Transcriptomic alterations in APP/PS1 mice astrocytes lead to early postnatal axon initial segment structural changes

Abstract

Alzheimer´s disease (AD) is characterized by neuronal function loss and degeneration. The integrity of the axon initial segment (AIS) is essential to maintain neuronal function and output. AIS alterations are detected in human post-mortem AD brains and mice models, as well as, neurodevelopmental and mental disorders. However, the mechanisms leading to AIS deregulation in AD and the extrinsic glial origin are elusive. We studied early postnatal differences in AIS cellular/molecular mechanisms in wild-type or APP/PS1 mice and combined neuron-astrocyte co-cultures. We observed AIS integrity alterations, reduced ankyrinG expression and shortening, in APP/PS1 mice from P21 and loss of AIS integrity at 21 DIV in wild-type and APP/PS1 neurons in the presence of APP/PS1 astrocytes. AnkyrinG decrease is due to mRNAs and protein reduction of retinoic acid synthesis enzymes Rdh1 and Aldh1b1, as well as ADNP (Activity-dependent neuroprotective protein) in APP/PS1 astrocytes. This effect was mimicked by wild-type astrocytes expressing ADNP shRNA. In the presence of APP/PS1 astrocytes, wild-type neurons AIS is recovered by inhibition of retinoic acid degradation, and Adnp-derived NAP peptide (NAPVSIPQ) addition or P2X7 receptor inhibition, both regulated by retinoic acid levels. Moreover, P2X7 inhibitor treatment for 2 months impaired AIS disruption in APP/PS1 mice. Our findings extend current knowledge on AIS regulation, providing data to support the role of astrocytes in early postnatal AIS modulation. In conclusion, AD onset may be related to very early glial cell alterations that induce AIS and neuronal function changes, opening new therapeutic approaches to detect and avoid neuronal function loss.

Keywords: ADNP; AnkyrinG; Astrocytes; Axon initial segment; Neurodegeneration; P2X7; Retinoic acid.

Fig. 1

Early loss of ankyrinG density and AIS shortening in P21 APP/PS1 mice cortex. A Representative images of somatosensory cortex S1 region brain sections of wild-type or APP/PS1 mice compared in the same litters at postnatal day 15 (P15), 21 (P21), or 30 (P30). Brain sections were stained with antibodies against ankyrinG (AnkG, green) and β-amyloid (6E10, red). Nuclei were stained using bisbenzimide H33342 (blue). Inversed greyscale images show ankyrinG staining at each age. B Graphs show ankyrinG fluorescence intensity in WT (black) and APP/PS1 (red) mice (n = 4) normalized to the corresponding WT mice at different brain ages. C Normalized ankyrinG fluorescence intensity in WT (black) and APP/PS1 (red) mice normalized to P30 WT mice (n = 4). P30 WT sections were used as a reference in all immunofluorescence studies. D AIS length in WT (black) or APP/PS1 (red) mice at different ages (n = 4). Scale bar = 50 μm. Each data point represents the mean ± SEM of 4 mice (open symbols) and the total number of AISs is shown in B at the bottom for each point. *p < 0.05, n.s. (not significant), Mann–Whitney test

Fig. 2

Early loss of ankyrinG density and AIS shortening in P21 APP/PS1 mice hippocampus A Representative images of CA1 hippocampal sections of wild-type or APP/PS1 mice compared in the same litters at postnatal day 15 (P15), 21 (P21) or 30 (P30). Brain sections were stained with antibodies against ankyrinG (AnkG, green) and β-amyloid (6E10, red). Nuclei were stained using bis-benzimide (blue). Inversed greyscale images show ankyrinG staining at each age. B Graphs show ankyrinG fluorescence intensity in CA1 region of WT mice (black points) and APP/PS1 (red points) mice normalized to the corresponding WT mice at different brain ages. C AnkyrinG fluorescence intensity in WT (black) and APP/PS1 (red) mice normalized to P30 WT mice. P30 WT sections were used in all immunofluorescence studies as a reference. Data were obtained from 3 to 5 WT or APP/PS1 mice from the same litter and analyzing three CA1 sections in each mouse. D AIS length in WT (black circles) or APP/PS1 (red squares) mice at different ages. Each data point represents the mean ± SEM of the corresponding number of mice (open symbols). Total number of AISs is shown at the bottom. Scale bar = 75 μm. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Mann–Whitney test

