Pro-inflammatory interleukin-6 signaling links cognitive impairments and peripheral metabolic alterations in Alzheimer's disease

Abstract

Alzheimer's disease (AD) is associated with memory impairment and altered peripheral metabolism. Mounting evidence indicates that abnormal signaling in a brain-periphery metabolic axis plays a role in AD pathophysiology. The activation of pro-inflammatory pathways in the brain, including the interleukin-6 (IL-6) pathway, comprises a potential point of convergence between memory dysfunction and metabolic alterations in AD that remains to be better explored. Using T2-weighted magnetic resonance imaging (MRI), we observed signs of probable inflammation in the hypothalamus and in the hippocampus of AD patients when compared to cognitively healthy control subjects. Pathological examination of post-mortem AD hypothalamus revealed the presence of hyperphosphorylated tau and tangle-like structures, as well as parenchymal and vascular amyloid deposits surrounded by astrocytes. T2 hyperintensities on MRI positively correlated with plasma IL-6, and both correlated inversely with cognitive performance and hypothalamic/hippocampal volumes in AD patients. Increased IL-6 and suppressor of cytokine signaling 3 (SOCS3) were observed in post-mortem AD brains. Moreover, activation of the IL-6 pathway was observed in the hypothalamus and hippocampus of AD mice. Neutralization of IL-6 and inhibition of the signal transducer and activator of transcription 3 (STAT3) signaling in the brains of AD mouse models alleviated memory impairment and peripheral glucose intolerance, and normalized plasma IL-6 levels. Collectively, these results point to IL-6 as a link between cognitive impairment and peripheral metabolic alterations in AD. Targeting pro-inflammatory IL-6 signaling may be a strategy to alleviate memory impairment and metabolic alterations in the disease.

Conflict of interest statement

The authors declare no competing interests.

Figures

Fig. 1. T2-weighted signal hyperintensities in the AD hypothalamus and hippocampus correlate inversely with MMSE scores and gray matter densities.

a Representative transversal (left) and coronal (right) T2-weighted images from control (CTL, top) and AD (bottom) patients. Arrows indicate the placement of right and left ROIs in the hypothalamus (yellow arrows) and hippocampus (white arrows). b Bars represent means ± SEM of T2-intensity signal ratios (hypothalamus relative to putamen) in control subjects (n = 18) and AD patients (n = 16); symbols represent individual subjects; p-value was calculated from student’s t-test and is shown in the graph. c Correlation between hypothalamus/putamen T2-intensity signal ratios and MMSE scores. Pearson correlation coefficient (r) and p-value are shown in the graph (n = 34). d Bars represent means ± SEM of T2-intensity signal ratios (hippocampus relative to putamen) in control subjects (n = 18) and AD patients (n = 16); symbols represent individual patients; p-value (Mann-Whitney test) is shown in the graph. e Correlation between hippocampus/putamen T2-intensity signal ratios and MMSE scores. Spearman correlation coefficient (r) and p-value are shown in the graph (n = 34). Bars represent means ± SEM of hypothalamic (f) and hippocampal (g) gray matter densities measured by voxel-based morphometry in control subjects (n = 17) and AD patients (n = 13); symbols represent individual patients; p-value (Student’s t-test) is shown in graph. h Correlation between hypothalamus gray matter density and hypothalamus/putamen T2-intensity signal ratios. Pearson correlation coefficient (r) and p-value are shown in the graph (n = 30). i Correlation between hippocampus gray matter density and hippocampus/putamen T2-intensity signal ratios. Spearman correlation coefficient (r) and p-value are shown in the graph (n = 30).

