A novel mechanism of synaptic and cognitive impairments mediated via microRNA-30b in Alzheimer's disease

Abstract

Background: It is widely accepted that cognitive and memory deficits in Alzheimer's disease (AD) primarily result from synaptic failure. However, the mechanisms that underlie synaptic and cognitive dysfunction remain unclear.

Methods: We utilized molecular biology techniques, electrophysiological recordings, fluorescence in situ hybridization (FISH), immuno- and Golgi-staining, chromatin immunoprecipitation (CHIP); lentivirus (LV)-based microRNA overexpression and 'sponging', and behavioral tests to assess upregulated miR-30b causing synaptic and cognitive declines in APP transgenic (TG) mice.

Findings: We provide evidence that expression of miR-30b, which targets molecules important for maintaining synaptic integrity, including ephrin type-B receptor 2 (ephB2), sirtuin1 (sirt1), and glutamate ionotropic receptor AMPA type subunit 2 (GluA2), is robustly upregulated in the brains of both AD patients and APP transgenic (TG) mice, an animal model of AD, while expression of its targets is significantly downregulated. Overexpression of miR-30b in the hippocampus of normal wild-type (WT) mice impairs synaptic and cognitive functions, mimicking those seen in TG mice. Conversely, knockdown of endogenous miR-30b in TG mice prevents synaptic and cognitive decline. We further observed that expression of miR-30b is upregulated by proinflammatory cytokines and Aβ42 through NF-κB signaling.

Interpretation: Our results provide a previously undefined mechanism by which unregulated miR-30b causes synaptic and cognitive dysfunction in AD, suggesting that reversal of dysregulated miR-30b in the brain may prevent or slow cognitive declines in AD. FUND: This work was supported by National Institutes of Health grants R01NS076815, R01MH113535, R01AG058621, P30GM103340 Pilot Project, and by the LSUHSC School of Medicine Research Enhancement Program grant (to C.C.).

Keywords: Alzheimer's disease; Dementia; Neuroinflammation; Nuclear factor kappa B; Small noncoding RNA; Synaptic failure; miRNA sponge.

Fig. 1

Non-coding small RNA miR-30b that targets the molecules important for synapses is upregulated in Alzheimer's disease (AD). (a) Expression of miR-30b is upregulated in the hippocampus of both AD patients and 5XFAD transgenic (TG) mice (Animals: ***P < 0.001, Student's t-test, n = 22/group; Humans: ***P < 0.001, n = 14/normal and n = 15/AD). The postmortem human hippocampal tissues were from normal controls at ages of 80.1 ± 3.3 years and from AD patients at ages of 76.3 ± 3.1 years with the mean post-mortem interval (PMI) 7.4 ± 1.2 h and 6.6 ± 0.9 h, respectively. (b) Binding sites of miR-30b in the 3′UTRs of ephB2, sirt1 and GluA2 and mutated binding sites in the 3′UTRs of ephB2, sirt1 and GluA2 recognized by the seed region of miR-30b. The luciferase reporter activity in 293 cells transfected with psiCHECK vector of wild-type (normal) and mutated binding sites in the 3′UTR of ephB2, sirt1, and GluA2 recognized by the seed region of miR-30b in the absence and presence of miR-30b. (**P < 0.01, n = 3/group). (c) Immunoblot analysis of ephB2, sirt1, and GluA2 expression in cultured hippocampal neurons transfected with LV-miR-30b or LV-control (**P < 0.0001, n = 7/group). (d) Immunostaining analysis of ephB2, sirt1, and GluA2 expression in cultured hippocampal neurons treated with LV-miR-30b or LV-control. Scale bar: 20 μm. (e) Endogenous miR-30b is sequestered by miR-30b ‘sponging’ (miR-30bs). FISH analysis was performed to detect expression of miR-30b in cultured hippocampal neurons treated with LV-scramble control or LV-miR-30bs. Scale bar: 30 μm. (f) qPCR analysis of knockdown of miR-30b by miR-30bs. **P < 0.01, n = 3/group with duplicates. (g) Spontaneous miniature excitatory postsynaptic currents (mEPSCs) recorded in cultured hippocampal neurons treated with LV-miR-30b, -LV-miR-30bs (‘sponge’) or -LV-control (*P < 0.05 compared with LV-control, ##P < 0.01 compared with LV-miR-30b, ANOVA with Bonferroni post hoc-test, n = 61 recordings in LV-control group, 58 in LV-miR-30b, and 51 in LV-miR-30bs, respectively). (h) Glutamate-induced currents in cell-free outside-out patches excised from cultured hippocampal neurons treated with LV-control, miR-30b, or miR-30bs (n = 25: LV-control group, n = 30: LV-miR-30b, and n = 28: LV-miR-30bs, respectively). Glutamate-induced currents were induced through a Burleigh fast solution switching system to deliver control and glutamate solutions via a theta glass tubing. (i) Stereotaxic injection of LV into the hippocampus and enlarged GFP expression in dentate granule neurons and CA1 pyramidal neurons. Scale bar: 200 μm. (j) EPSCs recorded at holding potential of +50 and − 70 mV in hippocampal slices from normal C57BL/6 mice injected with LV-control, −miR-30b, or –miR-30bs (** < 0.01, compared with LV-control, ##P < 0.01 compared with LV-miR-30b, ANOVA with Bonferroni post hoc-test, n = 18: LV-control group, n = 19: LV-miR-30b, and n = 22: LV-miR-30bs, respectively). (k) Input-output function of EPSCs at hippocampal synapses (n = 23: LV-control group, n = 25: LV-miR-30b, and n = 25: LV-miR-30bs, respectively).

