Exercise alters the immune profile in Tg2576 Alzheimer mice toward a response coincident with improved cognitive performance and decreased amyloid

Abstract

Background: Inflammation is associated with Abeta pathology in Alzheimer's disease (AD) and transgenic AD models. Previously, it has been demonstrated that chronic stimulation of the immune response induces pro-inflammatory cytokines IL-1beta and TNF-alpha which contribute to neurodegeneration. However, recent evidence has shown that inducing the adaptive immune response reduces Abeta pathology and is neuroprotective. Low concentrations of IFN-gamma modulate the adaptive immune response by directing microglia to differentiate to antigen presenting cells. Our objective was to determine if exercise could induce a shift from the immune profile in aged (17-19 months) Tg2576 mice to a response that reduces Abeta pathology.

Methods: TG (n = 29) and WT (n = 27) mice were divided into sedentary (SED) and exercised (RUN) groups. RUN animals were provided an in-cage running wheel for 3 weeks. Tissue was harvested and hippocampus and cortex dissected out. Quantitative data was analyzed using 2 x 2 ANOVA and student's t-tests.

Results: IL-1beta and TNF-alpha were significantly greater in hippocampi from sedentary Tg2576 (TGSED) mice than in wildtype (WTSED) (p = 0.04, p = 0.006). Immune response proteins IFN-gamma and MIP-1alpha are lower in TGSED mice than in WTSED (p = 0.03, p = 0.07). Following three weeks of voluntary wheel running, IL-1beta and TNF-alpha decreased to levels indistinguishable from WT. Concurrently, IFN-gamma and MIP-1alpha increased in TGRUN. Increased CD40 and MHCII, markers of antigen presentation, were observed in TGRUN animals compared to TGSED, as well as CD11c staining in and around plaques and vasculature. Additional vascular reactivity observed in TGRUN is consistent with an alternative activation immune pathway, involving perivascular macrophages. Significant decreases in soluble Abeta40 (p = 0.01) and soluble fibrillar Abeta (p = 0.01) were observed in the exercised transgenic animals.

Conclusion: Exercise shifts the immune response from innate to an adaptive or alternative response. This shift in immune response coincides with a decrease in Abeta in advanced pathological states.

Figure 1

IL-1β is significantly greater in sedentary Tg2576 mice than in WT sedentary mice (p = 0.006). Exercise results in a significantly lower level of IL-1β in the Tg2576 (p = 0.01). The level of IL-1β in exercised Tg2576 mice (TG_RUN) is no longer distingushable from the WT mouse (WT_SED). TNF-α is significantly greater in sedentary Tg2576 mice (TG sed) than in WT sedentary mice (p = 0.04). Exercise reduces TNF-α in TG mice (TG run) to a level indistinguishable from the WT (WT_SED). *Significantly different from sedentary WT † significantly different from sedentary Tg2576.

Figure 2

CD11b positive microglia (green immunofluorescence) in TG_SED(A). Higher magnification reveals some co-labeling with microglial marker Iba-1 (red) (B, arrowheads). CD11b positive glia are present in TG_RUN (C) and co-labeled with Iba-1 (red) in some cases (D, arrowheads). Overall levels of Iba-1 (normalized to actin) are not significantly different based on condition or genotype (E). High immunoreactivity for Iba-1 in WT is likely due to the advanced age of the animals used.

Figure 3

IFN-γ is significantly lower in the Tg2576 sedentary mice that in the WT sedentary mice (p = 0.03). Exercise resulted in increased levels of IFN-γ in the Tg2576 mouse (TG_RUN) to a level indistinguishable form the WT (WT). MIP-1α demonstrated a trend of being lower in TG_SED compared to the WT (p = 0.07), but was significantly increased by exercise (TG_RUN) (p = 0.05). *Significantly different from sedentary WT; † significantly different from sedentary Tg2576.

Figure 4

MHC II levels were significantly greater in TG_RUN than TG_SED (p = 0.04). CD40 is significantly greater in TG_RUN compared to TG_SED (p = 0.008). WT_SED tended to have greater levels of CD40 than TG_SED, but this difference failed to achieve significance (p = 0.10). † Significantly different from sedentary Tg2576.

Figure 5

CD11c positive microglia (green immunofluorescence) are present in TG_SED and colocalize with Iba-1 (arrowheads) but do not appear vascular (A). CD11c labeling in TG_RUN appeared in cells not labeled by Iba-1 (red) that were linearly arranged, perhaps within or around microvessels. (B, D). Larger vessels had CD11c labeling along the vessel wall, perhaps in the perivascular space (C). Using macrophage markers CD68, we observed microvascular labeling again only in TG_RUN (D). Double labeling for CD11c (green) and CD68 (red) revealed that CD11c+ cells were adjacent to CD68+ cells in and around vasculature (arrows) (E, F). Using mannose receptor antibody (red), specific for perivascular macrophages, we again observed vascular labeling only in TG_RUN (G-I). High magnification shows the mannose receptor labeled cells are within vessels (H, I)(arrowheads). Green indicates Iba-1 labeling for microglia in and around vessels (G-I).

Figure 6

Aβ analysis by multiplex and ELISA. Aggregated Aβ levels are not significantly lower in hippocampus of TG_RUN compared to TG_SED, though a 35% decrease is observed in means. (A) Aβ₄₀ but not Aβ₄₂ is significantly lower in soluble fractions from cortex of TG_RUN and TG_SED (p = 0.01)(B). There are no significant differences in insoluble fractions (C).

Figure 7

Aβ analysis by dot blot. No differences existed between TG_SED and TG_RUN for total Aβ in the soluble fraction of hippocampal samples, evaluated by 6E10 antibody (A). Aβ fibrils, detected by OC antibody, were significantly decreased in TG_RUN animals compared to TG_SED (p = 0.01)(B). A representative dot blot is shown (C).