14 December 2010. Since the approval of cholinesterase inhibitors and memantine, no new therapy for Alzheimer’s disease has passed the clinical trial litmus test. But that does not stop ideas from coming. At the 40th annual meeting of the Society for Neuroscience (SfN), held 13-17 November 2010 in San Diego, California, researchers laid several new strategies on the table, from glucagon-like peptide agonists that tweak insulin signaling pathways (see ARF related news story), to molecular tweezers and other compounds that prevent amyloid-β oligomerization and toxicity. Numerous compounds have been identified that seem to block Aβ aggregation, but the mechanism is often unclear and nothing has made it into the clinic. “The important thing is that we invest in trying to understand how these compounds work,” said Gal Bitan, University of California, Los Angeles, who chaired an SfN nanosymposium dedicated to therapies targeting Aβ and Aβ oligomers. “It is crucial to find out what those molecules bind, where they bind, and how they inhibit. This will help us find compounds that will be a success,” Bitan told ARF.
Case in point are small fragments from the C-terminus of Aβ42 (Aβ28-42 and smaller) that block Aβ aggregation (see Fradinger et al., 2008). Bitan co-founded Clear Therapeutics, and holds patents on lead compounds for AD therapy. Working in Bitan’s lab, Huiyuan Li and colleagues have used rational drug design to generate second-generation versions of these small peptides, systematically modifying the amino acids and looking at structure-function relationships. In her talk, Li noted that Aβ39-42 potently inhibits Aβ neurotoxicity, but is likely easily degraded and would make a poor drug candidate. Since those amino acids have mostly aliphatic side chains, one would not expect them to do anything other than act as a hydrophobic glue, said Bitan. However, a D-amino acid analogue of the peptide had little anti-Aβ activity, said Li, which suggests that specific peptide-peptide interactions are crucial for inhibiting Aβ neurotoxicity. Li showed that the side chain of isoleucine 41 is necessary, for example, and she has a variety of new derivatives that are being tested. So far, data from light-scattering experiments suggest that these compounds work by facilitating formation of non-toxic “hetero-oligomers” composed of a mixture of Aβ and the short peptide molecules. At SfN, collaborator Birgita Urbanc at Drexel University, Philadelphia, Pennsylvania, reported that the modified C-terminal peptides relax the β-strand structure of the full-length Aβ and hinder its aggregation.
Another strategy the Bitan lab pursues is to use molecular “tweezers” to block Aβ from forming multimers. In her talk, Aida Attar described the tweezers as C-shaped molecules that wrap around lysine side chains. Originally introduced by collaborators Thomas Schrader and Frank-Gerrit Klärner, University of Duisburg-Essen, Germany, the tweezers have hydrophobic arms that interact with lysine’s butylene side chain. A negatively charged cavity that separated the arms sits over lysine’s positively charged amino group. With this structure, the tweezers bind to lysine with high affinity, said Attar. Bitan, Schrader, and colleagues have been modifying these compounds to find molecules that can reduce protein aggregation. Attar focused on CLR01, a compound showing some efficacy in transgenic AD mice and in-vitro assays, she said.
In thioflavin T fluorescence tests of protein aggregation carried out by coauthor Sharmishta Sinha, CLR01 blocked the assembly of Aβ into β-sheet structures. Electron microscope pictures showed that the compound prevented Aβ from forming fibrillar structures in vitro. Similarly, while Aβ incubation normally yields oligomers that react with the A11 antibody on dot blots (A11 is specific for protein oligomers), Attar and Sinha found no A11 cross-reactivity when CLR01 was in the mix (the 6E10 antibody that reacts with most forms of Aβ tested positive). CLR01 also blocked tau aggregation as judged by thioflavin assays and electron microscopy, Attar said.
At concentrations up to 400 μM, CLR01 had little effect on the viability of PC12 neuroblastoma cells, but at substantially lower concentrations it completely rescued cell loss driven by 10 μM Aβ. The compound almost completely prevented Aβ-induced dendritic spine loss in primary hippocampal neurons, as well. And in single-cell recordings in primary hippocampal neurons, CLR01 rescued loss of mini-excitatory post-synaptic current (mEPSC) frequency that occurs in the presence of Aβ.
