Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2019 Feb;11(2):e9665.
doi: 10.15252/emmm.201809665.

Inhibition of Stat3-mediated astrogliosis ameliorates pathology in an Alzheimer's disease model

Nicole Reichenbach  1 , Andrea Delekate  1 , Monika Plescher  1 , Franziska Schmitt  1 , Sybille Krauss  1 , Nelli Blank  1 , Annett Halle  1   2 , Gabor C Petzold  3   4
Affiliations

Inhibition of Stat3-mediated astrogliosis ameliorates pathology in an Alzheimer's disease model

Nicole Reichenbach et al. EMBO Mol Med. 2019 Feb.

Abstract

Reactive astrogliosis is a hallmark of Alzheimer's disease (AD), but its role for disease initiation and progression has remained incompletely understood. We here show that the transcription factor Stat3 (signal transducer and activator of transcription 3), a canonical inducer of astrogliosis, is activated in an AD mouse model and human AD Therefore, using a conditional knockout approach, we deleted Stat3 specifically in astrocytes in the APP/PS1 model of AD We found that Stat3-deficient APP/PS1 mice show decreased β-amyloid levels and plaque burden. Plaque-close microglia displayed a more complex morphology, internalized more β-amyloid, and upregulated amyloid clearance pathways in Stat3-deficient mice. Moreover, astrocyte-specific Stat3-deficient APP/PS1 mice showed decreased pro-inflammatory cytokine activation and lower dystrophic neurite burden, and were largely protected from cerebral network imbalance. Finally, Stat3 deletion in astrocytes also strongly ameliorated spatial learning and memory decline in APP/PS1 mice. Importantly, these protective effects on network dysfunction and cognition were recapitulated in APP/PS1 mice systemically treated with a preclinical Stat3 inhibitor drug. In summary, our data implicate Stat3-mediated astrogliosis as an important therapeutic target in AD.

Keywords: Alzheimer's disease; Stat3; astrocytes; astrogliosis; glia.

PubMed Disclaimer

Figures

Figure 1
Figure 1. Stat3 activation in an AD model is efficiently reduced using a conditional knockout strategy
  1. A, B

    To verify the specificity of Cx43‐CreERT mice for astrocytes, mice were crossed to a tdTomato reporter line to induce Cre‐dependent tdTomato expression after tamoxifen administration. TdTomato fluorescence was enhanced with an antibody against red fluorescent protein, and astrocytes were stained with antibodies against GFAP or S100β (not shown). Merged images and quantification show efficient and widespread tdTomato expression specific for astrocytes (arrowheads indicate tdTomato and GFAP overlap) in 8‐ and 11‐month‐old animals. Expression was low in neurons (stained with NeuN, not shown) and negligible in NG2 cells and oligodendrocytes. Scale bars, 250 μm (upper panel) and 50 μm (lower panel).

  2. C

    The majority of peri‐plaque astrocytes in APP/PS1‐Stat3WT were positive for activated/phosphorylated Stat3 (pStat3) in the cortex (Cx) and hippocampus (HC), whereas the fraction of pStat3‐positive astrocytes was significantly lower in APP/PS1‐Stat3KO mice in both regions (n = 8 mice (three females and five males) per group; age, 8 months; *P < 0.05, Mann–Whitney test). The occurrence of pStat3 in WT‐Stat3KO and WT‐Stat3WT mice was negligible (n = 8 mice (four females and four males) per group; age, 8 months; Mann–Whitney test).

  3. D, E

    Examples of Stat3 activation in APP/PS1‐Stat3KO and APP/PS1‐Stat3WT. (D) pStat3 (arrowheads) was abundantly present in reactive astrocytes (identified by GFAP) around plaques (identified by IC16). (E) Only few astrocytes were positive for pStat3 in APP/PS1‐Stat3WT mice. Scale bars, 50 μm.

