Cellular senescence in white matter microglia is induced during ageing in mice and exacerbates the neuroinflammatory phenotype

Cdkn2a-positive microglia are primarily annotated in cluster 7, a DAM-related cluster, in addition to clusters 3, 4 and 8.

“showed that Cdkn2a-positive microglia were primarily annotated in cluster 7 (a DAM-related cluster) in addition to other clusters (clusters 3, 4 and 8) (Fig. 4e, f).”

p16INK4a-expressing cells are predominantly accumulated in the white matter of old mice but not in that of young mice.

“p16INK4a immunohistochemical staining using a specific antibody (Supplementary Fig. 1d, e) confirmed that the p16INK4a-expressing cells were predominantly accumulated in the white matter of old mice but not in that of young mice (Fig. 1e, Supplementary Fig. 1f, g).”

Most microglia are primarily separated into three cluster groups characterised by homoeostatic genes, inflammatory cytokines, or DAM-related genes.

“Most microglia were primarily separated into three cluster groups characterised by different sets of transcription factors: clusters expressing homoeostatic genes (Clusters 0, 1 and 3), inflammatory cytokines (Clusters 2, 4, 5 and 7), or DAM-related genes (Clusters 6 and 8)5,7 (Fig. 3d, e, Supplementary Fig. 5a–f).”

• cites [5] Krasemann, S. et al. The TREM2-APOE pathway drives the transcriptional phenotype of dysfunctional microglia in neurodegenerative diseases. Immunity 47, 566–581.e9 (2017). sourceNot checked yet

  Not checked in the free run.

• cites [7] Keren-Shaul, H. et al. A unique microglia type associated with restricting development of Alzheimer’s disease. Cell 169, 1276–1290.e17 (2017). sourceNot checked yet

  Not checked in the free run.

The numbers of p16INK4a-expressing cells, DAM and p16INK4a-positive DAM are significantly lower in old GF mice than in old SPF mice despite no change in the total number of microglia.

“immunostaining analysis revealed that the numbers of p16INK4a-expressing cells, DAM and p16INK4a-positive DAM were significantly lower in old GF mice than those in old SPF mice despite no change in the total number of microglia (Fig. 4k-p).”

In the spinal cord white matter of old mice, about half of the microglia are p16INK4a-positive.

“Immunostaining revealed that in the spinal cord white matter of old mice, about half of the microglia were p16INK4a-positive, and most ( ~ 90%) of the Lgals3-positive DAM were p16INK4a-positive (Fig. 5e–g).”

p16INK4a-positive microglia in old mouse brains incorporate much less EdU than p16INK4a-negative microglia.

“However, p16INK4a-positive microglia in old mouse brains incorporated much less EdU than p16INK4a-negative microglia.”

Bulk RNA-seq of FACS-sorted microglia shows that DAM- and cell cycle-related gene expressions are suppressed in microglia of old GF mice compared to old SPF mice.

“bulk RNA-seq analysis performed on FACS-sorted microglia further supported that DAM- and cell cycle-related gene expressions were suppressed in the microglia of old GF mice compared to old SPF mice (Supplementary Fig. 8a-e).”

Increased p16INK4a and Il1b expression is observed in old spinal cords.

“Increased p16INK4a and Il1b (an inflammatory SASP factor) expression was also observed (Fig. 5c, d).”

Senescent DAM accumulation in the spinal cord is observed in the paralysis phase in SOD1 G93A mice.

“Furthermore, senescent DAM accumulation in the spinal cord was observed in the paralysis phase in SOD1 G93A mice47, recapitulating amyotrophic lateral sclerosis (ALS) pathogenesis (Supplementary Fig. 9a–e).”

• cites [47] Gurney, M. E. et al. Motor neuron degeneration in mice that express a human Cu,Zn superoxide dismutase mutation. Science (80-.). 264, 1772–1775 (1994). sourceNot checked yet

  Not checked in the free run.

Over 70% of Lgals3-positive DAM are positive for p16INK4a in the corpus callosum of old mice.

“Over 70% of Lgals3-positive DAM were positive for p16INK4a in the corpus callosum of old mice (Fig. 3l).”

In the spinal cord white matter of old mice, most (~90%) Lgals3-positive DAM are p16INK4a-positive.

“Immunostaining revealed that in the spinal cord white matter of old mice, about half of the microglia were p16INK4a-positive, and most ( ~ 90%) of the Lgals3-positive DAM were p16INK4a-positive (Fig. 5e–g).”

In forced microglial proliferation experiments, the number of EdU-positive microglia during repopulation is significantly lower in p16INK4a-positive microglia than in p16INK4a-negative microglia.

