Human amyotrophic lateral sclerosis/motor neuron disease: The disease-associated microglial pathway is upregulated while APOE genotype governs risk and survival

Abstract

A key role for inflammation in amyotrophic lateral sclerosis/motor neuron disease (ALS/MND) has been identified. It is vital to assess which central nervous system structures are most affected and which inflammatory processes are responsible in humans. The inflammatory transcriptome was characterized in the cervical spinal cord and motor cortex in post-mortem frozen and formalin-fixed paraffin-embedded specimens from human sporadic ALS/MND and control cases using the nCounter® Neuroinflammation Panel. Archival data were reanalyzed and compared with the nCounter data. Immunohistochemistry was used to examine the inflammatory response in the spinal cord and motor cortex and validate changes found during transcriptomic analyses. In the spinal cord, marked inflammation was observed, while less inflammation was detected in the motor cortex. Examination of differentially expressed genes in the spinal cord highlighted TREM2, TYROBP, APOE, and CD163, as well as phagocytic pathways. In sporadic ALS/MND spinal cord, significant microglial reactivity and involvement of TREM2, ApoE (encoded by APOE), and TYROBP were confirmed, suggesting the involvement of the disease-associated microglial (DAM) phenotype. The corticospinal tracts showed greater inflammation than the ventral horns. The precentral gyrus of ALS/MND again showed less immune reactivity to disease when compared to controls. Finally, in the largest cohort assessed to date, we demonstrate an association between the APOE variant and ALS/MND risk, age of onset, and survival. We find confirmed associations between APOE ε3/ε3 and disease and between ε2/ε2 and absence of disease. Further, ε4/ε4 appears to be associated with earlier disease onset and a more aggressive course. We conclude that while there is widespread inflammation in the CNS in sporadic ALS/MND, this is more marked in the spinal cord, especially the corticospinal tract. The specific markers stress the DAM phenotype as having a key role together with a possible influx of somatic macrophages. In addition, APOE function and genotype may be relevant in ALS/MND.

Keywords: APOE; amyotrophic lateral sclerosis; disease‐associated microglia; inflammation; motor neuron disease.

FIGURE 1

sMND is associated with an increase in inflammatory signaling in the spinal cord—frozen and FFPE tissue. (A) Heatmap displaying the normalized and row‐scaled expression of 128 differentially expressed genes (p < 0.05 and FC >1.5) between MND and neurologically healthy cases. Genes are shown on the y‐axis. (B) Heatmap displaying the normalized and row‐scaled expression of 219 differentially expressed genes (p < 0.05 and FC >1.5) between MND and neurologically healthy cases. Genes are shown on the y‐axis. (C) Volcano plot for frozen spinal cord tissue showing fold change against the significance value. Of all the significant genes, 72 were upregulated, and 13 genes were downregulated (green points p < 0.05; FC >1.5; orange points FDR adjusted p < 0.05; FC >1.5, labeled points FDR adjusted p < 0.05; FC >2). (D) Volcano plot for FFPE spinal cord tissue showing fold change against the significance value. Of all the significant genes, 62 were upregulated, and 38 genes were downregulated (green points p < 0.05; FC >1.5; orange points FDR adjusted p < 0.05; FC >1.5, labeled points FDR adjusted p < 0.05; FC >2).

FIGURE 2

The motor cortex shows little inflammatory signaling in sMND. (A) Heatmap displaying the normalized and row‐scaled expression of 42 differentially expressed genes (p < 0.05 and FC >1.5) between MND and neurologically healthy cases. (B) Volcano plot showing fold change against the significance value. Green Points p < 0.05; FC >1.5. No genes reached significance following FDR correction.

FIGURE 3

IBA1 in the spinal cord of controls and ALS/MND cases. IBA1 labels perivascular macrophages (black arrows) and microglia (open arrows). In a control spinal cord, microglia tend to be ramified. In MND spinal cord, perivascular macrophages were swollen in some cases compared to the control. Microglia tended to show a mixture of ramified (open arrows) and amoeboid (gray arrows) morphology. (A) control ventral horn; (B) MND/ALS ventral horn; (C) control lateral corticospinal tract (LCST); D, sMND/ALS corticospinal tract. Scale bars = 50 μm.