Fig. 2. Intraneuronal hyperphosphorylated tau aggregates, Aβ deposits and inflammatory markers in post-mortem AD hypothalamus.

a Representative photomicrographs showing immunostaining of Aβ deposits (6E10 antibody; red) in the lateral hypothalamus of AD cases (AD1: Female 77 years old; AD2: Female, 88 years old; AD3: Female, 82 years old) and age-matched control (CTL1: Female, 71 years old; CTL2: Female, 80 years old). Graph represents means ± SEM of number of plaques in AD patients or control individuals; p-value was calculated from student’s t-test, and is shown in the graph. b Representative photomicrographs of the lateral hypothalamus showing double immunofluorescence staining for Aβ (Abcam ab134022; red) and astrocytes (GFAP; green) in post-mortem AD (AD1: Female 77 years old; AD2: Female, 88 years old) and age-matched control (CTL: Female, 71 years old). Magnified images of specific fields are indicated by the dashed rectangles in the main panels. c Representative photomicrographs of the lateral hypothalamus showing positive intraneuronal immunostaining for AT8 (p-tau antibody; red) in neurons in AD cases (AD1: Female, 82 years old; AD2: Female 77 years old) and absence of AT8-positive cells in age-matched control (CTL1: Female, 71 years old; CTL2: Female, 80 years old). Graph represents means + SEM of number of plaques in AD patients or control individuals; p-value (Student’s t-test) is shown in the graph. d, e Representative micrographs of triple immunofluorescence staining for Aβ (Abcam ab134022; red), astrocytes (GFAP; green), and IL-6 (blue) in AD lateral hypothalamus (Female, 77 years old). White arrows indicate a cell with triple staining for GFAP, Aβ and IL-6, and yellow arrows show a GFAP-positive cell that does not colocalize with either Aβ or IL-6 immunoreactivities.

Fig. 3. IL-6 is upregulated in AD brain and plasma, and correlates inversely with MMSE.

Representative photomicrographs of triple immunofluorescence staining for Aβ (Abcam ab134022; red), astrocytes (GFAP; green), and IL-6 (blue) in post-mortem cingulate cortex from AD subjects (AD1: Female, 82 years old; AD2: Female, 88 years old) (a, b) and an age-matched control (CTL: Female, 80 years old) (c). d Means ± SEM of IL-6 protein levels measured by ELISA in post-mortem prefrontal cortex from AD (n = 8) or control subjects (n = 5); symbols represent different individuals; p-value was calculated from student’s t-test and is shown in the graph. e Representative immunoblot of SOCS3 (β-actin used as loading control) in post-mortem prefrontal cortex (PFC) from AD (n = 8) or control (n = 9) subjects. Graph shows means ± SEM of SOCS3/β-actin ratios; symbols represent different individuals; p-value (Student’s t-test) is shown in the graph. f Means ± SEM of plasma IL-6 in AD patients (n = 12) or control individuals (n = 14) from the same cohort that was studied by MRI in Fig. 1; symbols represent different individuals; p-value was calculated from unpaired t-test with Welch’s correction, and is shown in the graph. g Correlation between plasma IL-6 and hippocampus/putamen T2-intensity signal ratios for patients in the studied cohort (n = 26). Spearman correlation coefficient (r) and p-value are shown in the graph. h, i Correlations between plasma IL-6 and MMSE or FAB scores for patients in the studied cohort. Symbols represent different individuals (n = 26). Pearson correlation coefficients (r) and p-values are shown in the graphs.

Fig. 4. Brain IL-6 signaling is upregulated and mediates memory impairment in AD mouse models.