Fig. 3

AnkyrinG expression in cultured APP/PS1 hippocampal neurons. A Schematic representation of the culture of WT and APP/PS1 hippocampal neurons in the presence of WT astrocytes. Hippocampal neurons were obtained from the same litter and the genotype was revealed after quantifications. Neurons were cultured during 6, 14, or 21 DIV. B Representative images of 14 and 21 DIV hippocampal neurons (WT or APP/PS1) cultured in the presence of WT astrocytes. Neurons were stained with antibodies against MAP2 (red) and ankyrinG (green). Scale bar = 25 μm. C Normalized ankyrinG density in 6, 14, and 21 DIV WT and APP/PS1 neurons. Data represents the mean ± SEM of the number of neurons indicated in the graph and obtained from three independent experiments containing each one 6, 14, and 21 DIV neurons. *p < 0.05, **p < 0.01, n.s. (not significant), Mann–Whitney test. D Mean ± SEM AIS length of neurons represented in C. *p < 0.05, (n.s. not significant), Mann–Whitney test

Fig. 4

APP/PS1 astrocytes decrease ankyrinG and AIS length of WT and APP/PS1 neurons. A Schematic representation of the culture of WT and APP/PS1 hippocampal neurons in the presence of WT or APP/PS1 astrocytes. Hippocampal neurons were obtained from the same litter and the genotype was revealed after quantifications. Neurons from every embryo were cultured in the presence of WT or APP/PS1 astrocytes for 21 DIV. B Representative images of wild-type 21 DIV hippocampal neurons cultured in the presence of WT or APP/PS1 astrocytes. Neurons were stained with antibodies against MAP2 (red) and ankyrinG (green). Scale bar = 25 μm. C Normalized ankyrinG density in 21 DIV WT or APP/PS1 neurons cultured in the presence of WT or APP/PS1 astrocytes. Data were normalized to WT neurons and WT astrocytes co-culture. D AIS length in different WT and APP/PS1 neurons or astrocytes combinations. Data number for each condition is indicated together with their mean ± SEM. Data were obtained from at least 5 independent experiments. **p < 0.01, ***p < 0.001, p < 0.0001, n.s. (not significant). 2-way ANOVA test followed by Bonferroni test for multiple comparisons. 2-way ANOVA analysis for AnkyrinG intensity or AIS length (GraphPad) shows an extremely significant effect of astrocytes genotype (p < 0.0001 and p < 0.0001), while neurons genotype effect is considered not significant (p = 0.1767 and p = 0.6688, respectively)

Fig. 5

Comparative transcriptomic analysis between wild-type and APP/PS1 astrocytes. A Volcano plot showing the global transcriptional changes in APP/PS1 astrocytes vs wild-type astrocytes. All genes identified were plotted. Each circle represents one gene. The log fold change (log2(FC)) in APP/PS1 genotype versus wild type is represented on the x-axis. The y-axis shows the log10 of the p value. A p-value of 0.03 (log10(p-value) = 1.5) and a log2(fold change) of 0.3 and – 0.3 of 2 are indicated by grey dotted lines. Up-regulated and down-regulated genes are shown in blue and red, respectively. Black points represent genes without a significative p-value and a log2(FC) between -0.3 and 0.3. B GO analysis of genes shown in red and blue in A. C Tables show the genes found in each biological process category after gene ontology analysis. D Volcano plot showing the transcriptional changes in genes identified in GO analysis and other genes with a significative change. E Heatmap showing the normalized expression of genes shown in D, in each one of the three WT or APP/PS1 mice used to obtain cultured astrocytes

Fig. 6

Retinoic acid synthesis enzymes downregulation in APP/PS1 astrocytes contribute to axon initial segment alterations. A Volcan plots of Rdhs and Aldhs genes identified in astrocytes. Rdh1 and Aldh1b1 are shown in red and are the only ones to significantly change their expression. All other Rdhs and Aldhs genes are shown in black and named in the graph. B Western-blot showing Aldh1b1 (green) expression in WT and APP/PS1 astrocytes in culture. Data are normalized to their respective tubulin levels (red) in 3 independent experiments. *p < 0.01, t-test. C Images show Aldh1b1 expression (green or greyscale). Astrocytes morphology was identified by acetylated tubulin staining (red) and nuclei are shown in blue. Scale bar = 100 μm. The graph represents the normalized Aldh1b1 protein expression quantified by immunofluorescence in 50 WT astrocytes and 50 APP/PS1 astrocytes obtained from 3 WT or APP/PS1 mice. *p < 0.05, Mann–Whitney test. D Axon initial length and normalized ankyrinG intensity in WT neurons co-cultured with WT astrocytes or APP/PS1 astrocytes, and treated with vehicle or an inhibitor of CYP26 enzyme for 3 days. Each condition represents the mean ± SEM of the neurons indicated in panel D obtained from 3 independent experiments. *p < 0.05, **p < 0.01. 2-way ANOVA and multiple comparisons Bonferroni test.