  4. F, G

    In postmortem brain sections from AD patients, we detected nuclear pStat3 (arrows) in the majority of peri‐plaque reactive astrocytes (identified by GFAP) in the cortex and hippocampus (plaques were stained with methoxy‐XO4, arrowheads). Scale bar, 50 μm. In total, n = 4 cortical and n = 3 hippocampal sections were analyzed.

Data information: Data are represented as mean ± SEM.
Figure 2
Figure 2. Astrocyte‐specific Stat3 deletion induces greater astrocytic and microglial complexity
  1. A, B

    Overall astrocytic and microglial coverage, as assessed by GFAP and Iba1 stainings, remained unchanged in APP/PS1‐Stat3WT versus APP/PS1‐Stat3KO mice (n = 8 mice (four females and four males) per group; age, 8 months; Mann–Whitney test for each comparison). Scale bars, 500 μm.

  2. C, D

    The volume of peri‐plaque reactive astrocytes was increased by the Stat3 deletion (*P < 0.05, Mann–Whitney test), and there was a nonsignificant trend toward more astrocytic branches and junctions and longer total process length (Mann–Whitney test; APP/PS1‐Stat3WT, n = 5 (two females and three males) mice; APP/PS1‐Stat3KO, n = 7 (three females and four males) mice; age, 8–10 months). Scale bar, 20 μm.

  3. E, F

    Microglial volume was not affected by the Stat3 deletion (Mann–Whitney test), but peri‐plaque microglia had significantly more microglial branches and junctions per plaque, and the total process length of peri‐plaque microglia was increased (*P < 0.05, Mann–Whitney test; same mice as in C and D). Scale bar, 20 μm.

Data information: Data are represented as mean ± SEM.
Figure EV1
Figure EV1. No evidence for proliferation of near‐plaque reactive glia
  1. A, B

    The total density of astrocytes and microglia remained unchanged in APP/PS1‐Stat3KO compared to APP/PS1‐Stat3WT mice (Mann–Whitney test for both comparisons; APP/PS1‐Stat3WT, n = 8 (three females and five males) mice; APP/PS1‐Stat3KO, n = 8 (five females and three males) mice; age, 8–9 months).

  2. C, D

    Using an anti‐Ki67 antibody as a marker for cellular proliferation, we detected Ki67‐positive cells (arrows) in the hippocampal dentate gyrus as a positive control. However, no Ki67 signal was detected around plaques (marked by arrowheads) of either APP/PS1‐Stat3KO or APP/PS1‐Stat3WT mice, indicating little‐to‐no glial proliferation (scale bars, 50 μm; same mice as in A and B).

Data information: Data are represented as mean ± SEM.
Figure EV2
Figure EV2. No significant morphological changes in glial cells remote from plaques
  1. A–D

    The morphology of hippocampal astrocytes remote from plaques was largely unaltered in APP/PS1‐Stat3KO compared to APP/PS1‐Stat3WT mice.

  2. E–H

    Similarly, the morphology of hippocampal microglia remote from plaques was also similar in APP/PS1‐Stat3KO compared to APP/PS1‐Stat3WT mice.

Data information: Data are represented as mean ± SEM. Mann–Whitney test for all comparisons; scale bars, 20 μm; APP/PS1‐Stat3KO, n = 10 (four females and six males) mice; APP/PS1‐Stat3WT, n = 6 (three females and three males) mice; age, 8–9 months.
Figure 3
Figure 3. Astrocyte‐specific Stat3 deletion reduces Aβ plaque burden and Aβ levels without altering APP metabolism
  1. A, B

    Aβ plaque burden, as assessed by plaque load and size using an anti‐Aβ antibody, was strongly reduced in APP/PS1‐Stat3KO versus APP/PS1‐Stat3WT mice (*P < 0.05, Mann–Whitney test; scale bar, 300 μm).

  2. C

    Plaque density, assessed by staining brain sections with thioflavin (yellow), remained unchanged in APP/PS1‐Stat3KO versus APP/PS1‐Stat3WT mice (Mann–Whitney test; scale bar, 250 μm).

