iPSC-derived PSEN2 (N141I) astrocytes and microglia exhibit a primed inflammatory phenotype

Abstract

Background: Widescale evidence points to the involvement of glia and immune pathways in the progression of Alzheimer's disease (AD). AD-associated iPSC-derived glial cells show a diverse range of AD-related phenotypic states encompassing cytokine/chemokine release, phagocytosis and morphological profiles, but to date studies are limited to cells derived from PSEN1, APOE and APP mutations or sporadic patients. The aim of the current study was to successfully differentiate iPSC-derived microglia and astrocytes from patients harbouring an AD-causative PSEN2 (N141I) mutation and characterise the inflammatory and morphological profile of these cells.

Methods: iPSCs from three healthy control individuals and three familial AD patients harbouring a heterozygous PSEN2 (N141I) mutation were used to derive astrocytes and microglia-like cells and cell identity and morphology were characterised through immunofluorescent microscopy. Cellular characterisation involved the stimulation of these cells by LPS and Aβ₄₂ and analysis of cytokine/chemokine release was conducted through ELISAs and multi-cytokine arrays. The phagocytic capacity of these cells was then indexed by the uptake of fluorescently-labelled fibrillar Aβ₄₂.

Results: AD-derived astrocytes and microglia-like cells exhibited an atrophied and less complex morphological appearance than healthy controls. AD-derived astrocytes showed increased basal expression of GFAP, S100β and increased secretion and phagocytosis of Aβ₄₂ while AD-derived microglia-like cells showed decreased IL-8 secretion compared to healthy controls. Upon immunological challenge AD-derived astrocytes and microglia-like cells showed exaggerated secretion of the pro-inflammatory IL-6, CXCL1, ICAM-1 and IL-8 from astrocytes and IL-18 and MIF from microglia.

Conclusion: Our study showed, for the first time, the differentiation and characterisation of iPSC-derived astrocytes and microglia-like cells harbouring a PSEN2 (N141I) mutation. PSEN2 (N141I)-mutant astrocytes and microglia-like cells presented with a 'primed' phenotype characterised by reduced morphological complexity, exaggerated pro-inflammatory cytokine secretion and altered Aβ₄₂ production and phagocytosis.

Keywords: Alzheimer’s disease; Amyloid-beta; Astrocytes; Glia; Microglia; Morphology; PSEN2; Pro-inflammatory; iPSC.

Fig. 1

A Overview of differentiation timeline for iPSC-derived NPCs showing representative brightfield images on the last day of each stage. Representative immunofluorescence images of iPSC-derived NPCs from healthy control lines and familial AD lines harbouring a PSEN2 (N141I) mutation. The cells were stained for B the neural progenitor markers Pax-6 (red) and Nestin (green), C a pluripotency marker Oct3 (green) and all nuclei were counterstained with DAPI (blue). Insert shows higher magnification. All scale bars = 50 μm. The percentage of iPSC-derived NPCs positive for D Pax6 and E nestin, with each bar displaying the mean ± SD of three cell lines with n ≥ 2 independent experiments per line. Immunofluorescence staining of all iPSC and astrocyte cell lines can be found in Additional file 1: Fig. S2, S3 and S4

Fig. 2

A Representative immunofluorescence images of iPSC-derived astrocytes from healthy control lines and familial AD lines harbouring a PSEN2 (N141I) mutation. The cells were stained for astrocyte markers GFAP (red) and S100β (green) and B the NPC marker nestin (green) and all nuclei were counterstained with DAPI (blue). Insert shows higher magnification. All scale bars = 50 μm. The percentage of iPSC-derived astrocytes positive for C) GFAP and D S100β. Mean fluorescence intensity of E GFAP and F S100β per astrocyte. An unpaired t-test was used to test whether there were statistically significant differences between the means of AD-derived and healthy control astrocytes. Each bar displays the mean ± SD of three cell lines with n ≥ 2 independent experiments per line (**p < 0.01). Immunofluorescence staining of all cell lines can be found in Additional file 1: Fig. S4

Fig. 3

A Overview of differentiation timeline for iPSC-derived microglia-like cells showing representative brightfield images on the last day of each stage. B Representative immunofluorescence images of iPSC-derived microglia from healthy control lines and familial AD lines harbouring a PSEN2 (N141I) mutation. The cells were stained for microglial markers IBA1 (red), TREM2 (green) and all nuclei were counterstained with DAPI (blue). Insert shows higher magnification. All scale bars = 50 μm. The percentage of iPSC-derived microglia-like cells positive for C IBA1 and D TREM2. Mean fluorescence intensity of E IBA1 and F TREM2 per cell. An unpaired t-test was used to test whether there were statistically significant differences between the means of AD-derived and healthy control microglia. Each bar displays the mean ± SD of three cell lines with n ≥ 2 independent experiments per line (**p < 0.01). Immunofluorescence staining of all cell lines can be found in Additional file 1: Fig. S6

Fig. 4

Morphological characterisation of iPSC-derived astrocytes from three healthy control and three familial AD lines harbouring a PSEN2 (N141I) mutation. Quantification of A average cell perimeter, B average cell circularity and C average cell volume were determined using GFAP and S100β immunofluorescence and 3D reconstruction within FIJI. D Individual cells were categorised and based on morphological appearance, with representative phenotypes of ‘arborised’, ‘polarised’ and ‘process devoid’ cells illustrated in E. An unpaired t-test (A–C) or two-way ANOVA with a post hoc Dunnett’s test (D) were used to test whether there were statistically significant differences between the means of AD-derived and healthy control astrocytes for each morphological category. The figure displays the mean ± SD of three cell lines with n ≥ 2 independent experiments per line (*p < 0.05, **p < 0.01, ***p < 0.001). Scale bar = 50 μm