Fig. 2

miR-30b regulates ephB2, sirt1 and GluA2 in the hippocampus. (a) Expression of ephB2, sirt1, and GluA2 is significantly reduced in the hippocampi of both AD patients and TG mice (Humans: **P < 0.01, Student's t-test, n = 4/group; Animals: **P < 0.01, n = 3/group). (b) Changes in expression levels of hippocampal miR-30b by injections of LV-miR-30b in WT mice or LV-miR-30bs in TG mice. LV-control, LV-miR-30b or miR-30bs were stereotaxically injected into the hippocampus of WT or TG mice. qPCR analysis of hippocampal miR-30b expression was conducted 8 weeks after injections. Injection of LV-miR-30b raises hippocampal expression levels of miR-30b in WT mice, while injection of LV-miR-30bs results in a reduction of endogenous hippocampal miR-30b in TG mice. **P < 0.01 compared with WT LV-control and ##P < 0.01 compared with TG LV-control (n = 6/group). (c) Overexpression of miR-30b in the hippocampus represses its targets ephB2, sirt1, and GluA2 in WT mice, while knockdown of miR-30b restores reduced targets in TG mice (ANOVA with Fisher's PLSD test, **P < 0.01 compared with WT-LV-control, ##P < 0.01 compared with TG-LV-control, n = 3/group). (d) Knockdown of endogenous miR-30b does not alter hippocampal production of Aβ42 and the enzymes responsible for synthesis of Aβ. Immunoblot analysis of Aβ42 and and expression of α (ADAM-10), β (BACE1)- and γ (Nicastrin)-secretases in 6-month-old WT or TG mice that received LV-miR-30b or LV-miR-30bs. **P < 0.01 compared with WT LV-control (n = 3 mice/group). (e) Knockdown of endogenous miR-30b by miR-30b sponge does not alter Aβ formation. Immunostaining analysis of total Aβ (all forms) detected by anti-4G8 antibody in the hippocampus of 6-month-old WT mice that received LV-control or LV-miR-30b at 4 months of age or in 6-month-old TG mice that received LV-control or LV-miR-30bs at 4 months of age (n = 5 mice/group).