  3. D, E

    Electrochemiluminescence ELISA after sequential extraction from whole‐brain homogenates using RIPA and SDS buffer revealed that soluble Aβ1–42 and Aβ1–40 were significantly reduced (*P < 0.05, Mann–Whitney test), whereas there was a nonsignificant trend toward reduced levels of insoluble Aβ1‐42 and Aβ1‐40 (Mann–Whitney test).

  4. F–H

    Western blot quantification of full‐length amyloid precursor protein (APP) and its C‐terminal fragments (CTF) showed no differences between both groups (Mann–Whitney test).

Data information: Data are represented as mean ± SEM. For all datasets, n = 9 male mice for both groups (age, 11 months).Source data are available online for this figure.
Figure 4
Figure 4. Astrocyte‐specific Stat3 deletion increases microglial Aβ internalization and degradation, and reduces apoE expression, dystrophic neurites, and detrimental cytokines
  1. A

    Internalization of Aβ (stained with IC16 antibody or methoxy‐XO4) was assessed using an engulfment assay, in which glial and Aβ structures were surface‐rendered and Aβ volumes co‐localized with glial volumes were quantified. Scale bars, 10 μm.

  2. B, C

    Microglia (left Y axes) from APP/PS1 mice internalized significantly more Aβ positive for IC16 or methoxy‐XO4 when Stat3 was deleted in astrocytes (*P < 0.05, Mann–Whitney test), whereas no changes were seen in astrocytes (right axes; APP/PS1‐Stat3WT, n = 8 (four females and four males) mice; APP/PS1‐Stat3KO, n = 11 (five females and six males) mice; age, 11 months; Mann–Whitney test).

  3. D–H

    (D–F) Western blot quantification of protein levels of the Aβ‐degrading enzymes neprilysin/CD10 and CD36, as well as the Aβ‐binding apolipoprotein E (apoE), revealed a significantly increased expression of neprilysin and CD36 and a decreased expression of apoE (APP/PS1‐Stat3WT, n = 9 (five females and four males) mice; APP/PS1‐Stat3KO, n = 9 (five females and four males) mice; age, 11 months; *P < 0.05, Mann–Whitney test for all comparisons). (G) In contrast, TREM2 expression remained unchanged (APP/PS1‐Stat3WT, n = 8 (four females and four males) mice; APP/PS1‐Stat3KO, n = 7 (four females and three males) mice; age, 11 months; Mann–Whitney test). (H) Western blots for proteins analyzed in (D‐G).

Data information: Data are represented as mean ± SEM.Source data are available online for this figure.
Figure 5
Figure 5. Stat3 deletion triggers a phenotypical switch in reactive astrocytes
  1. A, B

    Quantitative PCR from cortex of APP/PS1‐Stat3KO compared to APP/PS1‐Stat3WT mice revealed lower expression of “A1” markers Amigo2 and C3, whereas Ggta1 remained unchanged. In turn, the “A2” marker Tm4sf1 was upregulated and there was a nonsignificant trend for a higher expression of B3gnt5 (n = 6 mice (three females and three males) per group; age, 8 months; *P < 0.05, Mann–Whitney test).

  2. C–F

    Confirming lower expression of the “A1” marker C3d, Western blot analysis indicated lower protein levels of C3d in APP/PS1‐Stat3KO. Immunohistochemistry using an antibody against C3d revealed that lower expression of C3d particularly occurred in peri‐plaque reactive astrocytes (arrows indicate C3d and GFAP colocalization; arrowheads indicate plaques visualized with methoxy‐XO4; scale bars, 50 μm; APP/PS1‐Stat3WT, n = 6 (two females and four males) mice; APP/PS1‐Stat3KO, n = 6 (three females and three males) mice; age, 8 months; *P < 0.05, Mann–Whitney test).

