What is the latest on the immunization of mice and men with Aβ? Will we be celebrating victory over Alzheimer’s disease in a few years (and looking for new jobs), or has the story been overrated? One would predict to see a number of posters and talks at this year’s meeting on Alzheimer’s immunization paradigms, and we weren’t disappointed. Which form of immunization works best, passive or active? Are the results similar in APP-Swedish mice versus APP-PS1 doubly transgenic mice? Are there clear behavioral impairments in AD-transgenic mice and can these deficits be reversed by immunization?
Two recent reports have raised hope that immunization with Aβ peptides may block deposition of β-amyloid or even promote removal of existing plaques. Schenk and colleagues (Nature 1999) were the first to show that active immunization of PDAPP transgenic mice (which overexpress a mutant form of APP) with Aβ peptide could prevent or reverse the accumulation of β-amyloid plaques. If immunization was given prior to plaque formation (six weeks), virtually no plaques were detected in adult mice. Later immunization greatly reduced the amount of plaque material if administered after plaque formation was well under way (11 months). Weiner and Lamere from Selkoe’s group have reported similar results in PDAPP transgenic mice when the antigen is administered via a nasal route (Annals of Neurology 2000).
With this background, some key questions are whether these results would be replicated, and replicated in other transgenic models, whether passive immunization can yield similar results, whether there is evidence for IgG mediated clearance of amyloid, and whether there are behavioral consequences to immunization therapy.
The short summary is that yes, immunization with Aβ leading to a reduction in plaques has been replicated now by other groups in both PDAPP mice and APP-PS1 doubly transgenic mice. Both active (using Aβ peptide as antigen) and passive (injecting anti-Aβ IgG produced in another animal) methods are effective. Fc receptor anti-Aβ IgG mediated clearance of Aβ via microglia has been demonstrated. And clear behavioral consequences have been demonstrated. Since it would be too time consuming to report on each of these aspects, the remainder of this report focuses on two behavioral studies demonstrating that immunization with Aβ attenuates cognitive deficits in Tg AD-like mice.
Yu and colleges at the University of Toronto presented an excellent poster (Abstract 181.18) on a long-term behavioral study of APP-CRND8 transgenic mice (C3H/B6) overexpressing (by fivefold) the Swedish mutant form of APP. At 11 weeks, these mice show increases in levels of soluble Aβ via SDS gels and cortical Aβ positive plaques. They first demonstrated that injection of complete Freund's adjuvant or incomplete adjuvant had no effect on learning acquisition or reversal learning in control mice. After examining a variety of tasks, they determined that APP-CRND8 Tg mice exhibited an age-related cognitive impairment in acquisition and reversal of spatial learning using a reference memory version of the Morris water maze. They replicated the immunization protocol of Schenk (1999) in six-week-old mice (e.g., injections at 6, 8, 12, 16 and 20 weeks) using Aβ42 and islet associated polypeptide (IAPP) as an additional control. At 11 weeks, all mice received nonspatial pre-training in the water maze.
At 11 weeks, there was already a difference in the performance of Aβ immunized Tg mice when compared to IAPP immunized mice, although Aβ immunized mice did not perform as well as non-Tg controls. The Aβ immunized mice continued to improve on the task at 15, 19 and 23 weeks, while the IAPP immunized mice improved somewhat at 15 weeks, began to get worse at 19 weeks and much worse at 23 weeks. At 23 weeks, IAPP-mice could learn, but required more sessions. And at no time did the IAPP mice perform as well as the Aβ immunized mice. By the end of testing, the Aβ immunized mice were performing almost as well as the non-Tg mice. Following sacrifice at 25 weeks, Aβ load was measured in hippocampus and cortex, and both regions had significantly less Aβ deposition compared to IAPP-mice. Because there were only six Aβ immunized mice, a correlation between Aβ load and cognitive impairment was difficult to establish; however, for four mice, there was a “perfect correlation between Aβ load and performance … with one borderline animal and one outlier” (data not shown).
In a parallel study (also not shown), in collaboration with Paul Matthews at the Nathan Kline Institute in NY, similar mice were immunized and Aβ measured. The first changes to be detected were a decrease in plaque loads, which were later followed by decreases in Aβ as measured by ELISA. This poster left me convinced that immunization with Aβ in Tg mice can reduce cognitive impairment in an animal model. In particular, the control experiments conducted were extensive and convincing. One take-home message was that the selection of the behavioral test is critical in detecting cognitive impairment in the Tg mice. It remains to be seen whether the immune system in man can be primed sufficiently without going into overload, or whether there will be autoimmune problems. (However, passive immunization studies reported at this meeting suggest this can be avoided.)
Moving on to a doubly transgenic model (APP/PS1 mutant mice) from David Morgan and colleagues at the University of South Florida, there were a series of talks that covered the gamut, from the effects of calorierestriction (181.15) to blueberries (664.3), to when maximal titer is achieved (397.5) to behavioral effects of immunization (398.6). This is one well-studied group of mice. My notes indicate that David Morgan presented Abstract 397.5, not Gordon, but I admit it's all a blur now (and this has no correlation to number of Hurricanes consumed, n=1). My notes do indicate that following 3 to 4 vaccinations, the titer of Aβ IgG has reached maximum in these APP/PS1 mice (following a similar vaccination procedure to Schenk 1999). In the short-term study (three injections), there were no significant decreases in Aβ measures (ELISA and Aβ load), however, in the long term study (eight monthly injections), there was a 20% reduction in Aβ. As a percentage of control, the observed reductions are greater in singly mutant APP mice compared to APP/PS1 doubly mutant mice. However, in absolute terms, the reduction in area was greater in APP/PS1 mice compared to APP-only mice. David suggested that there was perhaps an upper limit in the amount of Aβ that could be removed. When asked if there was a relationship between microglia activation and Aβ load, David indicated no, but cautioned that they had only looked in short-term mice where there was less Aβ.
A separate talk on the behavioral performance in these same mice prior to and following vaccination was presented by Arendash (Abstract 397.6). As mentioned above, and at their poster (275.5), task selection is critical if one is to detect cognitive deficits in these mice. For example, Tg positive mice are not impaired in the Y-maze or the standard Morris water maze tests (circular platform performance). However, they are impaired in a radial arm water maze acquisition (RAWM). And there is also an age-related deficit in short-term (working) memory. And then Arendash talked about the effects of vaccination on RAWM task in these APP/PS1 doubly transgenic mice. Non-Tg mice exhibit no age effect on this task, while Tg mice are indistinguishable from controls until 15.5 months when their performance declines (and when Aβ levels are high). There is also a correlation between performance on the RAWM and Aβ load. In the current study being reported at the Neuroscience meeting, Tg and non-Tg mice were trained and tested between six and seven months on the RAWM and then immunized with Aβ peptide (between 7.5 and 11.5 months the animals received one vaccination/month). Retesting at 11.5 months (e.g., after four months of immunization) showed no differences between Tg-, Tg+ and Tg++Aβ mice. From 11.5 months to 15.5 months, additional vaccinations followed. At 15.5 months (following eight injections), the Tg++Aβ mice were tested again on the RAWM and while it still took them longer to reach the performance of controls, they were indistinguishable from Tg- control mice on the last two trials of block 3 of testing. This study is particularly nice because all animals started out the same in performance at 11.5 months and differences were only seen over time.
While no one reported on human studies at this year’s meeting, you can be sure trials are under way. Hey, does this mean that those of us weighing out and inhaling Aβ dust all these years can expect some measure of protection?—Brian Cummings
No Available References
No Available Further Reading