mTOR activation induces endolysosomal remodeling and nonclassical secretion of IL-32 via exosomes in inflammatory reactive astrocytes

Abstract

Astrocytes respond and contribute to neuroinflammation by adopting inflammatory reactive states. Although recent efforts have characterized the gene expression signatures associated with these reactive states, the cell biology underlying inflammatory reactive astrocyte phenotypes remains under-explored. Here, we used CRISPR-based screening in human iPSC-derived astrocytes to identify mTOR activation a driver of cytokine-induced endolysosomal system remodeling, manifesting as alkalinization of endolysosomal compartments, decreased autophagic flux, and increased exocytosis of certain endolysosomal cargos. Through endolysosomal proteomics, we identified and focused on one such cargo-IL-32, a disease-associated pro-inflammatory cytokine not present in rodents, whose secretion mechanism is not well understood. We found that IL-32 was partially secreted in extracellular vesicles likely to be exosomes. Furthermore, we found that IL-32 was involved in the polarization of inflammatory reactive astrocyte states and was upregulated in astrocytes in multiple sclerosis lesions. We believe that our results advance our understanding of cell biological pathways underlying inflammatory reactive astrocyte phenotypes and identify potential therapeutic targets.

Keywords: Astrocytes; Endolysosomal system; Extracellular vesicles; IL-32; Inflammatory reactive astrocytes; Neuroinflammation; mTOR.

Fig. 1

Endolysosomal function is perturbed in inflammatory reactive astrocytes. a Schematic of modeling inflammatory astrocyte reactivity in vitro. b Heatmap of changes in the expression of genes encoding lysosome-localized proteins in different hiPSC-derived astrocyte models treated with ITC or similar treatments vs vehicle control. c Top GO Cellular Component terms enriched among the top 1000 downregulated genes in iAstrocytes treated with ITC; endolysosomal system-related terms are highlighted by asterisks. d Representative images of LysoTracker staining in live iAstrocytes or immunostaining of LC3 or LAMP1 in fixed and permeabilized iAstrocytes; scale bar = 75 μm. e Quantification of imaging experiments shown in d (n = 3 wells per condition). f, g Representative immunoblot bands against LAMP1 (f) or LAMP2 (g). h, i Quantification of immunoblot experiments shown in f, g (n = 3 wells per condition). j Measurement of the acidity of LAMP1⁺ endolysosomal compartments using FIRE-pHLy via flow cytometry (see Methods, Additional file 8). k Measurement of autophagic flux with the LC3ΔG-RFP/LC3-GFP fluorescent reporter from Kaizuka et al. [25]. l Immunoblot against LC3 demonstrating LC3-I and LC3-II bands. m Quantification of autophagic flux from the LC3-II bands in l; error bars reflect the 95% confidence interval associated with the standard error of the mean; individual data points not shown because the quantities of interest are differences between means, with no biologically meaningful pairing of individual data points across conditions. P values where shown were calculated using two-sided Student’s t test

Fig. 2

Perturbation of endolysosomal function is accompanied by remodeling of the endolysosomal proteome. a Schematic of endolysosomal proteomics workflow (n = 3 wells per condition). b Volcano plot of endolysosomal proteomic data. c Scatterplot comparing log₂-fold change of endolysosomal vs total cell lysate protein abundance in ITC-treated iAstrocytes compared to vehicle-treated iAstrocytes (IP: immunoprecipitation). d Representative histograms of cell-surface LAMP1 staining intensity in vehicle- or ITC-treated iAstrocytes compared to isotype control staining intensity measured by flow cytometry. e Median fluorescence intensity (MFI) of cell-surface LAMP1 measured by flow cytometry (n = 3 wells per condition); P value by two-sided Student’s t test. f Cell-surface LAMP1 or LysoTracker staining in iAstrocytes treated with increasing doses of bafilomycin A1 (n = 3 wells per condition; error bars reflect the standard error of the mean)

