25 Apr 2011

Huddled around conference tables in an $80-million building designed by the star architect Frank Gehry, some 65 representatives from academia, industry, advocacy groups, and government agencies met 25-26 March 2011 for dialogue whose concentrated intensity evoked the neon glare of the Las Vegas strip just a few miles south. Their discussions tried to illuminate a path toward treatments for frontotemporal dementia (FTD). FTDs afflict 200 times fewer people than does Alzheimer’s disease, but lead to death sooner and are as common as AD in people 65 or younger. Many FTD patients are still misdiagnosed as having AD or a psychiatric disorder, and frontotemporal dementias have as yet no specific therapies.

However, recent research advances have ignited the field—so much so that last year a handful of academic and pharmaceutical scientists launched a study group to speed drug discovery for these rare disorders. The FTD Treatment Study Group (FTSG) met for the first time in April 2010 at the annual meeting of the American Academy of Neurology in Toronto, then again in October at the 7th International Conference on Frontotemporal Dementias in Indianapolis, Indiana (see ARF related news series). Last month, the fledgling organization held its first independent meeting, “Frontotemporal Dementia: The Next Therapeutic Frontier,” in Las Vegas.

The Frontotemporal Dementia Treatment Study Group held its first independent meeting 25-26 March 2011 at the Cleveland Clinic Lou Ruvo Center for Brain Health, which opened in Las Vegas in 2009. Image credit: Matt Carbone Photography

The conference took place at the Cleveland Clinic Lou Ruvo Center for Brain Health, whose 18,000 curving stainless steel shingles and 199 uniquely shaped windows arguably make it a Lady Gaga among neurodegenerative disease facilities. “You’re in one of the greatest buildings in the world,” director Jeffrey Cummings told attendees. “At this meeting, we should have thoughts commensurate with this building.” To facilitate outré thinking, Cummings and others on the FTSG steering committee packed the one-and-a-half-day agenda with talks on preclinical animal models for FTD and lessons from collaborative therapeutic initiatives in other neurodegenerative disorders, and spiced up discussion with proposals ranging from a Web-based registry of FTD models to an independent clearinghouse for deciding which compounds should get tested in which models. The organizers are summarizing conference proceedings in a position paper to be published in a peer-reviewed journal. In anticipation of this detailed, formally referenced article, this story briefly recaps the key science reported in Las Vegas. See Part 3 for the collaborative drug discovery proposals and pharma’s response to them, and Part 1 for a broader vista on the Lou Ruvo Center and its mission to rejuvenate clinical studies of neurodegenerative disease.

Along with Cummings, the FTSG steering committee consists of Susan Dickinson, Association for Frontotemporal Degeneration; Howard Feldman, Bristol-Myers Squibb; Howard Fillit, Alzheimer’s Drug Discovery Foundation; Michael Gold, Allon Therapeutics; and Megan Grether, Bluefield Project to Cure Frontotemporal Dementia. Adam Boxer, University of California, San Francisco, chairs the committee and organized the scientific component of the recent meeting.

Nine pharmaceutical companies sent representatives to Las Vegas. Boxer explained why their industry should find it in their interest to invest in FTD. Beyond the incentives that come with FTD’s orphan drug status and new diagnostic criteria, frontotemporal dementias progress rapidly—much quicker than AD, whose slow progression has made trials on this disease long and costly. “Theoretically, one could do a shorter trial and have greater power to detect an effect,” Boxer said. People with FTD also tend to be younger and have fewer comorbidities than AD patients. Furthermore, Boxer noted, some FTD syndromes have a clearer link to specific molecular pathology than in AD, where a mixture of amyloid, tau, synuclein, and even TDP-43 pathology may contribute to the clinical phenotype. In this regard, FTD with parkinsonism linked to chromosome 17 (FTDP-17) and progressive supranuclear palsy (PSP)—both pure tauopathies—rank among the top FTD treatment candidates, Boxer said. Another is progranulin-related FTD, which looked therapeutically promising in a subsequent presentation by Fen-Biao Gao, University of Massachusetts Medical School in Worcester.

Speaking on behalf of Feldman, who could not attend, Gary Tong of Bristol-Myers Squibb reminded attendees of the abysmal odds faced by CNS drugs coming through the clinical development pipeline. Of compounds tested in Phase 1 studies, 1 or 2 percent eventually reach market. Among those surviving to Phase 3, a mere 15 percent make it all the way. Drug development for FTD is particularly difficult, in part because of the heterogeneity of clinical syndromes. In addition, Tong said, outcome measures are not optimized, and cellular and animal models are insufficiently validated. “How do we know these models will translate to clinical efficacy?” he asked. “This is a major issue.”