“This further confirmed that the number of EdU-positive microglia during repopulation was significantly lower in p16INK4a-positive microglia than in p16INK4a-negative microglia (Supplementary Fig. 4e-g).”

Genes associated with ageing or cellular senescence are strongly expressed in the DAM-related cluster.

“Genes associated with ageing or cellular senescence32 were also strongly expressed in the DAM-related cluster (Supplementary Fig. 5g)”

• cites [32] Hu, Y. et al. Replicative senescence dictates the emergence of disease-associated microglia and contributes to Aβ pathology. Cell Rep. 35, 109228 (2021). sourceNot checked yet

  Not checked in the free run.

Expression of SASP factor genes, including Il1b and Cxcl10, is increased in old mice's brain regions where p16INK4a is highly expressed.

“Furthermore, the expression of SASP factor genes, including Il1b and Cxcl10, was also increased, especially in the old mice’s brain regions where p16INK4a was highly expressed (Supplementary Fig. 1b, c).”

p16INK4a mRNA and protein are highly expressed in the brain's white matter regions, such as the corpus callosum and medulla oblongata, in old mice.

“Further evaluation by immunoblotting and RT-qPCR revealed that p16INK4a mRNA and protein were highly expressed in the brain’s white matter regions, such as the corpus callosum and medulla oblongata in old mice (Fig. 1c, d).”

Old p16-luc mice exhibit a marked increase in bioluminescence signal, especially in the pons and medulla oblongata regions, compared to young mice.

“Old p16-luc mice exhibited a marked increase in bioluminescence signal, especially in the pons and medulla oblongata regions, compared to young mice (Fig. 1a, b, Supplementary Fig. 1a).”

p16INK4a-expressing cells are positive for Iba1, a microglial marker.

“immunostaining analysis confirmed that the p16INK4a-expressing cells were positive for Iba1, a microglial marker (Fig. 2h, i).”

Microglia sorted from old mice's brains show higher p16INK4a mRNA and protein expression than those of young mice.

“microglia sorted from the old mice’s brains showed higher p16INK4a mRNA and protein expression than those of young mice (Fig. 2f, g, Supplementary Fig. 3a–c).”

Most DAM express p16INK4a at the EAE peak.

“Notably, p16INK4a-positive microglia and Lgals3-positive DAM accumulated at the EAE peak, and most DAM expressed p16INK4a (Fig. 5h–j).”

p16INK4a expression is significantly reduced in the medulla oblongata of old germ-free mice compared to old specific-pathogen-free mice.

“we first compared p16INK4a expression in the medulla oblongata of the old mice, raised under specific pathogen-free (SPF) or germ-free (GF) conditions, and found that p16INK4a expression was significantly reduced in the medulla oblongata of old GF mice (Fig. 4a).”

The composition of the major microglia subset in old mice shows a different pattern from that in young mice.

“The composition of the major microglia subset in the old mice showed a different pattern from that in the young mice (Fig. 3a-c).”

Bioluminescence signal is markedly increased in old mice's spinal cords.

“Analysis using p16-luc mice confirmed that the bioluminescence signal markedly increased in the old mice’s spinal cords (Fig. 5a, b).”

p16INK4a and Lgals3 are frequently co-expressed in Iba1-positive microglia.

“Immunostaining analysis also confirmed that p16INK4a and Lgals3, another DAM marker, were frequently co-expressed in Iba1-positive microglia (Fig. 3j, k).”

γH2AX levels increase in old mice's microglia.

“the γH2AX levels (another cellular senescence marker) increased in the old mice’s microglia (Supplementary Fig. 3d)”

Cd11c-positive microglia highly express p16INK4a and DAM-related genes, including Cd11c, Apoe and Axl, compared with Cd11c-negative microglia.

“Cd11c-positive microglia highly expressed p16INK4a in addition to various DAM-related genes, including Cd11c, Apoe and Axl, compared with Cd11c-negative microglia (Fig. 3i).”

More cells with high expression of Cdkn2a or DAM-related genes are observed in old SPF mice than in old GF mice.

“More cells with high expression of Cdkn2a or DAM-related genes were observed in old SPF mice than in old GF mice (Fig. 4g-j, Supplementary Fig. 7f-n).”

Cd4+ T cell infiltration is significantly reduced in the spinal cords of p16 KO mice during the EAE acute phase.

“In the spinal cords of p16 KO mice, Cd4+ T cell infiltration was significantly reduced during the EAE acute phase (Supplementary Fig. 10a–k), as was the demyelination progression (Fig. 5n, o).”

There is a difference in phenotype intensity between systemic p16-KO mice and microglia/macrophage-specific p16-KO mice.