a Means ± SEM of IL-6 protein levels in the hippocampus of 11-month-old APP/PS1 (n = 7) and WT littermates (n = 5); symbols represent individual animals; p-value (Student’s t-test) is shown in graph. b Representative immunoblot of pSTAT3, STAT3, and β-actin (used as a loading control) in the hippocampus of 11-month-old APP/PS1 (n = 6) and WT mice (n = 8). Graph shows pSTAT3/STAT3 ratio; symbols represent individual animals; p-value (Student’s t-test) is shown in graph. c, d Means ± SEM of IL-6 (n = 4 Veh; 6 AβOs) and SOCS3 expression (n = 7 Veh; 8 AβOs) (mRNA levels, expressed as fold-change relative to vehicle-infused mice) in mouse hippocampus 24 h after i.c.v. infusion of AβOs. β-actin was used as a housekeeping control. Symbols represent individual animals; p-values were calculated from student’s t-test and unpaired t-test with Welch’s correction, and are shown in graphs. e Representative immunoblot of pSTAT3, STAT3, and β-actin (loading control) in mouse hippocampus 24 h after i.c.v. infusion of AβOs (n = 8) (or vehicle, n = 7). Graph shows pSTAT3/STAT3 ratio; symbols represent individual animals; p-value (Student’s t-test) is shown in graph. f Representative immunofluorescence staining for pSTAT3 (red), MAP2 (green), and DAPI (blue) in a primary hippocampal culture exposed to AβOs (or vehicle) for 24 h. Yellow arrowhead indicates pSTAT3 staining (red) in the nucleus of a cell that is not positive for MAP2 (green). g Novel Object Recognition (NOR) task with 11-month-old male APP/PS1 mice and WT littermates (n = 7–12) after 3 i.c.v. injections of anti-IL-6 (αIL6) or anti-GFP (αGFP), as indicated. Symbols represent individual mice, n = 7–12 per experimental condition; *p < 0.05, **p < 0.01, ***p < 0.001, one-sample Student’s t-test. h NOR task with 3-month-old mice i.c.v-infused with AβOs (or vehicle) and AG490. Bars represent means ± SEM. Symbols represent individual mice, n = 7–10 per experimental condition; *p < 0.05, **p < 0.01, one-sample Student’s t-test. Data are representative of at least two independent experiments.

Fig. 5. Central IL-6 promotes peripheral metabolic dysfunction in AD mouse models.

a Expression of IL-6 in the mouse hypothalamus 4 h after i.c.v. infusion of AβOs (n = 5) (or vehicle, n = 4). β-actin was used as housekeeping control. Bars represent means ± SEM; symbols correspond to individual mice; p-value (Student’s t-test) is shown in graph. b IL-6 in the mouse hypothalamus 4 h after AβOs (n = 5) or vehicle (n = 5) were i.c.v.-infused. Bars represent means ± SEM; symbols correspond to individual mice; p-value (Student’s t-test) is shown in graph. c Representative in situ hybridization images for SOCS3 in the mouse hypothalamus 7 days after AβOs (n = 4) or vehicle (n = 4) were i.c.v.-infused. Graph shows integrated densities (means ± SEM) in the regions indicated by dashed ellipses in AβO- or vehicle-infused mice. Symbols correspond to individual mice; p-value (Student’s t-test) is shown in the graph. d IL-6 in hypothalamic homogenates from 11-month-old APP/PS1 mice (n = 6) or WT litttermates (n = 8). Bars correspond to means ± SEM. Symbols correspond to individual mice; p-value (Student’s t-test) is shown in the graph. e Representative immunoblot for SOCS3 in the hypothalamus of 11-month-old male APP/PS1 mice (n = 4) or WT littermates (n = 4). Bars in graph correspond to means ± SEM. Symbols correspond to individual mice; p-value (Mann–Whitney test) is shown in graph. f Glucose tolerance test in WT mice 36 h after i.c.v. infusion of AβOs (or vehicle) and treatment with AG490. Symbols correspond to individual mice and represent means ± SEM, n = 4–6 per experimental group; *p < 0.05 from two-way ANOVA followed by Tukey’s post-hoc test comparing Veh vs. AβOs and AβOs vs. AβOs + AG490). g Glucose tolerance test in 11-month-old APP/PS1 mice after 4 i.c.v. injections of anti-IL-6 (αIL6) or anti-GFP (αGFP). Mice received 1 g/kg of glucose (i.p.) and blood glucose was measured at the indicated timepoints. Symbols correspond to individual mice and represent means ± SEM, n = 4-6 per experimental group; ***p < 0.001, ****p < 0.0001 from two-way ANOVA followed by Tukey’s post-hoc test comparing WT + αGFP vs. APP/PS1 + αGFP and APP/PS1 + αGFP vs. APP/PS1 + αIL6). h Plasma IL-6 in 11-month-old APP/PS1 mice after 4 i.c.v. injections of anti-IL-6 (αIL6) or anti-GFP (αGFP). Bars represent means ± SEM. Symbols correspond to individual mice, n = 8–12 per experimental condition; p-values were calculated from two-way ANOVA followed by Tukey’s post-hoc test and are shown in the graph. Data are representative of at least two independent experiments.