Dillin and colleagues have reported previously that the Daf2 pathway controls aggregation and disaggregation of Aβ in worms (see ARF related news story). This Daf2 effect disappears if either the transcription factor Daf16 or the chaperone Hsf1 is also knocked out. Delving deeper, the Dillin lab discovered that the two knockouts affect Aβ dynamics quite differently, however. The Daf16 knockout results in fewer aggregates of high molecular weight, while the Hsf1 knockout results in more of them. Dillin concluded that Daf16 drives an aggregation pathway, while Hsf1 drives a disaggregation pathway. Relieving suppression of both pathways by knocking out Daf2, which lies upstream of Daf16 and Hsf1, offers maximal protection against Aβ toxicity, he said, presumably by removing the most toxic oligomeric species of Aβ.

Dillin next showed data to test the idea that Daf16 acts as an “aggregase.” He is using an in-vitro assay for aggregation into which he plants worm extract. The idea is that extracts from worms with less aggregase activity should contain fewer seeds and therefore accelerate aggregation poorly. This was borne out in tests of Daf16 knockouts—these animals have fewer aggregates and extracts from them seed aggregation poorly. Daf2 knockouts have more aggregates, however, and their extracts seed potently. To test the disaggregase idea, Dillin’s lab employs the reverse, i.e., measuring loss of thioflavin T fluorescence as Aβ aggregates are demolished. Extracts from wild-type and Daf16 knockout worms did have disaggregase activity, and boiling the extract abolished that, suggesting that the phantom disaggregase may be a protein. Hsf1 knockout animals, however, had much lower disaggregase activity than wild-type, in keeping with the idea that Hsf1 activity somehow promotes disaggregation.

Could the same mechanism that protects worms against Aβ toxicity be at play in mammals? To test this, Dillin and colleagues crossed an APP/PS1 mouse model (see Borchelt et al., 1997) with insulin-like growth factor receptor (IGFR) heterozygote (homozygote knockouts are lethal). While the AD/IGFR animals do poorly in a Morris water maze test of spatial learning and memory, AD/IGFR+/- mice escape from the maze about four times faster, suggesting that lowering IGF signaling does indeed protect mice against Aβ toxicity. To examine this in detail, Dillin and colleagues looked at the brains of the mice. They found that at 12-13 months, the AD/IGFR heterozygotes had more NeuN staining than the AD control mice, indicative of less neuronal damage, as well as less evidence of astrocyte activation. As for plaques, though their numbers were equivalent in both strains, the plaques were much denser in the heterozygotes. Dillin interpreted this to mean that these animals are packing the Aβ better, or that they have a different plaque morphology.

This then raised the question of whether the packing or morphology of plaques might have anything to do with aggregase/disaggregase activity. To address it, the researchers turned to the seeding assay and showed that extracts from the AD/IGFR+/- mice seeded better than from the AD control mice. “It seems that in these [AD/IGFR+/-] animals, the Daf16 pathway is going full blast and protecting in a better way than in controls,” Dillin said. Exactly why or how the Aβ dynamic shifts is unclear; however, Dillin showed that brain extract of the protected (IGFR+/-) animals contained fewer Aβ dimers and trimers. That not only supports the idea that aggregation is going full tilt, but given recent findings linking dimers with toxicity in humans (see ARF related news story), it may also help explain why the IGFR+/- animals show less cognitive impairment.

Whether downregulating IGFR signaling might have the same effects in humans is not clear. Interestingly, mutations in human IGFR have been found in the very oldest people (see ARF related news story), suggesting that signaling through this receptor affects longevity in us, as well. But there are reasons to believe that IGFR signaling is beneficial, too. In fact, because of its neurotrophic effects, insulin-like growth factor 1 (IGF-1) has been tried as a potential treatment for motor neuron disease. Unfortunately, clinical trials have been disappointing (see ARF related news story). Similarly, Merck’s MK-677, which boosts circulating IGF-1 levels, has been tested on AD patients, again with disappointing results (see same ARF related news story). The theory was that because AD patients are insensitive to insulin and IGF-1, have low circulating levels of the latter, and because IGF-1 seems to promote clearance of Aβ from the brain in rodents (see Carro et al., 2002), MK-677 might offer hope of a treatment.

In a final word on the link between IGF signaling and AD, Dillin hinted that knocking out the pathway early on may not be a prerequisite for protection. In worms, blocking Daf2 even after Aβ toxicity has begun stops progression of disease, even if it doesn’t bring the animals back to normal. He plans to try the same approach in mice using conditional IGFR knockouts that can be turned on and off.—Tom Fagan.