Fig. 3

Multi-phenotypic CRISPR-based screening identifies mTOR as a central regulator of endolysosomal system function. a Schematic of CRISPR-based screening workflow. b Pearson correlation of gene scores (see Methods, Additional file 8) of hits from the LAMP1 vs LysoTracker screens (n = 2 biological replicate screens per condition). c Enrichment analysis against MSigDB Hallmark Pathways terms of the top 20 hits from each screen; terms pertaining to mTOR are highlighted with stars. d Heatmap of gene scores of the hits overlapping with the highlighted MSigDB terms in c. e, f Median fluorescence intensity (MFI) of phospho-S6 (e) or phospho-4E-BP1 (f) staining in ITC- vs. vehicle-treated iAstrocytes measured by flow cytometry. g Representative immunoblot bands corresponding to mTOR, phospho-S6, total S6, phospho-ULK1, and phospho-AKT1 in ITC- vs vehicle-treated iAstrocytes in the presence of different mTOR inhibitors. h Quantification of immunoblotting experiments (n = 6 wells per condition for phospho-S6/S6, n = 3 wells per condition for phospho-ULK1, n = 6 wells per condition for phospho-AKT1/pan-AKT; P values from two-sided Student’s t test). k, l Cell-surface LAMP1 (k) or LysoTracker (l) MFI in ITC-vs. vehicle-treated iAstrocytes in the presence of mTOR inhibitors measured by flow cytometry (n = 6 wells for DMSO treated, n = 3 wells for all other conditions; P values calculated only for ITC-treated conditions by linear regression with adjustment for multiple testing by Holm’s method, shown only if significant)

Fig. 4

Cytokine-induced remodeling or pharmacological disruption of the endolyosomal system results in increased exocytosis of endolysosomal contents. a–d Cell-surface LAMP1 or LysoTracker median fluorescence intensity (MFI) measured by flow cytometry (a), extracellular CTSB concentration measured by electrochemiluminescence-based immunoassay (b), extracellular IL-32 concentration measured by ELISA (c), or abundance of extracellular mito-EVs measured by flow cytometry (d) in ITC- vs. vehicle-treated iAstrocytes transduced with non-targeting control (NTC) sgRNAs or sgRNAs targeting genes encoding common (MTOR) or unique mTORC1 (RPTOR) vs mTORC2 (RICTOR) subunits, with or without co-treatment with bafilomycin A1. P values were calculated by linear regression with correction for multiple testing using Holm’s method, shown only when significant

Fig. 5

Extracellular IL-32 co-fractionates with extracellular vesicles likely derived from multivesicular body exocytosis. a Immunoblots against consensus extracellular vesicle (EV) markers (CD63, CD81, Hsc70, Caveolin-1) or common contaminants (e.g. ApoA-I) in EVs isolated from iAstrocyte conditioned media or total cell lysate after vehicle vs. ITC treatment. b EV size distribution measured by nanoparticle tracking analysis. c Immunoblot against IL-32 in EVs isolated from iAstrocyte conditioned media or total cell lysate. d Representative images of dual immunostaining against LAMP1 together with IL-32 or CD63 together with IL-32; scale bar = 60 μm. e, f Extracellular IL-32 concentration measured by ELISA in conditioned media from ITC- vs. vehicle-treated iAstrocytes transduced with non-targeting (NTC) sgRNAs or sgRNAs targeting genes encoding proteins involved in multivesicular body exocytosis (e), or treated with small molecules known to inhibit (PI4KIIIβ inhibitor 3) or not inhibit (PI-273) exosome biogenesis (f). P values were calculated by linear regression with correction for multiple testing by Holm’s method, shown only when significant

Fig. 6

IL-32 regulates the polarization of inflammatory reactive astrocytes and is upregulated in astrocytes in neuroinflammatory conditions. a, b Proportion of IL-1/IL-6-responsive (VCAM1-/C3 +) or TNF/interferon-responsive (VCAM1 + /C3-, VCAM1 + /C3 +) inflammatory reactive astrocyte polarizations (a) or their associated cytokines (b) in ITC- vs. vehicle-treated iAstrocytes transduced with non-targeting (NTC) sgRNAs or sgRNAs targeting IL32. c Log-scaled IL32 expression in astrocytes found in normal tissue vs. multiple-sclerosis lesions derived from pseudobulk analysis of snRNA-seq data from Macnair et al.; n = 15 for healthy control gray matter, n = 15 for normal-appearing gray matter, n = 15 for gray matter lesion, n = 22 for healthy control white matter, n = 18 for normal-appearing white matter, n = 17 for active lesion, n = 27 for chronic active lesion, n = 13 for chronic inactive lesion, n = 8 for relapsing lesion, n = 23 for not specified. d Representative immunostaining of IL-32 and GFAP in white matter brain tissue from patients with hypoxic-ischemic encephalopathy (HIE); scale bar 50 μm. e Percent GFAP + , OLIG1 + , or NeuN + cells among IL-32 + cells in HIE brain tissue (n = 3 patients); P values calculated via beta regression. f, Schematic of ITC-induced, mTOR-dependent endolysosomal remodeling and associated exocytic activity. P values were calculated using the Mann–Whitney U test in a, c, and e, and using the two-sided Student’s t test in b