  3. G

    Whole‐brain levels of the pro‐inflammatory cytokines IL‐1β and TNF‐α were significantly reduced in APP/PS1‐Stat3KO mice (Mann–Whitney test), whereas no changes were seen for IL‐10 (APP/PS1‐Stat3WT, n = 13 (six females and seven males) mice; APP/PS1‐Stat3KO, n = 13 (eight females and five males) mice; age, 11 months; *P < 0.05, Mann–Whitney test).

  4. H

    These changes were paralleled by a decrease in the area covered by dystrophic neurites in APP/PS1‐Stat3KO compared to APP/PS1‐Stat3WT mice, as assessed by LAMP1 staining (*P < 0.05, Mann–Whitney test; n = 10 male mice for both groups; age, 8 months; scale bars, 300 μm).

Data information: Data are represented as mean ± SEM.
Figure 6
Figure 6. Astroglial Stat3 modulates network imbalance in APP/PS1 mice
  1. A

    For in vivo two‐photon imaging, astrocytes (arrows) and neurons (arrowheads) were labeled with the calcium indicator OGB‐1, and astrocytes were co‐labeled with sulforhodamine 101 (SR101; arrows). Aβ plaques were labeled with the intravital dye methoxy‐XO4 (open arrowheads). Scale bar, 50 μm.

  2. B

    Calcium imaging of anesthetized animals showed that the hyperactivity of astrocytes in APP/PS1‐Stat3KO mice was reduced to levels comparable to WT‐Stat3WT mice, but significantly increased in APP/PS1‐Stat3WT mice (*P < 0.05, one‐way ANOVA followed by Bonferroni's multiple comparison test; n = 6 mice (three females and three males) for each group; age, 8 months).

  3. C, D

    Similarly, neuronal activity was also reduced to levels comparable to WT‐Stat3WT mice in APP/PS1‐Stat3KO mice, but significantly increased in APP/PS1‐Stat3WT mice (*P < 0.05, one‐way ANOVA followed by Bonferroni's multiple comparison test; same mice as in B). The cumulative distributions of neuronal calcium transients in APP/PS1‐Stat3KO mice were not different from those of WT‐Stat3WT mice (P = 0.31, Kolmogorov–Smirnov test), but significantly different from those of APP/PS1‐Stat3WT mice (P = 0.001, Kolmogorov–Smirnov test).

  4. E

    To investigate a reciprocal scenario in which astroglial hyperactivity reduces Stat3 activation, we implanted osmotic minipumps into APP/PS1 mice and treated them with the network‐normalizing P2Y1R inhibitor MRS2179 or vehicle for 6 weeks.

  5. F–I

    Two‐photon imaging of these mice confirmed that astrocytic hyperactivity and propagating calcium waves in APP/PS1 mice were reduced by MRS2179 (n = 6 (four females and two males) mice) compared to vehicle‐treated APP/PS1 mice (n = 6 (two females and four males) mice; age, 8 months; *P < 0.05, Mann–Whitney test). This network normalization induced a reduction in activated Stat3 (pStat3) in astrocytes assessed in fixed brain sections of the same mice (*P < 0.05, Mann–Whitney test). Scale bars, 50 μm. Arrowheads indicate pStat3 and GFAP colocalization, arrows indicate plaques visualized with methoxy‐XO4 (M‐XO4).

Data information: Data are represented as mean ± SEM.
Figure 7
Figure 7. Stat3 deletion in astrocytes protects from spatial memory and learning deterioration
  1. A

    Spatial learning and memory were assessed in the Morris Water Maze paradigm. APP/PS1‐Stat3KO mice showed faster latencies to reach the hidden platform compared with APP/PS1‐Stat3WT on days 4 and 5, but were similar to WT‐Stat3WT and WT‐Stat3KO mice (*P < 0.05, two‐way repeated‐measures ANOVA followed by Bonferroni post hoc test; P‐values are for APP/PS1‐Stat3KO versus APP/PS1‐Stat3WT mice).

