabilizes the cell’s architecture, came up in a TDP-43 zebrafish model generated by Bettina Schmid and Christian Haass of the German Center of Neurodegenerative Diseases, Munich. This study reported that embryos lacking TDP-43 expressed more FLNC, and the same held true for FTLD-TDP postmortem brains (see Mar 2013 news story). Jonathan Janssens, a postdoc in Van Broeckhoven's group, sequenced FLNC in an initial group of 179 Belgian FTLD patients, and identified 11 missense mutations that did not occur in controls. Expression analysis of lymphoblast cells from mutation carriers showed that FLNC expression was reduced. Conversely, a subset of FTLD patients with GRN mutations had elevated FLNC expression in the frontal cortex. The results could suggest that altered expression of FLNC may somehow hasten FTLD via its interaction with TDP-43.

For her part, Rademakers also made a go at finding new causal genes in a family with unexplained FTD. She performed whole-exome sequencing on three affected relatives and looked for new genetic variants they had in common. She came up with a list of 27, and urged other researchers in the audience to scan their own family cohorts for these 27 in hopes of narrowing down the list. “One of these genes could very well contain the mutation that explains the disease in this family, but we have no idea which one,” Rademakers said. One promising candidate, KIF17, rose to the top of her list based on its function. KIF17 is a motor protein that shuttles a subunit of the NMDA receptor along dendrites to the synapse. The protein may therefore play a role in synaptic transmission.

John van Swieten of Erasmus Medical Center in Rotterdam, Belgium, uncovered a new gene by studying a family with a novel neurodegenerative disease. Twelve of its members dominantly inherited a syndrome marked by dementia and/or parkinsonism, and their brains were riddled with neuronal cytoplasmic inclusions of the intermediate neurofilament α-internexin. Van Swieten and colleagues performed linkage analysis and whole-exome sequencing, which produced thousands of potential variants. To whittle them down, the researchers dissected out neuronal inclusions and parsed their protein content via mass spectrometry. Combining genomics and proteomics, the researchers thus zeroed in on a missense mutation in protein kinase A type I-beta regulatory subunit (PRKAR1B), a protein the researchers found comingling with neurofilaments in the inclusions (see Wong et al., 2014). Van Swieten hypothesized that a missense mutation in the regulatory subunit could trigger mislocalization and heightened activity of the kinase, which in turn could hyper-phosphorylate neurofilaments. So far, this mutation has not turned up in people with FTD, PD, or AD, so van Swieten and colleagues have yet to determine if it is important beyond this single family (see Cohn-Hokke et al., 2014).

J. Paul Taylor of St. Jude’s Children’s Research Hospital in Memphis, Tennessee, shared striking new mechanistic findings on two mutations his lab had previously identified in people with multisystem proteinopathy (MSP) and ALS. MSPs trigger neurodegeneration as well as destruction of bone and muscle. Like people with ALS and subsets of FTLD, people with MSPs have inclusions that contain TDP-43 as well as other RNA-binding proteins that collectively belong to the family of heterogeneous nuclear ribonucleoproteins (hnRNPs). These hnRNPs contain prion-like domains that promote clustering. Taylor discovered missense mutations within the prion-like domains of two such proteins—hnRNPA1 and hnRNPA2B1—in families with MSP as well as one family with ALS (see Kim et al., 2013). Efforts to find the mutations in these particular hnRNPs in other cohorts of MSP, FTD, or ALS patients have come up empty so far (see Le Ber et al., 2014; Seelen et al., 2014). However, mutations in other hnRNPs—hnRNPDL and TIA-1—have been implicated recently in families with different forms of muscular dystrophy (see Viera et al., 2014Hackman et al., 2013), and it is possible that others will follow.

Despite the unknown breadth of hnRNPs’ involvement across the spectrum of neurodegenerative disease, the underlying mechanisms the mutations revealed were startling. At ICFTD, Taylor presented data revealing that the hnRNPs use a so-called steric zipper motif to assemble into lipid-like droplets, which clouded up test tubes and looked like condensation under the microscope. These clusters formed rapidly in the presence of RNA, which the droplets hungrily soaked up. “This family of proteins are able to undergo spontaneous assembly into lipid-like droplets, transitioning between two distinct phases at physiological concentrations,” Taylor told the crowd. Taylor hypothesized that the clusters could act like mini-organelles, ushering RNA through key events in its life cycle within the expanse of the cytoplasm. Normally these clusters disassemble when the steric zippers unzip, but Taylor reported that the disease-associated mutations in hnRNPs prevented their disassembly and led to the formation of cytoplasmic inclusions.

How might this mechanism play out in cases of FTD/ALS where there is not an hnRNP mutation? Taylor went on to demonstrate that the same RNA granules that contained hnRNPs also contained VCP, a protein that causes FTD/ALS in mutation carriers. Depleting granules of VCP prevented their disassembly in cells, and eventually led to apoptosis and death. Taylor proposed that VCP mutations could prevent the physiological disassembly of hnRNP clusters, which would promote the formation of toxic cytoplasmic inclusions. Taylor is still determining whether these hnRNP droplets also contain TDP-43, but his findings, if confirmed, could point to fundamental cellular mechanisms that may underlie a spectrum of neurodegenerative diseases.—Jessica Shugart


No Available Comments

Make a Comment

To make a comment you must login or register.


News Citations

  1. Ataxin 2 Variants Take Aim at Motor Neurons, Not Frontal Cortex
  2. Cloak and Dagger Clusters? How C9ORF72 Repeats Kill Is Still a Mystery
  3. Paper Alert: Zebrafish Say TDP-43 Causes ALS by Loss of Function

Paper Citations

  1. . TMEM106B protects C9ORF72 expansion carriers against frontotemporal dementia. Acta Neuropathol. 2014 Mar;127(3):397-406. Epub 2014 Jan 3 PubMed.
  2. . Genetic modifiers in carriers of repeat expansions in the C9ORF72 gene. Mol Neurodegener. 2014 Sep 20;9:38. PubMed.
  3. . PRKAR1B mutation associated with a new neurodegenerative disorder with unique pathology. Brain. 2014 May;137(Pt 5):1361-73. Epub 2014 Apr 9 PubMed.
  4. . Mutation frequency of PRKAR1B and the major familial dementia genes in a Dutch early onset dementia cohort. J Neurol. 2014 Nov;261(11):2085-92. Epub 2014 Aug 10 PubMed.
  5. . Mutations in prion-like domains in hnRNPA2B1 and hnRNPA1 cause multisystem proteinopathy and ALS. Nature. 2013 Mar 28;495(7442):467-73. PubMed.
  6. . hnRNPA2B1 and hnRNPA1 mutations are rare in patients with "multisystem proteinopathy" and frontotemporal lobar degeneration phenotypes. Neurobiol Aging. 2013 Oct 9; PubMed.
  7. . No mutations in hnRNPA1 and hnRNPA2B1 in Dutch patients with amyotrophic lateral sclerosis, frontotemporal dementia, and inclusion body myopathy. Neurobiol Aging. 2014 Aug;35(8):1956.e9-1956.e11. Epub 2014 Feb 6 PubMed.
  8. . A defect in the RNA-processing protein HNRPDL causes limb-girdle muscular dystrophy 1G (LGMD1G). Hum Mol Genet. 2014 Aug 1;23(15):4103-10. Epub 2014 Mar 18 PubMed.
  9. . Welander distal myopathy is caused by a mutation in the RNA-binding protein TIA1. Ann Neurol. 2013 Apr;73(4):500-9. Epub 2013 Feb 11 PubMed.

Further Reading

No Available Further Reading