Fig. 5

A An example of the processing of an immunofluorescence image of iPSC-derived microglia-like cells stained for IBA1 (red), the result of image processing is shown in binary and was followed by skeletonisation (green). An overlay of the original IBA1 image (red) and skeleton (green) is shown for comparison. Automated analysis of the mean number of B junctions, C branches, D endpoints and E longest branch length per cell. An unpaired t-test was used to test whether there were statistically significant differences between the means of AD-derived and healthy control microglia-like cells. The figure displays the mean ± SD of three cell lines with n = 2 independent experiments per line (*p < 0.05)

Fig. 6

Concentration of IL-6 secreted from healthy control or AD-derived astrocytes A basally and after exposure to 10 or 50 μg/mL LPS for 24 h. B Data from part A represented as individual cell lines. Cell lines were re-categorised based on their APOE ε3/ε3 (Ctrl-06, Ctrl-71 and fAD-08) or APOE ε3/ε4 (Ctrl-88, fAD-948 and fAD-950) genotype and shows the concentration of IL-6 secretion C basally and after exposure to 10 or 50 μg/mL LPS for 24 h. A two-way ANOVA with multiple comparisons and post hoc Dunnett’s test was used to test whether there were statistically significant differences between the means of AD-derived and healthy control astrocytes or APOE ε3/ε3 and APOE ε3/ε4 astrocytes (***p < 0.001). Statistical tests were not conducted for figure B. Figures A and C display the mean ± SD of three cell lines with n ≥ 2 independent experiments per line, B display the mean ± SD of n ≥ 2 independent experiments per line

Fig. 7

The concentration of A secreted IL-6 and B viability of AD-derived and healthy control astrocytes after stimulation with initially monomeric 0, 5 or 10 μM Aβ₄₂ for 24 h. Cell lines were re-categorised based on their APOE ε3/ε3 (Ctrl-06, Ctrl-71 and fAD-08) or APOE ε3/ε4 (Ctrl-88, fAD-948 and fAD-950) genotype and shows the concentration of C secreted IL-6 and D viability after exposure to 0, 5 or 10 μM Aβ₄₂ for 24 h. The figure displays the mean ± SD of three cell lines with n ≥ 2 independent experiments per line. A two-way ANOVA with post hoc Dunnett’s test was used to test whether there were statistically significant differences between mean of AD-derived and healthy control astrocytes and APOE ε3/ε3 and APOE ε3/ε4 astrocytes (**p < 0.01, ***p < 0.001)

Fig. 8

Multi-cytokine array of AD-derived or healthy iPSC-derived astrocytes A basally and B after 24 h exposure to 10 μM Aβ₄₂ and iPSC-derived microglia-like cells C basally and D after 24 h exposure to 10 μM Aβ₄₂. The figure displays the mean ± SD of three cell lines with the average of two experimental duplicates per line. Multiple unpaired, non-parametric Mann–Whitney t-tests adjusting for a 5% false discovery rate were used to test whether there were statistically significant differences between mean cytokine/chemokine release of AD-derived and healthy control astrocytes and microglia-like cells (*p < 0.05, **p < 0.01). Cytokines/chemokines with mean pixel density under 10% of the maximum in both basal and Aβ₄₂-stimulated conditions were considered background and are not shown, full cytokine array is shown in Additional file 1: Figs. S7 and S8

Fig. 9

Concentrations of A Aβ₄₂, B Aβ₄₀, C the ratio of Aβ_42:40 and D total Aβ protein quantified from iPSC-derived astrocyte supernatants 72 h after plating. Secreted Aβ concentrations were measured using a highly sensitive ELISA and normalised to total protein concentration determined by BCA. The figure displays the mean ± SD of three cell lines with n ≥ 2 independent experiments per line. An unpaired t-test was used to test whether there were statistically significant differences between mean of AD-derived and healthy control astrocytes (*p < 0.05)

Fig. 10

AD-derived and healthy A iPSC-derived astrocytes and B iPSC-derived microglia-like cells were exposed to 488-fluorescent-Aβ₄₂ fibrils for 2 h and phagocytosis was quantified through average fluorescent intensity per cell. The figure shows the percentage of B iPSC-derived astrocytes and D iPSC-derived microglia-like cells which show any uptake of Aβ₄₂. The figure displays the mean ± SD of three cell lines with n ≥ 2 independent experiments per line. An unpaired t-test was used to test whether there were statistically significant differences between mean of AD-derived and healthy control astrocytes (*p < 0.05). Representative overlay of brightfield and fluorescent images of E iPSC-derived astrocytes and F iPSC-derived microglia-like cells, fluorescent Aβ₄₂ fibrils shown in green, scale bar = 50 μm

Fig. 11

Graphical representation and summary of morphological and immunological features of iPSC-derived microglia-like cells (top) and astrocytes (bottom) from healthy (left) and PSEN2 (N141I)-mutant fAD (right) origin. Created with Biorender.com