  2. B

    The area under the curve (AUC) for the latency to reach the hidden platform was similar in APP/PS1‐Stat3WT compared to WT‐Stat3WT and WT‐Stat3KO mice, but significantly higher in APP/PS1‐Stat3KO mice (*P < 0.05, Kruskal–Wallis test followed by Dunn's multiple comparisons test).

  3. C

    The swimming velocity was similar in all groups (Kruskal–Wallis test followed by Dunn's multiple comparisons test).

  4. D, E

    In the probe trial 24 h after the last training day, the time mice spent in the target quadrant (TQ) was different from chance in all groups except for APP/PS1‐Stat3WT mice (P < 0.05, one‐tailed one‐sample t‐test). APP/PS1‐Stat3KO, WT‐Stat3WT, and WT‐Stat3KO mice spent significantly more time in the target quadrant compared to the mean of all other quadrants (AO), whereas APP/PS1‐Stat3WT spent equal times in the target and all other quadrants (*P < 0.05, Wilcoxon matched‐pairs signed rank test for each comparison).

Data information: Data are represented as mean ± SEM. WT‐Stat3WT, n = 15 (7 females and 8 males) mice; WT‐Stat3KO, n = 15 (10 females and 5 males) mice; APP/PS1‐Stat3WT, n = 15 (eight females and seven males) mice; APP/PS1‐Stat3KO, n = 15 (nine females and six males) mice; age, 8–9 months).
Figure EV3
Figure EV3. Persistence of cognitive protection and reduced amyloid pathology in late‐stage APP/PS1‐Stat3KO mice
  1. A

    Spatial learning and memory were assessed in the Morris Water Maze. APP/PS1‐Stat3KO mice showed faster latencies to reach the hidden platform compared with APP/PS1‐Stat3WT on day 5, but were similar to WT‐Stat3WT and WT‐Stat3KO mice (*P < 0.05, two‐way repeated‐measures ANOVA followed by Bonferroni post hoc test; P‐value for APP/PS1‐Stat3KO versus APP/PS1‐Stat3WT mice).

  2. B, C

    The area under the curve (AUC) for the latency to reach the hidden platform was similar in APP/PS1‐Stat3WT compared to WT‐Stat3WT and WT‐Stat3KO mice, but significantly higher in APP/PS1‐Stat3KO mice (*P < 0.05, Kruskal–Wallis test followed by Dunn's multiple comparisons test). The swimming velocity was similar in all groups (Kruskal–Wallis test followed by Dunn's multiple comparisons test).

  3. D

    In the probe trial, APP/PS1‐Stat3KO, WT‐Stat3WT, and WT‐Stat3KO mice spent significantly more time in the target quadrant (TQ) compared to the mean of all other quadrants (AO), whereas APP/PS1‐Stat3WT spent equal times in the target and all other quadrants (*P < 0.05, Wilcoxon matched‐pairs signed rank test for each comparison).

  4. E, F

    Plaque load and plaque size were reduced in APP/PS1‐Stat3KO compared to APP/PS1‐Stat3WT mice (*P < 0.05, Mann–Whitney test for both comparisons; same mice as in A–D).

Data information: Data are represented as mean ± SEM. WT‐Stat3WT, n = 12 (6 females and 6 males) mice; WT‐Stat3KO, n = 10 (4 females and 6 males) mice; APP/PS1‐Stat3WT, n = 12 (seven females and five males) mice; APP/PS1‐Stat3KO, n = 12 (eight females and four males) mice; age, 13–14 months.
Figure 8
Figure 8. The systemic Stat3 inhibitor SH‐4‐54 confers protection from cognitive decline in APP/PS1 mice
  1. A

    APP/PS1 mice were systemically treated with SH‐4‐54 for 6 weeks and compared to age‐matched APP/PS1 mice treated with vehicle. In the Morris Water Maze test, treatment with SH‐4‐54 led to significantly faster latencies to reach the hidden platform on the last training day (*P < 0.05, two‐way repeated‐measures ANOVA followed by Bonferroni post hoc test).

