Induction of astrocyte reactivity promotes neurodegeneration in human pluripotent stem cell models

Abstract

Reactive astrocytes are known to exert detrimental effects upon neurons in several neurodegenerative diseases, yet our understanding of how astrocytes promote neurotoxicity remains incomplete, especially in human systems. In this study, we leveraged human pluripotent stem cell (hPSC) models to examine how reactivity alters astrocyte function and mediates neurodegeneration. hPSC-derived astrocytes were induced to a reactive phenotype, at which point they exhibited a hypertrophic profile and increased complement C3 expression. Functionally, reactive astrocytes displayed decreased intracellular calcium, elevated phagocytic capacity, and decreased contribution to the blood-brain barrier. Subsequently, co-culture of reactive astrocytes with a variety of neuronal cell types promoted morphological and functional alterations. Furthermore, when reactivity was induced in astrocytes from patient-specific hPSCs (glaucoma, Alzheimer's disease, and amyotrophic lateral sclerosis), the reactive state exacerbated astrocytic disease-associated phenotypes. These results demonstrate how reactive astrocytes modulate neurodegeneration, significantly contributing to our understanding of a role for reactive astrocytes in neurodegenerative diseases.

Keywords: astrocyte; differentiation; disease; neurodegeneration; reactivity; stem cell.

Figure 1

Induction of a reactive phenotype in hPSC-derived astrocytes (A and B) Reactive astrocytes exhibited a hypertrophic profile as seen by S100β and GFAP staining. (C and D) Increased accumulation of C3 staining in the perinuclear region of reactive astrocytes. (E–H) Morphological analysis showed an increased number of branches and shorter processes in reactive astrocytes, as well as an overall increased number of process intersections through Sholl analysis. (I and J) Increased C3 expression exhibited by reactive astrocytes through western blot analysis. (K) qPCR analysis showed that reactive astrocytes express increased levels of pan-reactive associated genes, as well as genes that specifically characterize the phenotype of neurotoxic reactive astrocytes. Results are represented as fold change vs. control astrocytes in (J) and (K). Colored dots correspond to 3 different lines, H7 hPSC line in blue, mips2 hPSC line in gray, and WTC11 hPSC line in orange. ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001 vs. control samples, two-tailed paired Student’s t test. Scale bar: 30 μm in (A) and (B), 20 μm in (C) and (D). Data represents mean values ± SEM.

Figure 2

Transcriptional analysis of reactive hPSC-derived astrocytes compared to controls (A) Heatmap representing differentially expressed genes in reactive and control astrocytes. The up- and downregulated genes are represented as red and blue colored, respectively. (B) Volcano plot exhibiting differentially expressed genes in reactive astrocytes compared to controls. Significantly upregulated genes are shown in blue, while significantly downregulated genes are shown in green. (C and D) GO and pathway analysis based on differential gene expression of reactive and control astrocytes.

Figure 3

Functional changes in hPSC-derived astrocytes following induction of reactivity (A and B) Representative images of pHrodo particles engulfed by control and reactive astrocytes. (C) The percentage of astrocytes that had engulfed pHrodo particles significantly increased from 16 to 48 h. (D–F) Representative images of Fluo-4 AM-positive astrocytes following stimulation with ATP, and (F) quantification demonstrating a significant decrease in fluorescence intensity in reactive astrocytes compared to control cells. ANOVA followed by Tukey’s multiple comparison test or by two-tailed paired Student’s t test, in (C) and (F), respectively. (G) Meso Scale Discovery analysis showed an increased secretion of inflammatory factors by reactive astrocytes compared to control astrocytes. Two-tailed unpaired Student’s t test with Welch’s correction. ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001, and ^∗∗∗∗p < 0.0001. Data represent mean values ±SEM. Scale bar represents 100 μm in (B) and 30 μm in (E).

Figure 4

Reactive hPSC-derived astrocytes contribute to blood-brain barrier disruption Co-cultures were performed using a transwell system, with brain microvascular endothelial cells (BMECs) seeded on the transwell and astrocytes seeded in the bottom of the well. (A–I) Representative images of iPSC-derived BMECs expressing characteristic endothelial cell markers, including Claudin-5, glucose transporter 1 (GLUT-1), Occludin, and zonula occludens 1 (ZO-1). Reactive astrocytes induced a reduction in transendothelial electrical resistance (TEER) levels (G), increased fluorescein permeability (H), and elevated rhodamine 123 transport (I), compared to the effect promoted by control astrocytes. CsA, cyclosporin. ANOVA followed by Tukey’s multiple comparison test in (G) and (H) and by Šídák’s multiple comparisons test in (I). Data represent mean values ±SEM. ^∗p < 0.05 and ^∗∗p < 0.01, BMEC monoculture vs. BMECs co-cultured with control astrocytes; ^###p < 0.001, BMEC monoculture vs. BMECs co-cultured with reactive astrocytes; ^&&p < 0.01, BMECs co-cultured with control astrocytes vs. BMECs co-cultured with reactive astrocytes in (G). ^∗∗∗p < 0.001 and ^∗∗∗∗p < 0.0001 in (H) and (I). Scale bar: 50 μm.