Early indications are that the compound can prevent protein aggregation in vivo, too. Attar and colleagues used subcutaneous administration of CLR01 into 14- to 16-month-old triple transgenic mice (see Oddo et al., 2003). On autopsy 28 days later, plaques and tangles were reduced compared to placebo-treated controls. The reductions occurred in most affected regions of the brain, including the CA1 of the hippocampus. The researchers have no behavioral data because they did not find any memory deficits in the 3xTG animals at the recommended age for study, said Attar. (Other researchers have also reported loss of the memory phenotype in these mice.)
Whether these compounds will make it into human testing remains to be seen. Their ability to bind generically to any lysine might raise safety flags, requiring preclinical controls to ensure that they do not interfere with normal protein function. Attar reported that at 30 times the effective dose, CLR01 had no obvious toxic effects on mice.
Andreas Müller-Schiffmann presented a different strategy for blocking Aβ oligomerization. Müller-Schiffmann works with Carsten Korth at the Heinrich Heine University of Düsseldorf, Germany. The researchers are taking a synergistic approach, combining two different classes of molecule that independently prevent Aβ aggregation in the hope of finding a much more potent compound. One class, aminopyrazoles (APs), are small molecules that bind the cross-β-sheets stabilizing β amyloid and other amyloid structures. APs, which like the molecular tweezers were developed by Schrader’s group, recognize the KLVFF amino acid motif of Aβ, said Müller-Schiffmann, but while they decrease Aβ aggregation in vitro, they work less well in vivo.
The researchers combined APs with a dodecapeptide called D3 that was found in a mirror phage display screen by researchers at Dieter Willbold’s group, also at Heinrich Heine University (see ARF related news story). Molecular simulation suggested to the German scientists that adding APs to D3 would yield a compound that bound very stably to Aβ and could disrupt aggregation. A hybrid molecule, JM169, then prevented oligomerization of Aβ secreted from 7PA2 cells, which are used as a natural source of Aβ (see Walsh et al., 2002). Neither D3 nor aminopyrazoles by themselves had strong effects. The researchers reported some of these results in the November Angewandte Chemie (see Müller-Schiffmann et al., 2010).
In addition, JM169 rescued long-term potentiation deficits in hippocampal slices treated with Aβ. The hybrid also partially restored mESPC in these slices. Bitan complimented this rational approach to drug design and the synergy in the hybrid, but wondered whether the size of the molecule might limit its use in vivo.
In her talk, Susan Catalano from Cognition Therapeutics, Inc., Pittsburgh, Pennsylvania, took a different tack. Instead of addressing Aβ oligomerization/aggregation, Catalano focused on preventing downstream effects. Aβ oligomers affect the rate of membrane trafficking, for example, which is essential for proper synaptic transmission and plasticity. “Depending on the Aβ oligomer, they can accelerate exocytosis, they can inhibit endocytosis, they have a variety of different actions on membrane trafficking,” she told ARF. Catalano’s team used a membrane trafficking screen (dyed cycling vesicles) to search for compounds that prevent Aβ oligomers from interfering with these processes. The researchers tested a variety of Aβ oligomer preparations, including those derived postmortem from human AD patient brain. “We use every Aβ oligomer preparation we can get our hands on, because though there are certain hypotheses about what the active oligomer species is, I don’t think it is yet established which ones play what role at what point in the development of AD. So we take a comparative approach,” Catalano told ARF.
One of the compounds, CT0093, restored trafficking in primary hippocampal neurons that were treated with 130 pM Aβ oligomers derived from human tissue but had no effect in the absence of Aβ. Catalano said that the CT0093 seems to work by blocking Aβ from binding the cell surface. It also prevented synapse loss in Aβ-treated neurons.
These compounds seem to be effective in vivo. Catalano reported that they prevent learning and memory deficits in an acute toxicity model involving injection of Aβ into wild-type mouse hippocampus. “We find that pharma is increasingly interested in acute models,” Catalano told ARF. “What is going on with transgenic animals is not exactly clear, and the wild-type background offers advantages, including the ability to look at off-target effects. Clearly, both transgenic and wild-type animals are important.”
When the compounds were injected one hour prior to the Aβ oligomers in this model, the mice performed as well as controls in a fear-conditioning test of learning and memory, whereas mice treated with oligomers only did poorly. The company is currently testing the compounds in transgenic models.
Catalano does not know which cell surface receptors these compounds bind. They represent a first in class and readily cross the blood-brain barrier. The company is continuing preclinical studies, she said, but it will be a while before the required safety studies in advance of any clinical trial are done.—Tom Fagan.