That concern also surfaced in Gao’s talk. A big problem with many FTD models is that they rely on overexpression of proteins in order to produce disease phenotypes, Gao said. He and others skirt this problem by generating patient-specific induced pluripotent stem cells (see ARF iPS series). Collaborating with Bruce Miller and Robert Farese of UCSF, Gao reported collecting skin biopsies from 29 FTD patients and 17 controls. The team used the samples to generate around 200 iPS cell lines, a few of which have been confirmed pluripotent. Some of the iPS lines show reduced progranulin, the gene behind 5 to 10 percent of FTD cases. Scientists can use these cells to screen for compounds that enhance progranulin expression, including an FDA-approved histone deacetylase inhibitor (suberoylanilide hydroxamic, or SAHA), as reported in a freely downloadable paper published online March 23 in the Journal of Biological Chemistry (Cenik et al., 2011).

Also in Las Vegas, Ed Burton, University of Pittsburgh, Pennsylvania, reported on his lab’s efforts to develop tauopathy models in another animal that looks promising for screening potential therapeutic compounds. This is the zebrafish. The scientists made transgenic fish expressing human tau specifically in neurons (Bai et al., 2007). These animals looked normal up until around six months, when they started having trouble breeding. The trouble stemmed from behavioral and motor defects that progressed with age. Consistent with these deficits, the transgenic fish had smaller brains, hyperphosphorylated tau, and more microglia.

While these zebrafish look promising as a tauopathy model, therapeutic studies work more easily in fish that develop pathology at larval stages. “Larvae are possible to screen on 96-well plates,” Burton said. His team has used the Gal4/UAS system to create stable tau transgenic lines with high expression levels in the brain. Researchers led by Christian Haass and Bettina Schmid at the Ludwig-Maximilians University in Munich have used a similar approach to create transgenic fish that express tau in the spinal cord (see ARF related conference story on Paquet et al., 2009).

Are these models ready for pharmaceutical application? “I think we are close with the Gal4/UAS animals,” Burton said. In preliminary studies, he and colleagues have detected 30 percent rescue of movement defects with a sample size of four fish. Computer systems can detect larval eye movements, and these reflexive behaviors can be used to validate compounds that have a motor effect. (For reviews on tauopathy zebrafish models, see Bai and Burton, 2010 and Bandmann and Burton, 2010.)

On the fly front, George Jackson and coworkers at the University of Texas Medical Branch, Galveston, uncovered 41 tau modifiers in a recent genomewide screen covering 20 percent of the Drosophila genome. As expected, GSK3β and other tau kinases appeared among the gain-of-function modifiers. The autophagy gene ATG6 (aka Beclin-1) turned up on the loss-of-function list. The researchers then did an “unbelievable number of Westerns” with the AT8 phospho-tau antibody, hoping to link modifier effects with tau phosphorylation, but, surprisingly, found no such correlation, Jackson reported. Furthermore, his team made phosphorylation-resistant tau mutants and showed they were still toxic. Perhaps phosphorylation “is not a sine qua non for tau toxicity,” Jackson said. “The data suggest that kinase inhibition may not be the best target.”

More complex models were also on the agenda at the Lou Ruvo meeting. Kathleen Zahs works with Karen Hsiao Ashe at the University of Minnesota, Minneapolis, on developing tau transgenic mice. In the group’s Tg4510 mice with regulatable forebrain expression of the FTD-linked P301L human tau mutant, they previously showed that turning off tau expression rescued memory and neurodegeneration even as the mice continued to form tangles (see ARF related news story on Santacruz et al., 2005; Ramsden et al., 2005). This suggested that soluble tau disrupts synaptic structure and function prior to neurodegeneration. The scientists’ hypothesis was proved correct in recent analyses showing that mutant tau mislocalizes to dendritic spines, causing loss of glutamate receptors and weakened synaptic signaling early in disease in the Tg4510 model (ARF related news story on Hoover et al., 2010).

Thus far, research with these mice has focused on mechanisms. In preclinical studies, it is critical to understand what aspect of disease is being modeled in any particular mouse in order to use it appropriately, Zahs said. She noted that APP transgenic mice model largely the asymptomatic phase of AD, making them well suited for prevention, but not treatment, studies. In the case of the Tg4510 mouse, “we see neurodegeneration (likely due to the mutant P301L tau) superimposed upon a developmental delay (likely due to tau overexpression), and any preclinical studies in these mice should take into account that they are not a model of pure neurodegeneration,” Zahs said.

