Loss of homeostatic microglial phenotype in CSF1R-related Leukoencephalopathy

Abstract

Microglia are resident macrophages of the central nervous system, and their unique molecular signature is dependent upon CSF-1 signaling. Previous studies have demonstrated the importance of CSF-1R in survival and development of microglia in animal models, but the findings are of uncertain relevance to understanding the influence of CSF-1R on microglia in humans. Hereditary diffuse leukoencephalopathy with spheroids (HDLS) [also known as adult onset leukoencephalopathy with spheroids and pigmented glia (ALSP)] is a neurodegenerative disorder primarily affecting cerebral white matter, most often caused by mutations of CSF1R. Therefore, we hypothesized that the molecular profile of microglia may be affected in HDLS. Semi-quantitative immunohistochemistry and quantitative transcriptomic profiling revealed reduced expression of IBA-1 and P2RY12 in both white and gray matter microglia of HDLS. In contrast, there was increased expression of CD68 and CD163 in microglia in affected white matter. In addition, expression of selective and specific microglial markers, including P2RY12, CX3CR1 and CSF-1R, were reduced in affected white matter. These results suggest that microglia in white matter in HDLS lose their homeostatic phenotype. Supported by gene ontology analysis, it is likely that an inflammatory phenotype is a key pathogenic feature of microglia in vulnerable brain regions of HDLS. Our findings suggest a potential mechanism of disease pathogenesis by linking aberrant CSF-1 signaling to altered microglial phenotype. They also support the idea that HDLS may be a primary microgliopathy. We observed increased expression of CSF-2 in gray matter compared to affected white matter, which may contribute to selective vulnerability of white matter in HDLS. Our findings suggest that methods that restore the homeostatic phenotype of microglia might be considered treatment approaches in HDLS.

Keywords: Adult leukoencephalopathy with spheroids and pigmented glia (ALSP); CSF-1; CSF-1R; CSF-1R-related leukoencephalopathy; Hereditary diffuse leukoencephalopathy with spheroids (HDLS); Immunohistochemistry; Microglia; RNA expression profiling.

Fig. 1

Alterations in microglial and macrophage phenotype in HDLS. Microscopic comparison of 4 macrophage markers in a representative HDLS case (case #7) and a representative normal control (case #8). IBA-1 expression (a & b) is decreased in the white matter of HDLS. Similarly, P2RY12 expression (c & d) is decreased in white matter of HDLS. Many amoeboid CD68 (e & f) macrophages are observed in HDLS lesional white matter. Presence of parenchymal CD163 (g & h) positive cells are confined to the lesional white matter of HDLS, with almost no cells detected seen in normal parenchyma. GM = gray matter, WM = white matter. Dotted line separates gray and white matter. Scale bar = 100 μm

Fig. 2

Analysis of microglial phenotypes in different cortical layers. Graph panel of quantified (with Aperio software and specifically designed algorithms) DAB-based immunohistochemistry images of: macrophage markers IBA-1 (a), CD68 (b), CD163 (c) in HDLS (n = 12 for all stains and regions) and Normals (n = 12 for all stains and regions). Data represented as Tukey Box and Whisker plots. * indicates statistically significant p-value< 0.05

Fig. 3

Analysis of sub-population of macrophages in HDLS. a & b Immunofluorescence double stain of CD68 (Alexa Fluor 568 (red)), and IBA-1 (Alexa Fluor 488 (green)) with DAPI nuclear staining (blue). White arrowhead indicates an example of an IBA-1, CD68 double labelled amoeboid macrophage. White arrow indicates an example of a CD68 positive, IBA-1 negative amoeboid cell. Scale bar = 50 μm. c & d Immunofluorescent double stain of CD68 (Alexa Fluor 568 (red)), and S-100 (Alexa Fluor 488 (green)), with DAPI nuclear staining (blue). White arrowhead indicates an example of a CD68, S-100 double labelled cell with nuclear and cytoplasmic S-100 expression. Asterisk indicates example of a CD68 positive, S-100 negative cell. Scale bar = 25 μm

Fig. 4

Identification of brain region-specific transcriptional changes in HDLS. (a & b) Heat map and PCA map of total transcription profiles in different brain regions. The same color codes were used to display the same groups in (a) and (b). (a) The heat map shows that HDLS-cases are clearly separated from control-cases by their molecular signatures in frontal cortex white matter. Up- and down-regulated transcripts in HDLS are shown in red and blue, respectively. (b) The PCA components from FW are proximal to the CW than FG indicating the similarity in cellular structure-specific gene expression patterns. PCA 67.1%, PC1 = 33.3%, PC2 = 24%. (c-e) PCA analysis of total transcripts in each brain areas, FW, FG and CW. Control cases (n = 5) are shown in blue, and HDLS cases (n = 6) are shown in red marbles. FW, FG and CW samples were acquired from the same individual assigned with the number from 1 to 12 by individual, and assigned with the alphabet a-c by brain area. (c) PCA map of frontal cortex white matter, PCA 65.7%, PC1 81.5%, PC2 22.4% (d) PCA of frontal cortex gray matter, PCA 79.9%, PC1 46.6%, PC2 26.3% (e) PCA of cerebellum white matter, PCA 58.9%, PC1 26.7%, PC2 17.6%. FW, frontal cortex white matter; FG, frontal cortex gray matter; CW, cerebellum white matter; Ctl, control

Fig. 5

Enrichment analyses with differentially expressed genes in frontal cortex white matter. a The number of up- and down-regulated genes in each brain region. FW, FG and CW are frontal cortex white matter (green), frontal cortex gray matter (pink) and cerebellum white matter (blue), respectively. b The overlap between HDLS specific transcriptional changes in different brain areas. A Venn diagram displays the number of differentially expressed genes in HDLS. The same color code was used to indicate each brain region. The total numbers of altered genes are shown in parenthesis. Overlapping portions of circles indicates the number of genes shown in multiple brain regions. Seventy-five genes are FW-specific and 24 genes are FG-specific. c Enriched pathway maps in frontal cortex white matter. Enrichment analysis was performed using 75 FW-specific genes. The table with top 25 pathways displays multiple CSF-1-involved pathways and a lot of immune response-related pathways. d Gene ontology process in frontal cortex white matter. Enrichment analysis was performed using 75 FW-specific. The table with top 25 processes includes cell migration, cell differentiation and axonal regeneration-related processes

Fig. 6

Comparison of molecular microglial signature in different brain regions of HDLS patients. The relative expression level of microglial-specific transcripts, IBA-1, P2RY12 and CX3CR1. All three genes have a trend of decreasing their level in HDLS cases. However, it is statistically significant only in frontal cortex white matter (P2RY12, p = 0.0033, CX3CR1, p = 0.0086). The relative expression level of CSF-1 and CSF-1R. The level of CSF-1 significantly increased both in white and gray matter. However, the level of CSF-1R significantly decreased only in white matter while it showed trend of increasing in gray matter. The box and whisker plot displays median values with the error bars of maximum and minimum values. Two tailed student t-test were used *p < 0.05, **p < 0.01, ***p < 0.005