This is Part 3 of a four-part series. See also Part 1, Part 2, Part 4. View PDF of the entire series.
3 November 2010. Recent discoveries of the molecular underpinnings of nearly all frontotemporal dementias (FTDs), as well as at least one genetic cause for each major subtype, have nudged open a door for researchers to begin to unravel disease pathways—a necessary step toward developing new therapeutic approaches. At the 7th International Conference on Frontotemporal Dementias, held 6-8 October 2010 in Indianapolis, Indiana, many scientists presented new molecular and genetic findings on disease mechanisms, and described recently developed animal models for different forms of FTD. The upshot from Indianapolis: It’s too early for consensus pathways to be anointed, but at least scientists now know where to look, and data are beginning to pour in.
Strange Bedfellows: ALS and FTD
In the last few years, researchers have found that the proteins TAR DNA binding protein 43 (TDP-43) and fused in sarcoma (FUS) form deposits in both FTD and in amyotrophic lateral sclerosis (ALS), suggesting these disorders have more in common than was once thought. The disorders can also occur together, said Catherine Lomen-Hoerth of the University of California in San Francisco; up to 30 percent of ALS patients develop dementia, and up to 15 percent of FTD patients get ALS. “The thing that was most interesting to me [at the conference] was to see how the ALS field has now sort of merged with the FTD field,” said Rosa Rademakers of the Mayo Clinic in Jacksonville, Florida. One session of the meeting dealt especially with this FTD/ALS connection.
Teepu Siddique of Northwestern University in Chicago, Illinois, discussed evidence that TDP-43, FUS, and a third protein, optineurin, act in a common pathological pathway in both familial and sporadic ALS (see ARF related news story). Mutations in SOD1 cause many cases of familial ALS; however, aggregates of SOD1 protein do not include TDP-43 or FUS, Siddique said, and therefore SOD1 may represent a separate disease pathway. Recent work in the field suggests that SOD1 may have a role in sporadic ALS as well (see ARF related news story).
TDP-43 and FUS are both RNA/DNA binding proteins found predominantly in the nucleus, said Christopher Shaw of King’s College London, U.K. He suggested that mislocalization of these proteins to the cytoplasm might drive the disease process (see Rogelj et al., 2010 and also ARF related news story on FUS localization). Shaw and colleagues expressed mutant FUS in cell lines, and confirmed that the protein accumulated in cytoplasm. However, only a tiny percentage of ALS cases have mutations in their genes for FUS or TDP-43, Shaw said. He speculated that defects in a person’s nuclear import/export machinery might contribute to non-genetic forms of the disease, and is pursuing this possibility.
One of the most tantalizing genetic mysteries in this field at present is the association of a region on chromosome 9 with both FTD and ALS cases. It shows up in multiple linkage studies, but researchers have been maddeningly unable to pinpoint the gene hiding in there. The data are consistent with a single Scandinavian founder, said John Hardy of University College London, U.K. The locus covers three genes, and no one has yet reported a mutation that consistently segregates with the disease, although speakers discussed their efforts to find it.
More Evidence for Risk Factor TMEM106B
Several speakers, such as Hardy, Andrew Singleton of NIH in Bethesda, Maryland, and Gerard Schellenberg of the University of Pennsylvania in Philadelphia, talked about genomewide association studies (GWAS) that are underway to turn up risk factors for various forms of frontotemporal lobar degeneration (FTLD). These should provide further clues to the molecular mechanisms of the disorder. Vivianna Van Deerlin of the University of Pennsylvania discussed the previously reported identification of membrane protein TMEM106B as a risk factor for FTLD-TDP in an international GWAS (see ARF related news story). People who carry the high-risk allele show higher TMEM106B expression, Van Deerlin said, and the allele associates with mutations in the progranulin gene.
Julie van der Zee of the University of Antwerp, Belgium, announced that her group replicated this finding in a Flanders-Belgian cohort, where two copies of the TMEM106B risk allele double the risk of FTLD-TDP. Van der Zee added, though, that in her study the RNA levels of TMEM106B and disease risk did not correlate. NiCole Finch of the Jacksonville Mayo Clinic provided some clues to the mechanism with her report that genetic variants of TMEM106B regulate plasma progranulin levels in people with mutations in the progranulin gene. The protective minor allele of TMEM106B leads to less TMEM106B expression, more progranulin expression, and a later age of onset of FTLD, Finch said.
Animal Models Galore
Ever since TDP-43 popped up in FTD research, scientists have raced to develop overexpression, knockout, and mutant transgenic animal models for the disease. For a fairly complete summary of sundry TDP-43 mouse models, as discussed at the International Conference on Alzheimer’s Disease 2010, see ARF related news story. At FTD 2010, speakers put forth several more, reflecting a smorgasbord of approaches.
One strategy models FTLD-TDP by tinkering with the progranulin gene, the cause of many human cases of the disease. Lauren Herl, working with Robert Farese at UCSF, presented a mouse progranulin knockout. Both homozygous and heterozygous knockouts exhibited social deficits and impaired fear conditioning, suggesting they model features of FTD, Herl said. Both strains also lost more dopaminergic neurons and activated their microglia more potently after administration of a neurotoxin than wild-type mice did. The results suggest that progranulin could promote neuronal survival by suppressing inflammation, Herl said.
Philip Van Damme of the Catholic University of Leuven, Belgium, investigated the effects of progranulin in zebrafish. He used morpholinos, a form of antisense RNA, to knock down endogenous progranulin. In the fish with less progranulin, axons grew poorly. The same thing happened in zebrafish that overexpressed wild-type or mutant human TDP-43, Van Damme said. Injecting either of these models with human progranulin rescued axon growth, confirming progranulin’s importance in the pathway (see Laird et al., 2010).