Figure 5

hPSC-derived reactive astrocytes induce retinal ganglion cell neurotoxicity and hyperexcitability Astrocytes were co-cultured with RGCs, in either direct-contact or transwell cultures and induced to a reactive state through incubation with IL-1α, TNF-α, and C1q for 2 weeks. Representative images of RGCs (BRN3b:tdTomato) co-cultured with control or reactive astrocytes, in direct co-cultures (A and B), or transwell cultures (C and D). (E and F) Reactive astrocytes promoted a reduction in the number of primary neurites and total neurite extension in RGCs, in both direct-contact and transwell culture systems. (G) The overall complexity of RGC neurite outgrowth, assessed by Sholl analysis, was significantly decreased in both direct-contact and transwell culture systems. ^∗∗∗p < 0.001, ^∗∗∗∗p < 0.0001 in (E) and (F). ^∗∗∗∗p < 0.0001, RGCs co-cultured with control astrocytes vs. reactive astrocytes in contact-dependent cultures, and ^####p < 0.0001, RGCs co-cultured with control astrocytes vs. reactive astrocytes in transwell cultures, in (G). One-Way ANOVA followed by Šídák’s multiple comparisons test with selected pair. (H–J) Upon whole-cell patch-clamp recording, RGCs were less likely to be spontaneously active in co-culture with reactive astrocytes compared to controls yet fired more action potentials upon delivery of a depolarizing current. (K–M) RGCs co-cultured with reactive astrocytes had a higher frequency of action potentials fired, while they also exhibited a significantly lower action potential current threshold and capacitance. ^∗p < 0.05 and ^∗∗∗∗p < 0.0001. Scale bar: 100 μm. Data represent mean values ± SEM.

Figure 6

hPSC-derived reactive astrocytes also induce neurotoxic phenotypes upon cortical neurons and motor neurons (A) Schematic demonstrating the derivation of both cortical neurons and motor neurons from hPSCs, with resulting cortical neurons identified by CTIP2 and MAP2 expression, while motor neurons could be identified by HB9 and βIII-tubulin expression. Images in Panel A created with BioRender.com. (B–H) Cortical neurons were grown either in direct contact with astrocytes (B and C) or in transwell models (D and E). After two weeks of co-culture, reactive astrocytes resulted in a significantly decreased number of primary neurites from cortical neurons (F), a decrease in total neurite length (G), as well as in overall morphological complexity by Sholl analysis (H). (I–O) Similarly, motor neurons were grown either in direct contact with astrocytes (I and J) or in transwell models (K and L). Morphological analyses demonstrated that reactive astrocytes promoted a reduction in the number of primary neurites in motor neurons, only in direct-contact culture system (M), as well as a significant reduction in total neurite length (N) and overall outgrowth complexity (O). Data represent mean values ±SEM from at least four independent experiments per hPSC line. One-Way ANOVA followed by Šídák’s multiple comparisons test with selected pair; ^∗p < 0.05, ^∗∗∗p < 0.001, ^∗∗∗∗p < 0.0001, and n.s. means non-significant, in (F), (G), (M), and (N). ^∗∗∗∗p < 0.0001, cortical neurons (in H) or motor neurons (in O) co-cultured with control astrocytes vs. reactive astrocytes in contact-dependent cultures, and ^####p < 0.0001, cortical neurons (in H) or motor neurons (in O) co-cultured with control astrocytes vs. reactive astrocytes in transwell cultures. Scale bar: 100 μm.

Figure 7

Induction of a reactive phenotype exacerbates disease-associated phenotypes in patient-specific cell lines Astrocytes were differentiated from hPSC lines carrying mutations associated with either glaucoma OPTN(E50K), Alzheimer’s disease (PSEN1-N141I), or ALS (SOD1-N139K). (A–L) Upon induction of reactivity, astrocytes from all backgrounds developed a hypertrophic profile (A–F) and increased accumulation of C3 (G–L). Among astrocytes with the glaucoma OPTN-E50K mutation (M–O), reactivity led to a significant increase in the number of branches, along with a significant decrease in longest branch length as well as morphological complexity by Sholl analysis. (P) Induction of reactivity also exacerbated disease-related phenotypes such as a significant increase in autophagy proteins P62 and the LC3-II/I ratio. (Q–S) Similarly, among astrocytes with the Alzheimer’s PSEN1-N141I mutation, reactivity led to a significant increase in the number of branches along with a significant decrease in longest branch length, although no significant differences were observed in overall morphological complexity. (T–U) Astrocyte reactivity led to a decreased capacity for amyloid beta uptake compared to homeostatic PSEN1 astrocytes. (V–X) ALS SOD1-N139K astrocytes induced to reactivity exhibited an increased number of branches along with a significant decrease in the longest branch length, although no differences were observed in overall morphological complexity. (Y and Z) Reactive SOD1-N139K astrocytes exhibited more prominent aggregates of SOD1 protein and had significantly increased intracellular levels of SOD1. Data represent mean values ±SEM, two-tailed paired Student’s t test, ^∗p < 0.05, ^∗∗p < 0.01, and ^∗∗∗p < 0.001 vs. respective control astrocytes. Scale bar: 30 μm.