Blair Leavitt, University of British Columbia in Vancouver, Canada, spoke about his conditional progranulin knockout mice, which he first reported at last year’s International Conference on Alzheimer’s Disease in Honolulu (see ARF related conference story). These animals show subtle, sex-specific abnormalities in social behavior, as well as defective neuronal morphology and impaired synaptic plasticity and long-term potentiation; all this happens at eight months of age and about 10 months before evidence of neuropathology in this model. These mice are “not quite ready” for testing potential therapeutics, Leavitt said, “but we are getting to a point where we are finding real and practical endpoints.” Besides being expressed in mature neurons, progranulin normally shows up in microglia (but not astrocytes) and gets upregulated in activated microglia. To explore its role in these macrophages of the brain, graduate student Terri Petkau is generating mice with a conditional progranulin deletion in microglia.

On the theme of microglial-neuronal interactions, recent work by Cleveland Clinic colleagues Bruce Lamb and Richard Ransohoff suggests that signaling through the fractalkine receptor CX3CR1 on microglia could be “an interesting pathway to consider for therapeutic interventions,” Lamb said. The scientists found that mice lacking CX3CR1 show exacerbated tau hyperphosphorylation in response to microglial activation. Similar tau changes appeared in another strain generated by introducing wild-type human tau into a mouse tau knockout. These mice had behavioral impairments that correlated with increased levels of p38 MAP kinase (ARF related news story on Bhaskar et al., 2010). In essence, the scientists believe, healthy neurons signal via this fractalkine receptor to microglia to keep them from getting activated. “If we can induce this pathway, it may have a protective effect in tauopathy models,” Lamb said.

Erik Roberson, University of Alabama at Birmingham, first outlined key challenges for modeling FTD in mice and then allayed each of these concerns with his lab’s recent data on progranulin-deficient mice made by Farese’s group at UCSF (see ARF related conference story) and mutant tau transgenic mice made by Gerard Schellenberg when he was at the University of Washington, Seattle (McMillan et al., 2008). One challenge with mouse models is that FTD involves the prefrontal cortex, a brain area that is dramatically expanded in humans compared to mice. Another is that FTD patients show dysfunctional social and emotional behaviors that seem too complex to find in rodents (see ARF related conference story). How much empathy for others, fidelity to a spouse, or financial prudence does the mouse show in the first place?

To the first point, Roberson reminded attendees that the field is shifting toward a network-based view of FTD. The disease’s primary target is not really the prefrontal cortex, but rather connectivity throughout the brain’s salience network, which includes the anterior cingulate, insula, amygdala, striatum, and brain stem areas—all of which are well represented in the mouse brain (ARF related news story on Seeley et al., 2009). This change in thinking also came up in a talk by Bruce Miller of UCSF. He stressed that brain atrophy patterns for various FTDs map to functional networks whose neurons “are not only born together and fire together, but also die together.”

On the second issue—the complexity of social and emotional behaviors in FTD—Roberson described several outcome measures that do seem to reveal rodent versions of the emotional blunting, social dysfunction, and repetitive behaviors common to FTD patients. On the former, progranulin-deficient mice show unusually high interest in inanimate objects in a sociability test, and subdued responses to aversive stimuli in a fear-conditioning test. Other measures reveal hints of social disinhibition and age-dependent repetitive behavior in tau transgenics. These mice spend considerable time exploring open, exposed arms of an elevated-plus maze—“dangerous” areas that dark-preferring wild-type mice avoid. In addition, tau transgenic mice develop facial lesions from compulsive grooming that intensifies with age. The behavioral impairments seem to correspond with anatomical abnormalities, as these animals show specific electrophysiology and morphological defects in the ventral striatum, the brain area at the root of repetitive behaviors in people with FTD.

Beyond progranulin and tau, the gene for valosin-containing protein (VCP), which is implicated in autophagy, causes some cases of familial FTD. Toward the end of his keynote address, Frank LaFerla of the University of California, Irvine, shared preliminary data on his lab’s brain-specific VCP transgenic mice. Relative to wild-type controls, these animals show spatial memory defects in the Morris water maze, as well as abnormal astrogliosis, protein ubiquitination, reduction of the autophagy marker LC3II, and accumulation of cytosolic inclusions containing TAR DNA-binding protein 43 (TDP-43).

“It’s exciting that there appear to be a number of models at different levels that are looking mature in terms of recapitulating FTD phenotype,” Boxer told ARF.—Esther Landhuis.

This is Part 2 of a three-part series. See also Part 1 and Part 3. See PDF of entire series.