Multiple factors contribute to the peripheral induction of cerebral β-amyloidosis

Abstract

Deposition of aggregated amyloid-β (Aβ) peptide in brain is an early event and hallmark pathology of Alzheimer's disease and cerebral Aβ angiopathy. Experimental evidence supports the concept that Aβ multimers can act as seeds and structurally corrupt other Aβ peptides by a self-propagating mechanism. Here we compare the induction of cerebral β-amyloidosis by intraperitoneal applications of Aβ-containing brain extracts in three Aβ-precursor protein (APP) transgenic mouse lines that differ in levels of transgene expression in brain and periphery (APP23 mice, APP23 mice lacking murine APP, and R1.40 mice). Results revealed that beta-amyloidosis induction, which could be blocked with an anti-Aβ antibody, was dependent on the amount of inoculated brain extract and on the level of APP/Aβ expression in the brain but not in the periphery. The induced Aβ deposits in brain occurred in a characteristic pattern consistent with the entry of Aβ seeds at multiple brain locations. Intraperitoneally injected Aβ could be detected in blood monocytes and some peripheral tissues (liver, spleen) up to 30 d after the injection but escaped histological and biochemical detection thereafter. These results suggest that intraperitoneally inoculated Aβ seeds are transported from the periphery to the brain in which corruptive templating of host Aβ occurs at multiple sites, most efficiently in regions with high availability of soluble Aβ.

Keywords: Abeta; cerebral beta-amyloidosis; peripheral induction; seeding.

Figure 1.

Time course reveals earlier and stronger induction of cerebral β-amyloidosis in APP23 versus R1.40 Tg mice with the same peripherally applied Aβ-containing seeding extract. A, Experimental design: 1- to 2-month-old male and female APP23 and R1.40 Tg mice were intraperitoneally injected (i.p. inj.; two times with 100 μl, 1 week apart) with brain extracts derived from either aged non-Tg mice (WT extract) or aged Aβ-depositing APP Tg mice (Tg extract). Brains of inoculated APP23 mice were analyzed 1, 4, 6, 7, and 8 months and brains of inoculated R1.40 mice at 1, 6, 8, 10, and 12 months after inoculation [incubation (inc.)]. Male and female mice were combined for the analysis (see Material and Methods). B–E, Immunohistochemical analysis for Aβ deposition. B, In intraperitoneally inoculated APP23 mice, the first Aβ deposits were noted at 6 months and increased sharply with the longer incubation times of 7 or 8 months. In contrast, mice that received the WT extract and were incubated for 8 months showed no or very few Aβ deposits, which are typical for APP23 Tg mice at 8–10 months of age, which is the age at analysis. Stereological quantification of neocortical Aβ load is shown in C. In R1.40 Tg mice, the first Aβ deposits occurred 8 months after inoculation in all mice that received the Tg extract albeit to a minor extent and increases with longer incubation times of 10 or 12 months. In contrast, mice that received the WT extract showed no or very little Aβ deposits after a 12 month incubation (D) that is a typical finding for R1.40 Tg mice at 10–14 months of age, which is the age at analysis. Stereological quantification of neocortical Aβ load is shown in E. n = 5–9 for APP23 and R1.40 Tg mice per incubation time and extract; ***p < 0.001, ****p < 0.0001, one-way ANOVA, Bonferroni's post hoc test. Scale bar, 200 μm. F, Neocortical Aβ deposits in both lines were further differentiated into vascular and parenchymal deposits, and results revealed that, in APP23 Tg mice, the majority of deposits (measured by area occupied) are parenchymal deposits with a high variability for vascular deposits. Interestingly, in R1.40 mice, vascular deposition seemed prominent during early time points but decreased thereafter and parenchymal deposits became more dominant. G, Differentiation of the neocortical Aβ deposits into congophilic (Congo red positive) and non-congophilic deposits revealed that, in APP23 Tg mice, 5–15% of induced deposits are congophilic, whereas in R1.40 mice, almost exclusively diffuse deposits are induced at least up to 12 months incubation. (However, note that Congo red stains only the core of a plaque and thus the area fraction is low compared with the Aβ stain of a given amyloid plaque.) H–K, Biochemical analysis of brains 1 month after inoculation. Representative immunoblot for APP, differentiated into mature APP (APPm) and immature APP (APPim) as well as C99 levels, both recognized by the 6E10 antibody specific to human APP (H). Relative APP levels (APPm and APPim combined) were quantified from immunoblots and normalized to the housekeeping protein GAPDH. Mean APP value for WT extract-injected APP23 Tg mice was set to 100%. Quantification shows no difference between APP23 mice injected with WT or Tg extract. Similarly, no difference was found in R1.40 mice. However, APP levels in R1.40 Tg mice were lower than in APP23 Tg mice representing ∼30% of APP levels of APP23 mice (I). Overall, similar results were found for C99 levels (J). Human Aβ levels in brain homogenates were assessed with electrochemiluminescence-linked immunoassay, and again, similar to APP and C99 levels, human Aβ levels did not differ between mice of the same Tg line treated with either WT or Tg brain extract. Between the two lines, human Aβ levels in APP23 mice are significantly higher and are approximately threefold to fourfold in APP23 compared with R1.40 mice (K). n = 8 for APP23; n = 5 for R1.40. Note that the same animals (1 month incubation) were used for biochemical analysis and histology. *p < 0.05, ***p < 0.001, ****p < 0.0001, one-way ANOVA, Sidak's multiple comparison test. nctx, Neocortex.

