Complement 3+-astrocytes are highly abundant in prion diseases, but their abolishment led to an accelerated disease course and early dysregulation of microglia

Abstract

Astrogliosis and activation of microglia are hallmarks of prion diseases in humans and animals. Both were viewed to be rather independent events in disease pathophysiology, with proinflammatory microglia considered to be the potential neurotoxic species at late disease stages. Recent investigations have provided substantial evidence that a proinflammatory microglial cytokine cocktail containing TNF-α, IL-1α and C1qa reprograms a subset of astrocytes to change their expression profile and phenotype, thus becoming neurotoxic (designated as A1-astrocytes). Knockout or antibody blockage of the three cytokines abolish formation of A1-astrocytes, therefore, this pathway is of high therapeutic interest in neurodegenerative diseases. Since astrocyte polarization profiles have never been investigated in prion diseases, we performed several analyses and could show that C3⁺-PrP^Sc-reactive-astrocytes, which may represent a subtype of A1-astrocytes, are highly abundant in prion disease mouse models and human prion diseases. To investigate their impact on prion disease pathophysiology and to evaluate their potential therapeutic targeting, we infected TNF-α, IL-1α, and C1qa Triple-KO mice (TKO-mice), which do not transit astrocytes into A1, with prions. Although formation of C3⁺-astrocytes was significantly reduced in prion infected Triple-KO-mice, this did not affect the amount of PrP^Sc deposition or titers of infectious prions. Detailed characterization of the astrocyte activation signature in thalamus tissue showed that astrocytes in prion diseases are highly activated, showing a mixed phenotype that is distinct from other neurodegenerative diseases and were therefore termed C3⁺-PrP^Sc-reactive-astrocytes. Unexpectedly, Triple-KO led to a significant acceleration of prion disease course. While pan-astrocyte and -microglia marker upregulation was unchanged compared to WT-brains, microglial homeostatic markers were lost early in disease in TKO-mice, pointing towards important functions of different glia cell types in prion diseases.

Keywords: A1-astrocytes; Microglia; Neurotoxicity; Prion diseases.

Fig. 1

A1-astrocytes are highly abundant in human prion diseases and a prion mouse model. a Representative immunofluorescence staining with the A1-astrocyte marker C3d showed high abundance of A1 astrocytes in different brain regions in terminally prion diseased mice (n = 3 individual animals per group), (age matched control mice; n = 3); red = C3d; blue = DAPI; scale bar: 25 μm. b Representative frontal cortex sections of control or sCJD individuals showed that GBP2-positive A1-astrocytes can be only detected in human prion disease brains. GBP2-signal co-localize with GFAP-positive astrocytes; (CJD cases; n = 5), (age and gender matched control; n = 5); red = GBP2; green = GFAP; scale bar: 50 μm. c Representative fluorescence staining of human frontal cortex brain sections show that Iba1-positive microglia do not co-localize with GBP2, but keep close contact with GBP2⁺ astrocytes (see high magnification insert); GFAP (green) or Iba1 (green) and GBP2 (red); scale bar: 50 μm, close up: 20 μm

Fig. 2

Triple-KO leads to significant acceleration of prion disease course a Western blot analysis of PrP^Sc after proteinase K digestion of brain tissue at 80 and 110 days p.i. and at clinical prion disease. Quantification of signal intensity showed that WT and Triple-KO mice do not show significant differences at all three investigated time points (n = 4 independent animals per group and time point) day80 p = 0.347; day 110 p = 0.126; clinical prion disease p = 0.297. b Kaplan-Meier survival curve of RML-prion infected WT- (n = 8 individual animals) and TKO-mice (n = 10 individual animals), Mantel-Cox log rank *** p = 0.0003. c Titers of prion infectivity as measured by bioassay in tga20-mice are similar in brains of WT- or TKO-mice 80 days post prion infection (p = 0.1451). Brain homogenates of two individual infected mice per group were injected into 4 individual tga20-mice, each