“We found that there is a difference in phenotype intensity between systemic p16-KO mice and microglia/macrophage-specific p16-KO mice (Fig. 5k, p).”

Demyelination progression is reduced in the spinal cords of p16 KO mice.

“In the spinal cords of p16 KO mice, Cd4+ T cell infiltration was significantly reduced during the EAE acute phase (Supplementary Fig. 10a–k), as was the demyelination progression (Fig. 5n, o).”

Not all DAM are senescent microglia.

“However, our results indicate that although senescent microglia are present in DAM, not all DAM are senescent microglia; some senescent microglia are not classified as DAM (Figs. 3k, 4o, 5f, 5i, Supplementary Fig. 9d).”

Blockade of cellular senescence does not significantly decrease the Cd11c-positive DAM number.

“Furthermore, blockade of cellular senescence did not significantly decrease the Cd11c-positive DAM number (Supplementary Fig. 10b).”

Some senescent microglia are not classified as DAM.

“However, our results indicate that although senescent microglia are present in DAM, not all DAM are senescent microglia; some senescent microglia are not classified as DAM (Figs. 3k, 4o, 5f, 5i, Supplementary Fig. 9d).”

The degree of paralysis in EAE mice is significantly attenuated in p16 KO mice compared to WT mice.

“In contrast to WT mice, the degree of paralysis in EAE mice was significantly attenuated in p16 KO mice (Fig. 5k–m).”

Clusters expressing WAM-specific genes were not found in the scRNA-seq analysis.

“Although we could not find clusters expressing WAM-specific genes in our scRNA-seq analysis (Supplementary Fig. 13)”

The proportion of Cdkn2a-expressing T-cells is increased in the white matter of aged mice.

“Interestingly, our data showed that the proportion of Cdkn2a-expressing T-cells was increased in the white matter of aged mice (Supplementary Fig. 2f)”

Microglia expressing DAM-related genes have a senescent signature in the white matter of old mice.

“indicating that microglia expressing DAM-related genes have a senescent signature in the white matter of old mice.”

Cellular senescence is induced in microglia in the white matter during ageing.

“These results, together with the observation that the γH2AX levels (another cellular senescence marker) increased in the old mice’s microglia (Supplementary Fig. 3d), indicated that cellular senescence is induced in the microglia in the white matter during ageing.”

Microglial senescence is predominantly induced in DAM.

“Microglial senescence is predominantly induced in DAM (Figs. 3l, 4p, 5g, 5j, Supplementary Fig. 9e)”

Microglia in the white matter represent the main cell type that undergoes cellular senescence in the CNS during ageing and EAE.

“We demonstrated that microglia in the white matter represent the main cell type that undergoes cellular senescence in the CNS during ageing and EAE.”

Microglia with high p16 promoter activity exhibit signatures associated with cytokine production and T cell infiltration.

“Moreover, microglia with high p16 promoter activity exhibited signatures associated with cytokine production and T cell infiltration37.”

• cites [37] Talma, N., Gerrits, E., Wang, B., Eggen, B. J. L. & Demaria, M. Identification of distinct and age-dependent p16High microglia subtypes. Aging Cell 20, 1–13 (2021). sourceNot checked yet

  Not checked in the free run.

The induction of DAM and cellular senescence is influenced by commensal bacteria.

“We have also shown, by comparing GF and SPF mice, that the induction of DAM and cellular senescence is influenced by commensal bacteria (Supplementary Fig. 12).”

• AssumptionRests on an assumption: Differences in microglial senescence between germ-free and specific-pathogen-free mice are attributable to commensal bacteria rather than to other differences between the husbandry conditions.

It is unlikely that the mechanisms that lead to DAM are the same as those that cause cellular senescence.

“It is therefore unlikely that the mechanisms that lead to DAM are the same as those that cause cellular senescence.”

The DAM transition and cellular senescence processes are independent of each other.

“However, these two processes are independent of each other.”

Common triggers can lead to a parallel transition from homoeostatic microglia to DAM and cellular senescence in ageing and diseased states.

“These findings suggest that common triggers can lead to a parallel transition from homoeostatic microglia to DAM and cellular senescence in ageing and diseased states.”

Microglial senescence is mostly observed in the white matter DAM.

“Notably, microglial senescence was mostly observed in the white matter DAM.”

Suppression of senescence induction by knockout of the p16INK4a gene in microglia ameliorates the neural impairment phenotype in the EAE model.

“Suppression of senescence induction by knockout of the p16INK4a gene in microglia ameliorated the neural impairment phenotype in the EAE model (Fig. 5p-r).”