Continuing the model menagerie with flies, Jane Wu of Northwestern University described a series of Drosophila strains that express wild-type human TDP-43 in neuronal subpopulations and develop neurodegeneration (see Li et al., 2010). When the flies express mutant human TDP-43, their neurodegeneration is more severe (unpublished data). That even wild-type TDP-43 is toxic suggests to Wu that both overproduction or delayed clearance of TDP-43 could lead to disease.
Researchers in the laboratory of Virginia Lee of the University of Pennsylvania developed transgenic mice that overexpress wild-type human TDP-43 or mutant human TDP-43 with a defective localization signal. The mice develop rare TDP-43 deposits, but profound neuron loss and motor defects. Endogenous mouse TDP-43 was downregulated in these transgenics, Lee said, which suggests the mice may be modeling a loss of function of TDP-43, rather than a toxic gain of function. The mice Lee described share many features with transgenic human TDP-43 mice made by Leonard Petrucelli of the Jacksonville Mayo Clinic (see ARF related news story). Petrucelli also discussed a C. elegans model that overexpresses human TDP-43 and has a motor defect (see Ash et al., 2010), as well as the expression of truncated TDP-43 in a cell culture model that lends support to a toxic gain-of-function mechanism (see ARF related news story).
Have the animal models available to date produced any overarching insights? “I think there’s a consensus in the field that basically all the transgenic mouse models we’ve developed by overexpressing wild-type protein or mutant proteins lead to loss of function,” Lee said. “In other words, the cells rapidly die. So the challenge for the future is, Can we produce animal models that develop pathology, in addition to the loss of function?” Although many of the mouse models develop TDP-43 deposits, Lee said, these deposits are quite rare and cannot account for the massive neuron loss. What might explain it? Lee summed up the current data: “If you overexpress TDP, you get neuron loss; if you knock down TDP, you get neuron loss, so that suggests that TDP-43 is very finely regulated.” In support of the idea that regulation of TDP-43 levels is crucial for brain health, Emanuele Buratti of the International Center for Genetic Engineering and Biotechnology in Trieste, Italy, found that TDP-43 can regulate its own levels through a negative feedback loop. The protein’s sequestration in aggregates may co-opt this self-regulation and result in its overproduction, Buratti speculated.
Rademakers said the value of these model organisms is they will allow scientists to begin to tease out the functions of TDP-43 and progranulin. “Now we can start doing very specific studies and hopefully find some new targets for therapy using these models,” she said. Many researchers are already doing this. For example, Philipp Kahle of the University of Tubingen in Germany examines the physiologic role of TDP-43. In Indianapolis, he presented data to suggest that one of its targets is HDAC6, a histone deacetylase enzyme that is also a major mediator of toxic protein turnover in the cell. Higher levels of TDP-43 reduce HDAC6 levels and activity, Kahle said, which in turn leads to increased aggregation of toxic polyQ proteins, the hallmark of several neurodegenerative diseases, including Huntington’s disease (see Fiesel et al., 2010).
Adrian Isaacs of University College London pointed out that several genes in the endosomal sorting pathway are mutated in human neurodegenerative diseases. He cited as examples the genes VCP, CHMP2B, Alsin, Rab7, Fig4, and Spastin, associated with a gamut of neurological conditions including FTDs, upper motor neuron disease, peripheral neuropathies, and paraplegia. A large body of evidence implicates sorting proteins such as SorL1, sortilin, and Vps35 in Alzheimer’s disease (e.g., see ARF related news story and ARF news story). Isaacs suggested that defects in endosomal sorting and protein degradation might be a common disease mechanism in neurological disorders. Supporting this idea, Rademakers announced the discovery of yet another endosomal sorting gene as a regulator of plasma progranulin levels (paper in preparation—stay tuned).
Do Neurodegenerative Proteins Act Like Prions?
A repeated finding from imaging studies is that pathology can spread through brain networks, but how that happens is unknown. Michel Goedert of the MRC Laboratory of Molecular Biology in Cambridge, U.K., presented his finding that tauopathies can be transmitted from cell to cell in a prion-like fashion. It is based on experiments in which brain extracts from diseased mice seeded tau aggregation in healthy mice overexpressing tau (see ARF related news story). Goedert also described new work in which extracts from human patients with various FTDs were injected into the tau-overexpressing mice, with disease-specific results. For example, mice that received extracts from people with a dementing tauopathy called argyrophilic grain disease developed grain-like structures that mirrored the human disorder. “It is fascinating that there is some sort of specificity in terms of the pathology,” Lee said. “[Goedert] was able to show that the pathologies that developed in the mice recapitulated those of the human disease.” Goedert’s work adds evidence to the emerging idea that many pathological proteins, such as Aβ, huntingtin, and α-synuclein, can propagate themselves in a prion-like way by seeding misfolding and aggregation (see ARF Live Discussion).
Masato Hasegawa of the Tokyo Institute of Psychiatry, Japan, suggested that a similar process may happen in FTLD-TDP and ALS. He showed in cell culture that C-terminal fragments of TDP-43 can form inclusions that recruit full-length TDP-43 (see Arai et al., 2010). TDP fragments vary from patient to patient, but are indistinguishable between the brain regions of a given patient, Hasegawa said. This suggests that the pathology may propagate from cell to cell. Hasegawa referred to amyloid-like proteins in neurodegenerative diseases as “protein cancers,” and speculated that they might be transmitted through synapses. The audience was intrigued by the hypothesis, but wanted to see more cases than the handful presented. For a discussion of possible FTD therapies and current clinical trials, see Part 4 of this series.—Madolyn Bowman Rogers.
This is Part 3 of a four-part series. See also Part 1, Part 2, Part 4. View PDF of the entire series.