Tau Snapshots from Neuroscience 2017
This year’s Society for Neuroscience meeting, held November 11–15 in Washington, D.C., drew 23,526 attendees from around the globe. For researchers in the Alzheimer’s disease field, the meeting has lost some of its vibrancy of late, and this year seemed no exception. Only one of 10 press briefings featured AD, a nanosymposium on “tauopathies” focused on brain effects of anesthetics and only two of its nine abstracts mentioned tau. Many leading AD researchers nowadays devote their travel to other meetings, but SfN remains a must for the basic science relevant to AD, and for students with broad interests across neuroscience.
This year, our SfN news coverage balances two stories on emerging areas—sleep and the microbiome in brain health and AD (see Parts 2 and 3 of this series)—with two stories on preclinical and mechanistic advances on the old but insufficiently understood topic of tau. Some researchers tried to curtail toxic forms of tau using intrabodies and zinc-finger suppressors (see below). Other labs revealed how tau compromises mitochondria, the nucleus, and the vasculature, and how microglia and synapses may contribute to the prion-like spread of tau in the brain (see Part 4 of this series).
Zinc-Finger Vectors. AAV vector injected into the hippocampus (left) limits zinc finger to that locale, whereas a single retro-orbital injection of a PHP.B vector achieves brain-wide expression (right). [Courtesy of the Hyman lab.]
Cool Ways to Temper Tau
Sarah DeVos and Susanne Wegmann from Bradley Hyman’s lab at Massachusetts General Hospital, Charlestown, described a way to cut tau production in vivo using a zinc-finger-based gene-suppression system. DeVos and Wegmann collaborated with Bryan Zeitler and colleagues at Sangamo Therapeutics, a Richmond, California-based company that develops this technology for a variety of conditions. The researchers fused zinc-finger (ZF) proteins that specifically recognize the tau gene onto a transcriptional repressor. The Kruppel-associated box, aka KRAB domain, of the human ZF protein KOX1 served as the business end of these tau repressors. KRAB domains enlist co-repressors in the nucleus to shut down transcription.
Wegmann and DeVos tested different suppressor and Adeno-associated virus (AAV) vector combinations. Out of three different repressors, one dubbed ZFP-TF89 worked best to specifically knock down tau. Wegmann made a construct to drive expression of ZFP-TF89 from a cytomegalovirus promotor, and packaged it into an AAV9 virus. A nuclear localization signal ensured this engineered suppressor would be delivered next to DNA, while a Venus yellow fluorescent protein allowed her to track infected cells. In vitro, Wegmann saw increasing knockdown of tau with increasing dose of the vector.
To try ZFP-TF89 in wild-type mice, Wegmann injected the AAV vector into the hippocampus on one side of the brain and saline into the other. Venus fluorescence confirmed widespread ipsilateral expression of the vector in the hippocampus. Immunohistochemistry, western blot, and mRNA analysis indicated about 88 percent suppression of endogenous tau expression six weeks after injection. Remarkably, even 11 months later, tau mRNA levels were still down by 75 percent, while the neuronal marker NeuN indicated no major change in neurons. However, Wegmann saw a transient uptick in glial fibrillary acidic protein and in microglial activation, suggesting the AAV vector may cause a mild inflammatory reaction. A control vector expressing only GFP evoked a similar reaction.
To avoid untoward glia responses, the scientists swapped out the CMV promoter, which turns on ZFP-TF89 in most cell types, for promoters specifically active in neurons, including those for synapsin-1, CaMKIIa, and MeCP2. These barely expressed ZFP-TF89 in glia while efficiently suppressing tau in N2a neuroblastoma cells as well as in neurons in vitro and in vivo.
Would the vectors protect mice from AD pathology? Wegmann tested them in 4.5-month-old APPPS1 animals, which overproduce Aβ42. Previously, several groups had reported that Aβ was much less toxic when tau was knocked down or knocked out (May 2007 news). Wegmann injected AAV containing synapsin-1 ZFP-TF89 into one side of the brain, and a control vector expressing red fluorescent protein into the other. Ten weeks later, she recorded 35 percent fewer dystrophic neurites on the treated side of the brain than on the control side. Both the amount and size of Aβ plaques were the same on both sides of the brain. Wegmann said this fits with the idea that knocking down tau protects against various manifestations of Aβ toxicity, including seizure, hippocampal atrophy, and learning and memory deficits. Zeitler claimed this ZFP-TF based approach is the first to target all tau forms within neurons using a single AAV administration.
Could such a system someday treat tauopathies? It’s possible—the FDA just approved an AAV-based gene therapy for treating retinal dystrophy, and an AAV9-based therapy for spinomuscular atrophy 1 made runner-up for breakthrough of the year in Science magazine (see press release; Science, 22 Dec 2017).