Fig. 3

Over-expression of miR-30b impairs synaptic and cognitive function in WT mice and knockdown of miR-30b prevents synaptic and cognitive declines in TG mice. (a) Input-output function recorded at hippocampal perforant synapses (n = 7 animals/WT-LV-Con, n = 6 animals/WT-LV-30b, n = 6 animals/TG-LV-Con, and n = 6 animals/TG-LV-30bs). LV-control, LV-miR-30b, or LV-miR-30bs were stereotaxically injected into the hippocampus in WT or TG mice at 4 months of age. Synaptic and cognitive functions were assessed 2 months after LV injections. (b) Long-term potentiation (LTP) at perforant synapses (n = 6 animals/WT-LV-Con, n = 8 animals/WT-LV-30b, n = 8 animals/TG-LV-Con, and n = 6 animals/TG-LV-30bs). (c) Learning acquisition in the Morris water maze test (ANOVA with repeated measures, P < 0.001, n = 19 animals/group). (d) Crossing the platform in the probe test (ANOVA with Bonferroni post hoc-test, *P < 0.05, **P < 0.01 compared with WT-LV-Con, ^##P < 0.001 compared with TG-LV-Con, n = 19/group). (e) The percentage of time spent in the target quadrant during the probe trial. (ANOVA with Bonferroni post hoc-test, *P < 0.05 compared with WT-LV-Con, ^#P < 0.001 compared with TG-LV-Con, n = 19/group). (f) Novel object recognition test (ANOVA with Bonferroni post hoc-test. **P < 0.01 compared with WT-LV-Con, ^##P < 0.01 compared with TG-LV-Con, n = 10/group).

Fig. 4

miR-30b regulates the integrity of synaptic structure. (a) Golgi staining of dendritic spines in hippocampal CA1 pyramidal neurons and dentate granule neurons from WT or TG mice injected with LV-30b, LV-30bs or LV-Con for 8 weeks (ANOVA with Bonferroni post hoc-test, **P < 0.01 compared with WT-LV-Con, ^##P < 0.01 compared with TG-LV-Con, n = 6 animals/group). Scale bars: 50 and 5 μm. (b) hippocampal expression of glutamate receptor subunits and PSD-95 in WT or TG mice injected with LV-Con, LV-30b or LV-30bs (ANOVA with Fisher's PLSD test, *P < 0.05, **P < 0.01 compared with WT-LV-Con; ^#P < 0.05, ^##P < 0.01 compared with TG-LV-Con n = 3–6/group).

Fig. 5

Expression of miR-30b in the brain is triggered by neuroinflammation and regulated by NF-κB signaling pathway. (a) Age-dependent elevation of miR-30b expression in TG mice (ANOVA with Fisher's PLSD test, n = 3/group, *P < 0.05, **P < 0.01 compared with WT mice at one-month-old, n = 3/group). (b) Expression of miR-30b is upregulated by proinflammatory cytokines IL-1β (10 ng/ml), TNFα (10 ng/ml) or Aβ (10 μM) in cultured hippocampal neurons and the upregulation is blocked by SC-514 (SC, 100 μM). ANOVA with Fisher's PLSD test, ** < 0.01, compared with the vehicle control, ##P < 0.01 compared with Aβ, IL-1β, or TNFα (n = 3/group). (c) Increased phosphorylated NF-κB in the hippocampi of both AD patients and TG mice. Student's t-test, ** < 0.01, compared with normal subjects or WT controls (n = 4/group for human samples and n = 10/group for mouse samples). (d) Binding of the NF-ĸB p65 subunit in the promoter of the miR-30b gene assessed by chromatin immunoprecipitation (CHIP) analysis. (e) Silencing of the NF-ĸB p65 subunit blocks cytokines-induced increase in expression of miR-30b. Hippocampal neurons in culture were transfected with lentiviral vectors expressing control or NF-ĸB p65 shRNA (ANOVA with Fisher's PLSD test, ** < 0.01, compared with the vehicle control, n = 3/group with duplicates).

Fig. 6

Hypothetical signaling pathways involved in neuroinflammation triggered-upregulation of miR-30b that contributes to synaptic and cognitive deficits in AD.