FIGURE 4

Heterogeneity of HLA‐DR immunoreactivity in sMND cases. In the spinal cord, sMND cases varied greatly in the level of HLA‐DR staining present. Both cases (A & B) are sMND cases. However, as visible from the low magnification image, these cases showed very different levels of HLA‐DR‐positive microgliosis. (A1 and B1) show higher magnification images of the lateral corticospinal tract for these cases. In the low pathology case (A1) microglia were rounded (when compared to control) and showed thickened processes. In the high pathology case (B2) microglia were completely rounded and showed much denser staining. (A2 and B2) show higher magnification images of the ventral horn. In case A (low pathology) microglia again showed thick processes, swollen cell bodies, and an increased number of HLA‐DR‐positive microglia. In case B, HLA‐DR‐positive microglia were much rounder, many becoming ameboid, and microgliosis was much more severe. A and B scale bars =2.5 mm; A1, A2, B1, and B2 scale bars =100 μm.

FIGURE 5

Area density for a variety of monocyte/microglial markers in MND/ALS spinal cord. There is greater inflammation in the motor structures of the cord, namely the ventral horns (VH) and lateral corticospinal tracts (LCST) in MND/ALS, as shown by IBA1 (A), CD68 (B), HLA‐DR (C), and CD163 (D). There is some lower‐level inflammation seen in the sensory dorsal columns when assessed by CD163 and CD68.

FIGURE 6

ApoE immunoreactivity in the spinal cord. In control (A) and MND/ALS (B) spinal cord, at low power, ApoE is seen in higher levels in the gray matter compared to the white matter. In MND/ALS cases, ApoE is increased in the corticospinal tract (gray arrows) and ventral horns (black arrows). At higher power (C, control ventral horn; D, MND/ALS ventral horn), ApoE is present in endothelial cells (open arrows), perivascular macrophages (gray arrows) and parenchymal glial cells (black arrows) that have the appearance of astrocytes, as well as the background neuropil. There is also variable motor neuron staining (E): Some having a high signal (black arrow), others have a similar signal to the surrounding parenchyma (gray arrows) and a small number of neurons have no immunoreactivity (open arrows). This differential ApoE expression varied greatly between cases and was not associated with disease/control status. Scale bars: A,B = 2.5 mm; B,C = 50 μm; D = 100 μm.

FIGURE 7

Expression of TYROBP in the spinal cord. In control (A) and MND/ALS (B) spinal cord at low magnification, TYROBP immunoreactivity was present in the neuropil, at higher levels in the gray matter compared to the white matter. Glial staining in the white matter, particularly in the corticospinal tracts, was increased in MND/ALS cases. At higher magnification (C), TYROBP‐labeled motor neurons in the ventral horn (black arrow), and blood vessels (gray arrow), as well as various ramified (white arrow) and unramified glia (open arrow). Image taken from the ventral horn. In MND/ALS ventral horn (D,E), there was greater TYROBP expression in perivascular cells. Scale bars: A,B = 2.5 mm; D = 250 μm; C,E = 50 μm.

FIGURE 8

Area density for a variety of ApoE, TYROBP, and TREM2 in MND/ALS spinal cord. There is greater expression of ApoE (A) and TYROBP (B) in the motor structures of the cord, namely the ventral horns (VH) and lateral corticospinal tracts (LCST) in MND/ALS. There were no such intergroup differences in TREM2 (C). However, excess TREM2 expression was associated with slower disease progression (D).

FIGURE 9

TREM2 expression was minimal in the parenchyma. TREM2 immunoreactivity was minimal in the parenchyma, although some cases did show slightly greater labeling of the gray matter compared to white. At low power, there was little difference between control cases (A) and sMND cases (B). At higher power, TREM2 was seen in a few small, rounded cells, likely perivascular macrophages (C). The majority of neurons showed minimal TREM2 signal (open arrow, D). However, TREM2 immunoreactivity did label a small number of motor neurons more strongly in the ventral horn (black arrow). These were not present in all cases and did not appear to be associated with either sMND or control cases specifically. In the parenchyma (E), small TREM2 + ve granules were observed. Some granules were associated with glial or monocyte cells (open arrow). Others were not (gray arrow). Scale bars: B = 2.5 mm; C,D,E = 25 μm.

FIGURE 10

IBA1 in motor cortex. IBA1‐labeled perivascular macrophages (black arrows) and microglia (open arrows). In control cases, microglia were ramified with small cell bodies and fine processes. Similar patterns of expression were observed in the gray matter (A) and white matter (C). In MND/ALS (B, cortex; D, white matter), microglia had an activated morphology, with thicker cell bodies and swollen processes (open arrows); some microglia had also transitioned to the fully amoeboid state (gray arrows). Perivascular macrophages are labeled with black arrows. Expression was similar in gray matter (C) and white matter (D). Scale bar = 50 μm.