Figure 2.

Induction of cerebral β-amyloidosis by intraperitoneal seeding in the absence of peripheral APP expression. A, Male and female APP23 mice and APP23 mice on an App-null background (APP23/App^−/−; all 1–2 months of age) were intraperitoneally inoculated with 200 μl of brain extract from aged Aβ-depositing APP23 mice (Tg extract) or non-Tg WT mice (WT extract) and analyzed for cerebral Aβ deposition 8 months later. Male and female mice were combined for the analysis (see Material and Methods). Shown are representative coronal sections (scale bar, 500 μm), and higher-magnification images of the frontal cortex are shown below (scale bar, 100 μm). Immunohistochemical analysis revealed a similar, albeit somewhat lower, induction of Aβ deposition by the Tg extract in APP23/App^−/− mice compared with the APP23 mice. Mice that received the WT extract showed no induced Aβ deposits. Note that 9- to 10-month-old APP23 and APP23/App^−/− mice (age at preparation) do exhibit the first endogenous Aβ deposition at this age. B, Stereological quantification indeed revealed that the amyloid load of APP23/App^−/− was 29% less compared with APP23 mice, but the difference did not achieve significance (n = 4 for APP23 mice and APP23/App^−/− mice injected with WT extract and n = 8 for APP23/App^−/− inoculated with Tg extract; mean ± SEM; one way ANOVA, F_(2,13) = 6.136, p = 0.013, Tukey's multiple comparison, *p < 0.05; n.s., not significant). C, Neocortical Aβ load was differentiated into vascular and parenchymal deposits. No difference was noted between APP23 and APP23/App^−/− mice. D, Differentiation of neocortical amyloid in congophilic (Congo red positive) and non-congophilic Aβ lesions also revealed a similar distribution of deposits in both APP23 and APP23/App^−/− Tg mice. nctx, Neocortex.

Figure 3.