Still, the viruses generally don’t express that broadly in the human brain, and researchers are seeking alternatives. Last year, Benjamin Deverman and colleagues at the California Institute of Technology, Pasadena, reported that a viral capsid variant called AAV-PHP.B transfers genes throughout the CNS about 40 times more efficiently than the commonly used AAV9 (Deverman et al., 2016). When used with the neuron-specific synapsin-1 promoter, these researchers were able to widely express a TDP43 transgene in neurons in a mouse brain after a single intravenous injection of the virus (Jackson et al., 2016).
Taking advantage of this, DeVos tested a synapsin1 ZFP-TF AAV-PHP.B construct in mice. She reported that one week after a single retro-orbital injection to place the virus into the capillary bed behind the eye, tau mRNA and protein levels fell by up to 70 percent across the brain and spinal cord. CSF tau plummeted by 80 percent after 10 weeks and this persisted for six months until the end of the study.
DeVos injected 4.5-month-old APPPS1 mice once with AAV-PHP.B capsids carrying the ZF suppressor. Much like Wegmann, she found a 50 percent reduction in neuritic dystrophies around Aβ plaques 2.5 months later.
Researchers at SfN asked about off-target effects or changes to DNA. DeVos emphasized that the ZF strategy does not alter the DNA; it only prevents transcription, and that the transcription factor only binds the tau gene. The researchers tested for changes in cellular gene expression with Affymetrix microarrays, said DeVos, and found the tau suppression to be highly specific, with extremely low-off target gene modulation in the two regions they have examined so far, the frontal cortex and the hippocampus.
Others wondered if expression of the zinc finger in the periphery might be a problem, and DeVos noted that when using the synapsin-1 promoter, very little repressor is made outside the brain. Zeitler emphasized that they use human proteins for the transcription factor component, minimizing any chance for an immune response. How about microtubule function? Might it be compromised? The lab is working on this question, but DeVos noted that knocking down tau, or even completely knocking it out, has little effect (Aug 2013 news).
Another hybrid approach for targeting tau came from Marshall Goodwin, from Todd Golde’s lab at the University of Florida, Gainesville. He uses intrabodies, single-chain antibodies that can be expressed inside cells. The rationale is that because tau aggregates intracellularly, intrabodies could be better at binding and removing it than regular antibodies, which don’t easily penetrate cells, or even the brain.
Goodwin and colleagues first used tau antibodies generated by Peter Davies, including CP13, PHF1, and Tau5. They spliced DNA for the variable motifs from the heavy and light chains of the antibodies, and packaged them into viral vectors. When injected into the brains of newborn P301L mouse pups, these tau-transgenic mice expressed single-chain (scFv) intrabodies widely in the CNS, said Goodwin. However, while the CP13 intrabody reduced the amount of tau aggregates by six months, it did not make the mice live much longer, extending their survival from 10.2 months only to about 12. “These intrabodies would need to be optimized for therapeutics,” said Goodwin.
Goodwin has since fused the intrabodies to other domains to promote degradation of tau. He coupled the CP13 intrabody to the catalytic domain of several E3 ubiquitin ligases. He predicted that once these hybrid intrabodies bound tau, the ligases would tag it with ubiquitin, signaling the proteasome to degrade it. Similarly, in another hybrid, he spliced the scFV to a motif that targets proteins for chaperone-mediated autophagy (CMA).
Golde’s group has set up a three-step system to expedite screening of these antibodies. They test them first in HEK models of tau aggregation, then in wild-type brain slices infected with tau genes that created neurofibrillary tangles, and then finally in a mouse AAV model that accumulated tangles within three months. So far, Goodwin has tested the modified intrabodies only in the cell model. While the CP13 intrabody decreased insoluble tau in the cells, adding the CMA motif almost totally ablated tau aggregates. Several of the ligase versions also outperformed the normal intrabody.
Researchers at the meeting considered the approach exciting, but wondered about safety. Goodwin said that the parent intrabodies have been tested in normal mice and do not appear to be toxic and are being tested for any effects on behavior. He plans to test the modified versions in neurons and in vivo.—Tom Fagan
- Disturbed Sleep Exerts Toll on Memory and Neurodegeneration
- Gut Microbiome May Modify Neurodegeneration
- Is There No End to Tau’s Toxic Tricks?
- APP Mice: Losing Tau Solves Their Memory Problems
- In Adult Mice, Reduced Tau Quiets Agitated Neurons
Research Models Citations
- Deverman BE, Pravdo PL, Simpson BP, Kumar SR, Chan KY, Banerjee A, Wu WL, Yang B, Huber N, Pasca SP, Gradinaru V. Cre-dependent selection yields AAV variants for widespread gene transfer to the adult brain. Nat Biotechnol. 2016 Feb;34(2):204-9. PubMed.
No Available Further Reading
No Available Comments
Make a Comment
To make a comment you must login or register.