Fig. 3

Prion pathology in TKO- is similar to WT-mice at clinical time points a Determination of PrP^C expression by Western blot and subsequent quantification of PrP^C normalized to β-actin in age matched non-infected mice brain homogenates showed similar expression levels at the age of 200 days (corresponding to the clinical time point). b Representative H&E staining of hippocampus and thalamus of clinical prion diseased mice and age matched non-infected controls; scale bar: 50 μm. c Semi-quantitative determination of spongiosis levels showed no differences between clinical TKO versus WT mice (n = 4 individual mice each). In contrast, uninfected TKO or WT mice do not display spongiosis (n = 3 individual mice each)

Fig. 4

A1-astrocyte markers are significantly altered in TKO mice upon prion infection a qPCR expression analysis of microglia and astrocyte disease markers in thalamus tissue from terminally sick mice and age matched control confirmed the knockout of TNFα in TKO mice, but significant upregulation in diseased WT-mice (p = 0.0004). A1-astrocyte markers are significantly less upregulated in diseased TKO-mice (C3 p = 0.0006; GBP2 p = 0.0001). In contrast, microglia disease marker Clec7a was upregulated in both infected groups (p = 0.0023) n = 3 independent mice/group. b Representative staining of GFAP (red), C3 (green) and DAPI (blue) in Thalamus tissue at clinical prion disease. c Quantification of positive staining area (μm² × 1000) showed significant upregulation of GFAP in both, prion infected WT- and TKO-mice (p = 0.0003), while C3 is significantly upregulated in WT-mice only (p = 0.0005); n = 3 independent from Fig. 4a animals/group

Fig. 5

Analyses of immune cell and astrocyte marker expression. a Fold change induction of reactive astrocyte specific transcripts split to Pan-, A1-, and A2-specific cassettes for all experimental groups including age matched control groups. Data are mean +/−SEM using Nanostring nCounter data. b Fold change induction of reactive astrocyte specific transcripts split to Pan-, A1-, and A2-specific cassettes showing significant differences between infected WT- and TKO-mice. Data are mean +/−SEM using microfluidic qPCR. c Heatmap of transcripts of Nanostring encounter expression analysis from terminally sick prion infected WT- and TKO-mice. Although subtle, there are clustering differences between infected WT- versus TKO-mice. d PCA plot of Nanostring expression data. Unit variance scaling is applied to rows; singular value decomposition with imputation is used to calculate principal components. X and Y axis show principal component 1 and principal component 2 that explain 50.6 and 12% of the total variance, respectively. All control animals regardless of genotype cluster closely together, while prion infected animals are clustered according to genotype; n = 3 individual animals per group for all analyses

Fig. 6

GLP-1R is upregulated in TKO-mice but not in WT-mice after prion infection a Western blot analysis of GLP-1R at clinical prion disease and age-matched controls (n = 3 individual mice per group). b Quantification of GLP-1R levels normalized to β-actin showed that GLP-1R is significantly increased in TKO-mice at clinical prion disease (p = 0.0001); n = 3 independent animals per group. c Western blot analysis of GLP-1R at day 80 p.i. and age-matched controls (n = 3 individual mice per group). d Quantification of GLP-1R levels normalized to β-actin showed that GLP-1R is increased in TKO-mice at day 80 post infection, but significantly downregulated in WT-mice (p = 0.0022); n = 3 independent animals per group

Fig. 7

Homeostatic microglia markers are significantly altered in pre-clinical prion diseased TKO-mice at day 80 p.i. a Representative immunostaining of GFAP, Iba1, P2ry12, or TMEM119 in thalamus in brain sections of TKO- and WT-mice and age matched control; scale bar: 50 μm; high magnification: 25 μm. b Quantification of positive staining area (μm² × 1000) of GFAP, Iba1, P2ry12, and TMEM119 of brain sections at 80 days post infection. While GFAP and Iba1 staining intensities are unchanged between prion infected TKO and WT-mice, both are significantly upregulated compared to uninfected control (GFAP p = 0.0058; Iba1 p = 0.0030). In contrast, P2ry12⁺ and TMEM119⁺-microglia are significantly dysregulated in prion infected TKO-mice only (P2ry12 p = 0.0069; TMEM119 p = 0.0089) n = 3–4 individual mice/group