Concentration-dependent induction of cerebral β-amyloidosis by intraperitoneal seeding with Aβ-containing brain extract in APP23 mice. A, Tg extract and PBS dilutions 1:10 and 1:100 thereof (200 μl) were intraperitoneally injected into young 1- to 2-month-old APP23 mice. Control mice were inoculated with 200 μl of brain extract from non-Tg WT mice (WT extract). The last group received a Tg extract that was mixed with the anti-Aβ-specific antibody β1 before the injection. All mice were males. Brains were immunohistochemically analyzed for Aβ deposition 8 months after intraperitoneal injection. Shown are representative coronal sections (scale bar, 500 μm), and higher-magnification images of the frontal cortex are shown below (scale bar, 100 μm). The undiluted Tg extract induced the expected Aβ deposition (see also Fig. 1B). The 1:10 diluted Tg extract induced appreciable amyloid deposition, whereas the 1:100 diluted Tg extract did no longer induce noticeable Aβ deposits (the same was true for the 1:1000 dilution; data not shown) and was identical to the WT extract-injected mice. Amyloid deposition in mice that received the extract/β1 mixture was markedly reduced. Note that some endogenous Aβ plaques are typical for 9- to 10-month-old APP23 mice (age at analysis). B, Stereological quantification of neocortical Aβ load, n = 3–4 mice per group; for the analysis, the 1:100 and 1:1000 dilutions were combined; mean ± SEM is indicated). C, Neocortical Aβ load was differentiated into vascular and parenchymal deposits. Although the majority of Aβ deposits induced by the undiluted Tg extract were detected in the brain parenchyma, deposition induced by the 1:10 diluted Tg extract showed prominent percentage of vascular deposits. D, Neocortical amyloid load was differentiated in congophilic (Congo red positive) and non-congophilic Aβ deposits revealing that the induced Aβ deposits are primarily of a diffuse nature. (However, note that Congo red stains only the core of a plaque and thus the area fraction is low compared with the Aβ stain of a given amyloid plaque.) The increased percentage of Congo red staining in the 1:100 dilution group represents the endogenous Aβ plaques. nctx, Neocortex; undil, undiluted.

Figure 4.

Regional distribution of intraperitoneally induced cerebral β-amyloidosis resembles endogenous amyloid deposition, albeit the induced amyloid deposits appear more clustered. A, Intraperitoneally injected Aβ seeds predominantly induce Aβ deposits in the frontal cortex, followed by the parietal cortex and the hippocampus. Shown is a sagittal section of an APP23 mouse 8 months after the intraperitoneal inoculation. Stereological quantification of systematically sampled coronal sections (every 24th was taken) throughout the brain revealed a similar anterior-to-posterior distribution of plaque load in intraperitoneally seeded APP Tg mice compared with nonseeded APP23 Tg mice. For the analysis, five intraperitoneally seeded mice (7–8 months after inoculation; plaque load, 0.90 ± 0.09) were compared with five aged APP23 mice with a similar plaque load (12–14 months old; plaque load, 0.84 ± 0.25). Male and female mice were combined for the analysis (see Material and Methods). In the diagram, the mean ± SEM is shown. B, Coronal sections through an APP23 Tg mouse brain intraperitoneally inoculated with Tg extract and incubated for 8 months. Note the clustering of the induced Aβ deposits (arrowheads). This is most evident in the posterior sections in which the induced amyloid load is less extensive compared with the anterior sections. Scale bar, 1000 μm. C, Higher magnification of clustered Aβ deposits, in which typically a large plaque was surrounded by numerous smaller Aβ deposits. Shown is a representative image of an APP23 Tg mouse 8 months after inoculation and an R1.40 mouse 12 months after inoculation. D–F, Quantitative analysis of clustering of Aβ deposits. Aβ deposits were quantified in size [blue for large (L), green for medium (M), yellow for small (S), and red for very small (XS)], and their x–y position was determined (for details, see Materials and Methods); an example is shown for the entorhinal cortex for intraperitoneally seeded and nonseeded APP23 Tg mice (D). The frequency distribution of the XS deposits (20–800 μm²) around larger deposits (>800 μm²), within 100 μm from their surface, revealed higher numbers of the XS deposits surrounding the larger Aβ deposits in intraperitoneally seeded APP23 Tg mice compared with nonseeded APP23 Tg mice (E). Differentiation between XS plaques surrounding L, M, and S plaques within 20, 50, and 100 μm from the surface yielded a significant increase in the number of XS neighboring plaques in intraperitoneally seeded APP23 Tg mice. However, the increase was most pronounced around the L and M plaques. Additionally, an increase of the number of neighbors around plaques of all sizes was evident in the intraperitoneally seeded animals as the distance to the plaque surface increased. The difference in the number of neighbors between nonseeded and seeded animals, for any plaque size, was significant (p < 0.001, Mann–Whitney test); darker-to-lighter box shading denotes 20, 50, and 100 μm from plaque surface, respectively, for each category (L, M, S); box hinges are 25th–75th percentiles, whiskers are adjacent values, and points are outliers. Scale bars: C, D, 200 μm.

