We monitored the appearance and growth of Aβ plaques in Tg2576 mice by in vivo two-photon microscopy of animals pretreated with methoxy-X04, which stains Aβ plaques. As a model, we used Tg2576 mice, which overexpress a human APP transgene containing the Swedish FAD mutation (K670N/M671L), and display the slowest amyloid pathology [9
] of all the models analyzed by Meyer-Luehmann et al. [16
]. We reasoned that if we were able to detect amyloid plaque growth in this model, then it should also be detectable in other mouse models with a faster amyloid pathology.
In summary, we showed that in the early stages of amyloid pathology in Tg2576 mice, newborn plaques are initially small in size and grow over time. These findings are in contrast to the previous study by Meyer-Luehmann et al. [16
] which found that plaques appeared very rapidly (within 24 h) and did not change significantly in size thereafter. One possible reason for the differing results might be the fact that we primarily quantified volume, rather than area. However, even when we quantified plaque size as an area, that area steadily increased over 6 weeks (Figs. d, d). Thus, quantification of the plaques in three dimensions does not account for the observed difference. Another possibility is that open-skull surgery induces local inflammation and gliosis in the first weeks after surgery [8
], which may affect plaque growth [24
]. Therefore, in contrast to Meyer-Luehmann et al. [16
], we waited for 3 weeks before imaging, so that possible inflammatory responses could abate. Yet even when we started daily imaging immediately after cranial window surgery, we were able to measure plaque growth. Initial plaque volumes were 92.1 ± 64.5 μm3
(95% CI, 60.0–124 μm3
= 18). However, a significant increase in plaque volume was only detected if we monitored plaque sizes for more than 16 days (Supplementary Fig. S2). Thus, delayed imaging 3 weeks after surgery does not account for the observed differences, either. We identified, however, two factors which determined if plaque growth can be observed in Tg2576 mice; first, and most importantly, we found that plaque growth can only be detected after a sufficiently long observation period. We monitored plaques for six consecutive weeks and detected that a statistically significant change in plaque volume of newborn plaques was achieved not earlier than 2 weeks after their first appearance (Fig. c), whereas a statistically significant change in plaque volume of pre-existing plaques was achieved not earlier than 4 weeks after the initial measurement (Fig. c). Second, we also found that the age of the observed animals plays an important role; we detected plaque growth in 12 months old mice, whereas 18 months old mice showed no significant overall growth of plaques (Figs. c, ), which was shown before by Christie et al. [3
]. Glial interaction with amyloid deposits has previously been suggested as a possible explanation for this observation in old mice [3
]. However, we would further like to propose the possibility that the high number of existing amyloid plaques at this age and therefore the large total surface area where newly formed Aβ fibrils can attach to existing plaques could lead to a decelerated growth of each individual plaque up to a degree that it is not detectable by two-photon in vivo imaging anymore. It is furthermore interesting to note that plaque size varied widely in 18 months old animals (Fig. ); yet we observed no newborn plaques at that age (Fig. ). Individual plaques, however, seemed both to shrink and grow between measurements (Fig. ). While this suggests the exciting possibility that plaques might be in a dynamic equilibrium, increasing measurement errors with increasing plaque size (note the heteroscedasticity in Figs. c, c) preclude us from making a firm conclusion. Thus, our data show that both the appropriate age and an appropriate observation period are required in order to detect plaque growth. Furthermore, our findings corroborate results which showed plaque growth in different double-transgenic AD mouse models [2
In conclusion, we show that amyloid plaques do grow in 12 months old Tg2576 mice when amyloid pathology is in the early stages [9
]. Newborn amyloid plaques are initially small in size and grow over the following weeks. After the initial growth period, plaques continue to grow, albeit more slowly, until plaque growth is not detectable anymore in aged mice. These findings have important implications for AD diagnosis and therapy. A recent study [12
], for instance, showed that in patients with mild cognitive impairment (MCI), amyloid plaque load, as measured by 11
C-Pittsburgh compound B (PIB) retention in positron emission tomography, increased over 5 years, while the regional cerebral metabolic rate of glucose decreased. However, cognitive function remained stable in these patients. Patients with AD, in contrast, had higher PIB retention than patients with MCI, which did not increase further over the observation period, and showed a decline in the regional cerebral metabolic rate of glucose as well as in cognitive function. In combination with our results, these data suggest that plaque formation occurs early in the course of the disease, but before a loss of cognitive function becomes apparent. Therefore, drugs that target plaque formation should be most effective early in the disease, when plaques are growing. These results also suggest that a different mechanism may be responsible for the cognitive decline in patients with manifest AD. Thus, an effort has to be made to find the mechanism of cognitive decline in the late disease stages in order to tailor effective therapeutic strategies.