  2. B

    The area under the curve (AUC) for the latency to reach the hidden platform was also smaller in APP/PS1 mice treated with the Stat3 inhibitor (*P < 0.05, Mann–Whitney test).

  3. C

    The swimming velocity was not affected by the treatment (Mann–Whitney test).

  4. D

    In the probe trial test, APP/PS1 mice treated with the Stat3 inhibitor spent significantly more time in the target quadrant (TQ) compared to the mean of all other quadrants (AO), whereas APP/PS1 mice treated with vehicle spent equal times in the target and all other quadrants (*P < 0.05, Wilcoxon matched‐pairs signed rank test for all comparisons).

  5. E

    To verify target engagement in the brain, the same mice assessed in the Morris Water Maze were anesthetized and imaged using in vivo two‐photon microscopy of calcium activity. Systemic treatment with the Stat3 inhibitor reduced the hyperactive phenotype of cortical neurons (*P < 0.05, Mann–Whitney test).

Data information: Data are represented as mean ± SEM. APP/PS1‐Stat3WT, n = 12 (five females and seven males) mice; APP/PS1‐Stat3KO, n = 12 (four females and eight males) mice; age, 8 months).
Figure 9
Figure 9. Target engagement of the systemic Stat3 inhibitor in APP/PS1 mice
  1. A–D

    SH‐4‐54 significantly decreased plaque growth, as assessed by IC16 immunohistochemistry, while plaque load remained unchanged. There was also no significant change in dystrophic neurite area during the treatment time (Mann–Whitney test for all comparisons; scale bars, 500 μm).

  2. E–G

    The fraction of pStat3‐positive astrocytes in the peri‐plaque region was strongly reduced by the treatment with the Stat3 inhibitor (arrowheads indicate pStat3 signals; scale bars, 100 μm; Mann–Whitney test).

  3. H–K

    While no changes were seen in morphological parameters of peri‐plaque astrocytes, there was a significant increase in the process length of near‐plaque microglia, indicating higher microglial complexity (Mann–Whitney test for all comparisons).

Data information: Data are represented as mean ± SEM. APP/PS1 mice treated with SH‐4‐54, n = 12 (six females and six males) mice; APP/PS1 mice treated with vehicle, n = 12 (7 females and 5 males) mice; age, 8 months.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. Alonzi T, Maritano D, Gorgoni B, Rizzuto G, Libert C, Poli V (2001) Essential role of STAT3 in the control of the acute‐phase response as revealed by inducible gene activation in the liver. Mol Cell Biol 21: 1621–1632 - PMC - PubMed
    1. Barres BA (2008) The mystery and magic of glia: a perspective on their roles in health and disease. Neuron 60: 430–440 - PubMed
    1. Ben Haim L, Ceyzeriat K, Carrillo‐de Sauvage MA, Aubry F, Auregan G, Guillermier M, Ruiz M, Petit F, Houitte D, Faivre E et al (2015) The JAK/STAT3 pathway is a common inducer of astrocyte reactivity in Alzheimer's and Huntington's diseases. J Neurosci 35: 2817–2829 - PMC - PubMed
    1. Burda JE, Sofroniew MV (2014) Reactive gliosis and the multicellular response to CNS damage and disease. Neuron 81: 229–248 - PMC - PubMed
    1. Carter SF, Scholl M, Almkvist O, Wall A, Engler H, Langstrom B, Nordberg A (2012) Evidence for astrocytosis in prodromal alzheimer disease provided by 11C‐deuterium‐L‐deprenyl: a multitracer PET paradigm combining 11C‐pittsburgh compound B and 18F‐FDG. J Nucl Med 53: 37–46 - PubMed

Publication types