Figure 5.

Intraperitoneally injected Aβ seeds are detectable shortly after the injection in macrophages in the peritoneal cavity, blood, liver, and spleen. A, Electrochemiluminescence-conjugated immunoassay for human Aβ (Aβ₄₀ and Aβ₄₂ combined) in the cellular fraction of the peritoneal fluid 3 h, 24 h, 1 week, or 1 month after intraperitoneal inoculation with Tg or WT extract. Male or female 4- to 8-month-old APP23 mice, n = 5 per group and time point except n = 4 for the 1 month time point. Separate statistical analysis did not indicate a gender or age difference. ANOVA revealed significant main effects and a significant extract × post-inoculation time interaction (F_(3,30) = 14.02; p < 0.001). Post hoc Tukey's tests showed significantly more Aβ in the Tg extract-treated mice at 24 h compared with 3 h (p < 0.05) and 1 week (p < 0.001). Wilcoxon's comparisons revealed similar significances. Indicated is the mean ± SEM. B, Pappenheim's stain (purple) combined with Aβ immunostaining (dark blue) of lavaged peritoneal cells from a Tg extract-inoculated mouse disclosed Aβ-positive monocytes (arrowheads). No Aβ immunoreactivity was found in WT extract-inoculated mice. C, Electrochemiluminescence-conjugated immunoassay for human Aβ in the blood cellular fraction 1–24 h, 1 week, or 1 month after intraperitoneal inoculation. Male or female 4- to 8-month-old APP23 mice, n = 16 (Tg extract) and n = 11 (WT extract) for the 1–24 h time point and n = 5 per group for other time points. (Animals were added to the 1–24 h time period in an unsuccessful effort to reduce the high variability of Aβ measurements.) Separate statistical analysis did not indicate a gender or age difference. ANOVA indicated no significant differences, but Wilcoxon's test revealed a significant difference between Tg extract-treated and WT extract-treated mice in the 1–24 h post-inoculation group (median, 0.495 vs 0; p < 0.05). D, Aβ-positive monocytes (arrowheads) in a blood film from a Tg extract-inoculated mouse (Pappenheim's stain, purple; Aβ immunostain, dark blue). E, Double-immunofluorescence staining revealed that Aβ-positive (red) cells are positive for CD11b and CD45 (green; maximum projection of 10 μm z-stack). Scale bars, 10 μm. The same cells were also positive for CD68 (data not shown). F, Immunoprecipitation of the blood cellular fraction with an antibody against human Aβ followed by immunoblot revealed a robust 4 kDa Aβ band in Tg extract-inoculated mice, as well as weaker oligomeric Aβ bands (indicated by asterisks) and APP, very similar as seen in the Tg extract (Langer et al., 2011). Only a weak signal corresponding to monomeric Aβ was seen in WT extract-inoculated mice, possibly because of residual human endogenous Aβ from the plasma fraction. IG heavy chain (HC) and light chain (LC) signals are attributable to coelution of the antibodies used for immunoprecipitation. G, The amyloid-specific dye pFTAA stains Aβ in peritoneal monocytes isolated from Tg-extract inoculated mice. DAPI is used as a nuclear counterstain (scale bar, 5 μm). H, I, Peripheral organs of representative mice measured and stained in (A–D) were immunostained for Aβ and counterstained with nuclear fast red. Aβ-positive cells with the morphological characteristics of macrophages were found in the liver of Tg extract-inoculated mice (n = 3 analyzed mice) but not WT inoculated mice (n = 2 analyzed mice) 1 d after inoculation. Scale bar: 50 and 10 μm. In a separate study, non-Tg B6 mice were injected with Tg and WT extract (n = 2 for Tg and WT each) and analyzed 1 h, 1 d, 1 week, and 1 month later. Again, Aβ-positive cells with a macrophagic appearance were found after 1 and 7 d but not 1 month after inoculation (results not shown).