Fig. 7

P2X7 inhibition recovers ankyrinG and AIS length in APP/PS1 mice and cultured neurons. A, D Representative images of cortical or hippocampal (CA1) sections from WT or APP/PS1 mice treated with PBS or P2X7 inhibitor BBG from 1 to 3 months. Sections were stained with antibodies against ankyrinG (green) and β-amyloid (6E10, red). Nuclei are shown in blue. Greyscale images of ankyrinG staining are shown for the hippocampus. B, C Normalized ankyrinG intensity (C) and AIS length (D) in cortical sections of WT, APP/PS1 and APP/PS1 mice treated with BBG. Data were obtained from 6 mice per condition and the number of AISs analyzed is indicated in the graphs. E, F Normalized ankyrinG intensity (E) and AIS length (F) in the hippocampus (CA1). AnkyrinG intensity in CA1 was calculated as the ankyrinG staining region. Data were obtained from 6 mice per condition and the number of CA1 sections indicated in E. AIS length was calculated from the indicated number of AISs in the CA1 region. ***p < 0.001, n.s. (not significant). One-way ANOVA and Kruskal-Willis post-hoc test with Dunn´s test multiple comparisons. G Representative images of 21 DIV WT neurons cultured in the presence of WT or APP/PS1 astrocytes and treated for 24 h with PBS or the P2X7 inhibitor BBG (100 nM). Neurons were stained with MAP2 (red) and AnkyrinG (green) antibodies. AISs magnifications are shown in the right panels. Scale bar = 25 μm. H, I Normalized ankyrinG (H) intensity and AIS length (I) in WT or APP/PS1 (WTn or TGn) cultured neurons in the presence of WT or APP/PS1 astrocytes treated with PBS or BBG (100 nM). Two-way ANOVA and Bonferroni test for multiple comparisons was performed in neurons cultured with WT astrocytes or in neurons cultured with APP/PS1 astrocytes. **p < 0.01, ****p < 0.0001, n.s. (not significant). J Representative western-blot showing P2X7 expression in cultured WT and APP/PS1 astrocytes. Graph represents the mean ± SEM of 3 independent experiments. **p < 0.01, t-test. K Adnp expression in WT and APP/PS1 astrocytes treated with vehicle or BBG (100 nM) for 24 h. The image shows a representative Western-blot of Adnp (red) and tubulin (green) expression. Data were obtained from 6 independent experiments and lines connect points from the same experiment. *p < 0.05, **p < 0.01, n.s. (not significant), two-way ANOVA and Bonferroni test for multiple comparisons

Fig. 8

The Adnp-derived peptide, NAP, recovers ankyrinG and AIS length. A Representative images of WT and APP/PS1 astrocytes stained with Adnp (green or grey) antibody and phalloidin-Alexa 568 (Actin, red). Scale bar = 50 μm. B Normalized Adnp protein expression in WT or APP/PS1 cultured astrocytes and representative Western blot. Adnp was normalized in each sample to their corresponding tubulin intensity. Data were obtained from 6 WT and 13 APP/PS1 independent astrocytes cultures. C Western-blot showing Adnp, tubulin and GFP expression in WT astrocytes 7 days after transduction with lentiviral particles expressing scramble shRNA and three different ADNP shRNAs (sh1-3). The graph show the normalized Adnp expression normalized to tubulin expression in 3 independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001, t-test. D, E Representative images showing ADNP expression in WT astrocytes after 7 days of transduction with a scramble and ADNP shRNAs lentiviral particles. GFP and tubulin stainings of these cells are shown in supplementary Fig. 3. The graph represents the normalized ADNP expression. Data are the mean ± SEM of the indicated number of astrocytes analyzed from 3 independent experiments. ***p < 0.001, one-way ANOVA and Kruskal-Willis post-hoc test with Dunn´s test multiple comparisons. F, G Normalized ankyrinG intensity and AIS length in WT neurons co-cultured with astrocytes expressing scramble shRNA or ADNP shRNAs 1–3. *p < 0.05, ***p < 0.001, one-way ANOVA and Kruskal-Willis post-hoc test with Dunn´s test multiple comparisons. H Representative images of 21 DIV WT neurons cultured in the presence of WT or APP/PS1 astrocytes and treated for 24 h with vehicle or the NAP peptide (10^–9 M). Neurons were stained with MAP2 (red) and AnkyrinG (green) antibodies. AISs magnifications are shown in the bottom panels. Scale bar = 50 μm. I, J Normalized ankyrinG (G) intensity and AIS length (H) in WT cultured neurons in the presence of WT or APP/PS1 astrocytes treated with vehicle or NAP peptide. Data were acquired from 4 independent experiments and around 190 neurons in each condition as shown in the graph. **p < 0.01, ***p < 0.01, ****p < 0.0001. Two-way ANOVA and Bonferroni test for multiple comparisons